Handbook Tailings and Mine Waste 2008

TAILINGS AND MINE WASTE '08

Hinshaw_book.indb i 9/10/2008 3:58:37 AM

Hinshaw_book.indb ii 9/10/2008 3:58:37 AM

PROCEEDINGS OF THE 12TH INTERNATIONAL CONFERENCE, VAIL, COLORADO, USA, 19–22 OCTOBER 2008

Tailings and Mine Waste '08

Hinshaw_book.indb iii 9/10/2008 3:58:37 AM

CRC Press/Balkema is an imprint of the Taylor & Francis Group, an informa business

© 2009 Taylor & Francis Group, London, UK

Typeset by Vikatan Publishing Solutions (P) Ltd., Chennai, IndiaPrinted and bound in the USA by Edwards Brothers, Inc, Lillington, NC.

All rights reserved. No part of this publication or the information contained herein may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, by photocopying, recording or otherwise, without written prior permission from the publisher.

Although all care is taken to ensure integrity and the quality of this publication and the information herein, no responsibility is assumed by the publishers nor the author for any damage to the property or persons as a result of operation or use of this publication and/or the information contained herein.

Published by: CRC Press/Balkema P.O. Box 447, 2300 AK Leiden, The Netherlands e-mail: [email protected] www.crcpress.com – www.taylorandfrancis.co.uk – www.balkema.nl

ISBN: 978-0-415-48634-7 (hbk)ISBN: 978-0-203-88230-6 (ebook)

Hinshaw_book.indb iv 9/10/2008 3:58:38 AM

v

Table of Contents

Preface ix

Organization xi

Keynote paper

Future tailings management strategies – High time we took the high road 3A.B. Fourie

Design, operation, and disposal

Simple mine waste management planning for successful operations in the Canadian low arctic 17P.M. Bedell & K.S. Willis

James Creek Tailings Facility relocation for the gold quarry open pit expansion 25A. Boye & J. Young

Heat and mass balance modeling of a subaqueous tailings disposal facility 35G.R. Eykholt, J.B. Manchester, S.V. Donohue & J.C. Cherry

Water chemistry and metal cycling in a subaqueous tailings disposal facility 49J.B. Manchester, G.R. Eykholt, S.V. Donohue & J.C. Cherry

Heap leach lixiviant flow—myth versus reality 63J.R. Kunkel

Challenges in heap leach pad design: Consideration of thermal conditions 73A.L. Hudson & T. Meyer

Innovative mine waste disposal in two distinctly different settings 83I. Wislesky & A. Li

Pipeline design for paste and thickened tailings systems 95R. Cooke

Efficient dewatering solutions on vibrating screens 101M. Doerffer & R. Heinrich

High pressure washing technology Hydro-Clean 113S. Palombo & J. Varela

Remote monitoring of a high hazard coal waste impoundment in mountainousterrain case study 125J.D. Quaranta, L.E. Banta & J.A. Altobello

Tailings and Mine Waste '08© 2009 Taylor & Francis Group, London, ISBN 978-0-415-48634-7

Hinshaw_book.indb v 9/10/2008 3:58:38 AM

vi

Geotechnical considerations

Experimental characterization of the influence of curing under stress on thehydromechanical and geochemical properties of cemented paste backfill 139E. Yilmaz, T. Belem, M. Benzaazoua & B. Bussière

Guidelines for stabilizing historic mine workings 153J.F. Lupo

Landscape design for soft tailings deposits 165G. McKenna & V. Cullen

Liners, covers, and barriers

Liner system design for tailings impoundments and heap leach pads 177J.F. Lupo

Evaluation of geomembrane puncture potential and hydraulic performance inmining applications 189C. Athanassopoulos, A. Kohlman, M. Henderson & J. Kaul

Field performance of cover systems 199N.R. Amorim, R.F. Azevedo, O.R. Ferreira, A.G.C. Ribeiro & I.D. Azevedo

Capping the tailings impoundment at the Jamestown Mine 207J.C. Isham

Water management and geochemistry

Applying numerical hydrogeochemical models as decision support tools for mineclosure planning 221L.E. Eary, R. Jakubowski, J. Eshleman & A. Watson

Arsenic species & its binding forms in tailing sediments 233T. Naamoun & B. Merkel

Geochemical characterization of proposed waste dumps over time and space 243L. Breckenridge, A. Hudson, S. Poos & D. Thompson

Stochastic prediction of mine site water balance, Gilt Edge Mine Superfund Site,Lawrence County, South Dakota 253M. Nelson, S. Fundingsland, G. Hazen, P. Hight & V. Ketellapper

Using water balance tools for site design, operation and expansion management 265A. Trautwein

Desiccation and Rheology in cyclic surface deposition of gold paste tailings 269P. Simms, B. Fisseha, J. Henriquez & R. Bryan

Remediation and reclamation

Past, present and future for treating selenium-impacted water 281J. Gusek, K. Conroy & T. Rutkowski

Reclamation and closure cost planning and estimation and the mining life cycle 291L.E. Boxill & T.E. Martin

Hinshaw_book.indb vi 9/10/2008 3:58:38 AM

vii

Policies, procedures, and public safety

Uranium tailings facility design and permitting in the modernregulatory environment 305K. Morrison, J. Elliott, J. Johnson & B. Monok

Tailings structure closure for economic development in Ghana 315M.B. Thorpe, F. Nyame & B.A. Addo

Lessons learned from tabletop reviews of emergency action plans for high hazarddams in West Virginia, USA 325J.D. Quaranta, H.M. Childers & P. Myles

Working for responsible management of tailings facilities 337E. Gardiner & D. Gladwin

Developments in the safety and security of mining industry dams 345J.W. Fredland

Life cycle assessment and tailings management trade-off studies—concepts 355D. van Zyl

Case histories

Tailings closure at BHPBilliton’s San Manuel operation design and closureconstruction San Manuel, Arizona 361D. Ortman

Case study: Site-wide water balance of the Pierina Gold Mine, Peru 369L. George, W. Ludwick & J. Chahbandour

Reclamation of the Panna Maria uranium mill site and tailingsimpoundment: A 2008 update 381C.L. Strachan & K.L. Raabe

Mining impacts: A case study 393R.K. Will & W.E. Motzer

Design & construction of an evaporation pond at a historic uranium mining facility 401T.A. Chapel, C. Woodward & R. Jolley

Gold quarry North Waste Rock Facility slide investigation and stabilization 409R.J. Sheets & E.E. Bates

An overview of the Grouse Creek Mine tailings impoundment closure 423F. Moye, S. Rogers, D. Poulter & B. Tritthart

Stabilized upstream tailings dam and converted into a filtered tailings facility 437M.F. Veillette, T.E. Martin & S.A. Larreta

Author index 449

Hinshaw_book.indb viii 9/10/2008 3:58:38 AM


ix

Preface

This marks the twelfth annual Tailings and Mine Waste Conference. The purpose of these confer-ences is to provide a forum for discussion and establishment of dialogue among people involved in the mining industry and environmental community regarding tailings and mine waste. Previ-ous conferences have been successful in providing opportunities for formal and informal discus-sion, exhibits by equipment and instrumentation companies, technical exhibits, and general social interaction.

This year’s conference includes over 40 papers. These papers address the important issues faced by the mining industry today. These proceedings will provide a record of the discussions at the conference that will remain of value for many years.

Hinshaw_book.indb ix 9/10/2008 3:58:38 AM

Hinshaw_book.indb x 9/10/2008 3:58:38 AM


xi

Organization

Organized by the Department of Civil and Environmental Engineering, Colorado State University, Fort Collins, Colorado in conjunction with the University of Alberta, Edmonton, Alberta and the University of British Columbia, Vancouver, British Columbia.

ORGANIZING COMMITTEE

Daniel Overton(Committee Chair) Engineering Analytics, Inc., Fort Collins, ColoradoLarry Cope SRK Consulting, Inc., Fort Collins, ColoradoNeil Eurick Golder Associates, Inc., Lakewood, ColoradoChristopher Hatton URS Corporation, Denver, ColoradoMike Henderson Tetra Tech, Inc., Golden, ColoradoAndrew Robertson Robertson GeoConsultants, Inc., Vancouver, B.C.Charles D. Shackelford Colorado State University, Fort Collins, ColoradoClint Strachan MWH, Fort Collins, ColoradoBryan Ulrich Knight Piésold Consulting, Elko, Nevada

John D. NelsonHonorary Conference Chair Engineering Analytics, Inc., Fort Collins, Colorado

Hinshaw_book.indb xi 9/10/2008 3:58:38 AM

Hinshaw_book.indb xii 9/10/2008 3:58:38 AM

Keynote paper

Hinshaw_book.indb 1 9/10/2008 3:58:38 AM


3

Tailings and Mine Waste ‘08© 2009 Taylor & Francis Group, London, ISBN 978-0-415-48634-7

Future tailings management strategies – High time we took the high road

A.B. FourieUniversity of Western Australia, Perth, Australia

ABSTRACT: Managing the vast quantities of mine waste, particularly tailings, which are gener-ated by many of today’s mining operations will continue to provide an enormous challenge to the designers and operators of Tailings Storage Facilities (TSFs). This paper suggests that the correct approach to future management strategies is to resist the temptation to do only what is required, and to go beyond currently accepted standards, improving TSF design and operations so that these facilities do not become the Achilles heel of the mining industry. We should ultimately be able to claim that tailings are indeed managed in a responsible and benign manner. Anything less could ultimately result in the industry’s continuing licence to operate being placed in jeopardy.

1 INTRODUCTION

Procedures for the analysis and design of tailings storage facilities (TSFs) are now reasonably well established. There is an extensive bibliography of papers dealing with aspects of TSF perform-ance, and some Universities even incorporate modules in undergraduate courses that deal with some of the fundamentals of TSF analysis and design (of necessity these modules are usually very limited). On the occasion of the revival of the conference series dealing with Tailings and Mine Waste after a hiatus of four years, it is an opportune time to consider some of the improvements made in TSF management (where the term management is taken to encompass analysis, design, construction and operation) and where improvements are possible. The discussion will address two general themes; one dealing with specific technical issues and the other dealing with the less tangible, but equally important aspects of TSF management that perhaps are best described by the term stewardship.

The title of the paper is loosely based on a concept that will be well known to many South Africans who lived through the demise of the apartheid years and the transition to a democracy. In the years prior to the first democratic election in 1994, Clem Sunter provided two scenarios that he considered faced South Africa, with both being possible. The ‘High Road’ was the route of negotiation, leading to a political settlement, with the ‘Low Road’ being the route of confronta-tion, leading to a civil war and a resulting wasteland. Fortunately the High Road option prevailed. Although tailings management cannot in any sense be considered to be as significant in scope as the deliberations of a transition to democracy, the concept is a useful one to frame a debate about future tailings management strategies. In this paper, it is suggested that although we presently do many things quite well, it is certainly possible to do them a whole lot better. Technology is avail-able, both in conventional geotechnical engineering as well as other disciplines, such as rheology, fluid mechanics and offshore geotechnics, to mention a few, that we should continue to draw on to improve the design and management of TSFs. It is likely that in future the TSF will become the focus of groups opposed to new mining developments, as they often provide such an appealing target, being large, visible and unfortunately not aesthetically pleasing. In the rapidly changing sociological climate, where considerations of water allocation, land use and energy consumption are going to take centre stage to a greater and greater degree, we need to ensure that the TSF is designed, built, operated and closed according to the best technology currently available.


4

2 APPROPRIATE STABILITY ANALYSES

An effective stress analysis of the slopes of a TSF require determination of the effective stress strength parameters c′ and φ′. These parameters are determined in a routine fashion using labora-tory tests such as the direct shear box test or a triaxial compression test. To complete an effec-tive stress analysis, it is necessary to know the pore water pressure that is likely to be acting on any assumed potential failure plane. The pore pressure throughout the region of potential failure is usually evaluated by having a number of appropriately located piezometers around the TSF periphery, from which the depth to the phreatic surface may be determined. (Note this calculation is not always done correctly, as the effect of seepage on total head is sometimes ignored. How-ever, that is not the subject of this paper and will not be discussed further here). A series of either force or moment balance calculations are then carried out (usually using a commercially available computer programme) and the critical potential failure surface located and the factor of safety against failure evaluated. The vast majority of reports dealing with stability of TSFs rely solely on this approach to stability evaluation. It has largely served the industry well, but when dealing with contractant material (even if it does not undergo significant strain softening) it can be shown to be fundamentally and significantly unsafe. It is highly likely that some of the catastrophic TSF failures that have occurred can be traced to an over-reliance on effective stress methods of stabil-ity analysis.

The effective stress method of analysis (ESA) uses pore pressures derived from either calcula-tion or measurement (as above) in the evaluation of stability. What is more, it implicitly assumes that this pore pressure does not change during the shearing that takes place when a slope failure develops. It effectively assumes that shearing takes place at a rate that is slow enough for sufficient drainage to occur such that the pore pressure does not change.

It is difficult to tell whether many geotechnical engineers fully appreciate the underlying assumptions in the ESA method of analysis, but it is possible that some do not. Why then has this not resulted in more problems in the past? The reason can probably be attributed to the fact that most tailings, certainly the material that is deposited around the perimeter of a well-operated TSF in which segregation of particles occurs along the beach, are in a dilative state. Shearing of material such as this and hence the conventional ESA approach provides a reasonable approach for dilative materials, as it considers the more unsafe condition that exists when full dissipation of excess pore water pressures occur.

The response of a dilative material to undrained loading can be further considered in terms of a stress path plot, as shown in Figure 1. Here the axes are the mean effective stress ′ = +p ( ) /σ σ1 32 3′ ′ (the abscissa) and a measure of shear stress, q = (σ

1′−σ

3′). If we consider a specimen that is loaded

from an initially isotropic state of stress (point M) in an undrained triaxial compression test, the total stress path moves along a straight line at a slope of 3:1, as shown by the line MN. In terms of effective stresses, an initially dense (dilative) specimen will produce a stress path such as that

Figure 1. Illustration of effective stress paths for dilative (curve A) and contractive (curve B) material.

Shear stress

Mean effective stress

A

B

Total stress path

M

N

Failure envelope


5

shown by curve A, where the negative increments in pore water pressure result in the stress path deviating to the right of the total stress path. As loading continues, the effective stress path reaches the allowable limit, which is given by the effective stress failure envelope, after which further load-ing is accompanied by further dilation, with the stress path now moving up the failure envelope (still strain hardening) until ultimate shear failure occurs.

During shearing of contractant tailings the opposite occurs. As shown by path B in Figure 1, generation of positive excess pore water pressure during shearing results in an effective stress path that intersects the failure envelope at a lower shear strength than that implicitly assumed in the ESA method of analysis. Consistent with this observation, the ESA method of analysis would overestimate the available shear strength and hence the factor of safety of a TSF where a substantial volume of contractant material was within the potential failure zone. It is highly likely that this is exactly what happened in the Merriespruit TSF catastrophe in South Africa in 1994 (Wagener, 1997), where the judicial inquest held after the failure revealed that for extended peri-ods of operation, the decant pond had been in very close proximity to the slope of the TSF where the failure occurred. Deposition of tailings into a pond of water produces a very loose matrix of particles, which is invariably contractant when sheared and disturbance of the tailings in this state at the Merriespruit TSF appears to have triggered the flow failure that occurred, which resulted in 17 deaths.

There is now a great deal of literature that discusses the different response of a dilative mate-rial to undrained loading compared with a contractive material (see Been and Jefferies, 2006, Fourie and Papageorgiou, 2001). The importance of this difference in behaviour still appears to be unappreciated by many tailings engineers, perhaps because the vast majority of tailings behave in a dilative manner, particularly the coarser fraction that is traditionally deposited around the perimeter of a TSF.

The main consequence of the above difference in behaviour is that a material that is initially potentially contractant and is loaded undrained will experience strain-softening, during which process the load-carrying capacity of an element of this material decreases continuously. This results in a shedding of load to adjacent elements, with these adjacent elements consequently becoming overloaded, undergoing strain softening and shedding load to other elements or zones of material. This results in what has become known as progressive failure. Once it has initiated it is very difficult to prevent it developing further, and one essentially has to wait until the material to which load is transferred is sufficiently strong (or dilative) to accommodate the imposed load-ing. This is what happens when a TSF undergoes a liquefaction failure, where collapse and flow of material can occur for some minutes, or even longer, as load shedding results in an ever-increasing volume of material becoming overloaded and failing due to strain-softening. Once dilative mate-rial is exposed to loading, the process may be arrested. This simple description is consistent with the photographs of the aftermath of liquefaction failures in TSFs, where the perimeter of the failure scar typically has vertical or near-vertical faces. The material in these faces was initially sufficiently dense to not experience collapse failure, with a further improvement in strength prob-ably resulting from some drainage of the pore fluid occurring during the development of the flow failure.

To distinguish between material that is initially potentially liquefiable (contractant) and suf-ficiently dense to dilate upon undrained loading, it is common to present data in terms of a plot of the Steady State Line (SSL). This is illustrated in Figure 2, where the axes are again mean effective stress on the abscissa and void ratio on the vertical axis. The SSL divides material that is dilative (below the SSL) from material that is potentially contractive (above the SSL). The line is not hori-zontal, because as the mean effective (confining) stress increases, most materials can be at a lower void ratio and still be contractive.

The SSL is usually obtained by carrying out a number of undrained triaxial compression tests. The undrained nature of loading means that the void ratio of specimens remains unchanged during loading and in terms of the axes plotted in Figure 2, the paths can only be in a horizontal direc-tion. If the specimen is dilative, the path will move to the right, terminating on the SSL, whereas if it is contractive, the path will move to the left (with an associated decrease in mean effective


6

stress as the excess pore water pressure increments are positive) until it reaches the SSL. In prac-tice it is a little more difficult to accurately locate the SSL. Initially dense specimens tend to fail along localised slip planes and it is almost impossible to measure localised void ratios along these planes. This non-homogeneity of dilative specimens undergoing undrained loading has meant that most determinations of the SSL for materials described in the literature have been derived from contractant material, where the specimen tends to remain more uniform during undrained loading and it is thus easier to pick the termination value of the test in terms of mean effective stress. The difficulty with this latter approach is preparing specimens in the laboratory that are contractive in nature. Most material preparation techniques, especially when dealing with cohesionless materi-als, inevitably produce specimens that are relatively dense, resulting in dilative behaviour upon shearing. Despite these problems it is still possible to define the location of the SSL with sufficient certainty as long as a reasonable number of tests are carried out (as a rule of thumb, a minimum of five undrained triaxial tests are suggested).

It is important to realise that the SSL is not unique for a particular type of tailings material, although it is indeed unique for a particular particle size distribution of any tailings (at a particular pH value). Generally, as the percentage of fines increases, the SSL becomes lower in this plot, meaning that in order to be non-contractive, a material with a high fines content must be at a lower void ratio than material with little or no fines. Results illustrating this effect are given in Lade and Yamamura, 1997, and Fourie and Papageorgiou, 2001.

3 DAMBREAK ANALYSIS

The negative publicity surrounding the failure of a TSF that results in release of tailings and water into the surrounding environment is immense, particularly in the era of the instant newsbreak. Aside from the unacceptable risk of human injury or fatality, and the potential for environmental damage, the negative image of the entire mining industry that results from a TSF failure represents a real risk to the continuing social licence to operate in many countries. Although the statement reported in the MMSD study that, ‘mining is one tailings dam break from extinction’, may be an exaggeration, the passion generated by opposition to TSFs on the basis of the potential for a dambreak is immense. One only has to look at the reports posted on the internet by various NGOs opposed to mining, where the fear of a dambreak is regularly highlighted to generate opposition to a particular operation, to realise the risk to continuing operations posed by a perceived dambreak risk. Failures have long memories. For instance, anyone working on a TSF design in the Philip-pines will invariably be reminded (by opponents to the project) of Marcopper; in South Africa it is Merriespruit and Bafokeng; in Europe it is Baia Mare, Stava and Los Frailes.

The requirement to undertake a dambreak study for any TSF is mandatory in many countries and states. These studies are intended to consider the potential for a dambreak to occur, and more importantly, to consider the risk posed to downstream communities and the downstream environment. So how good are we are undertaking these dambreak studies? In reviewing reports

Figure 2. Illustration of steady state line (SSL) dividing contractive material from dilative material.

Void ratio

Mean effective stress

Steady state line (SSL)

Contractive zone

Dilative

Stress paths during undrained loading


7

describing dam break evaluations and associated hazard assessment studies, it became apparent that this question cannot be answered with conviction. The reason is that most consulting engi-neering companies who undertake dam break studies do so using propriety ‘in-house’ methodolo-gies. The claim is made (perhaps quite reasonably) that publishing the details of their particular procedure would jeopardise their competitive edge. The difficulty with this approach is of course that the client has no way of independently validating the reasonableness of the findings pre-sented. At a workshop of the Association of State Dam Safety Officials in April 1999, the number one research priority was identified as being dam failure analysis, with the need to address, ‘flow characteristics of various types of tailings’ as a key requirement. The workshop further recognised that the degree of difficulty was a key factor holding back this research. In other words, perhaps it was just too hard.

When considering our capabilities in this field, it is clear just how difficult this problem is. There have, of course, been many significant contributions to the field, such as those dealing with rock avalanches and debris flows (Iverson, 1997 and Hungr, 2000, 2006) and with mudflows (Coussot, 1997, and Coussot et al., 1998). However, the ability to accurately predict the likely inundation area downstream of a TSF failure still seems a long way off. For example, recent work by Naef et al. (2006) evaluated a number of flow-resistance relations using a one-dimensional finite ele-ment model, and although they were able to reproduce analytical results for a dam break event, this required calibration of the model. In addition, the simulation was only one-dimensional, with the extension to two (or three) dimensions still some way off. Thus even the highly constrained condi-tion analysed by Naef et al. (2006) presents predictive problems, as the model has to be extensively calibrated against other results, a luxury that is not readily available for TSFs. In the absence of the research needed to provide more reliable and robust predictive models, industry naturally requires an interim, pragmatic approach. It might be argued that this is currently the case, where propri-etary, in-house solutions are used, as discussed earlier. However, the need for pragmatism does not remove the need for rigour. All models should be subjected to scrutiny, and evaluated against previously published studies, or even better, against published laboratory and field data. Without this, we will probably be consigned to using models of water retaining dams as a fallback position, as this clearly constitutes the most conservative assumption possible. The trouble is, it tends to predict downstream inundation that is intolerable, and usually clearly impossible. This is when the ‘pragmatic’ approach has to be adopted, resulting in uncertainty and lack of transparency.

What then are possible solutions? Obviously focussed research would provide very useful input. However, this will take time and of course suitable funding. It is therefore not likely to hap-pen in the short term. In the interim, a workshop could be convened where modellers, researchers and consultants could be invited to predict the downstream inundation of a number of study sites. Such studies have proven useful in a range of disciplines, such as evaluation of various constitu-tive models, but require careful planning, as sufficient, and realistic, data must be provided to assist the participants to calibrate their models. Another pragmatic approach is to discard the concept of modelling as it is currently practised, at least until the required fundamental research has been completed, and use simpler simulations taking advantage of the power of tools such as geographical information systems (GIS). Using documented profiles of post-failure tailings dam breaks (e.g. Blight, 1998), together with an estimate of the volume of tailings likely to flow from a TSF, a digital terrain model (DTM) of the downstream area could be ‘filled up’ with the tailings lost from the breach. While this ignores kinematic effects of the mudflow, it provides a simple ‘cut and fill’ approach that many engineers would be familiar with. It also enables relatively rapid scenario evaluation, because the volume of failed material can be altered and a sensitivity study carried out.

This brings us to the issue of estimating the volume of material that might flow from a TSF breach. As discussed by Naef et al. (2006), the event volume (volume of flowing material) is vitally important in determining the extent of downstream inundation. In the two catastrophic failures that occurred in South Africa, the volumes of material were about 3 million m3 in the Bafokeng failure, and about 600 000 m3 in the Merriespruit failure. The reasons for the difference are not known, but expected to be a combination of the state of the tailings at the time of the failure (where ‘state’


8

is used in the geotechnical sense, indicating whether or not the void ratio was above or below the Steady State Line) and the volume of water stored on the TSF surface at the time of the failure.

4 ESTIMATING THE STATE OF TAILINGS USING IN-SITU TESTS

It is extremely difficult to obtain undisturbed samples of tailings material for use in laboratory testing programmes. This is true even for tailings containing some proportion of clayey material. Some promising work has been carried out in Canada on the use of in-situ freezing and coring of cohesionless tailings (see Hoffman et al., 2000), but the technique will likely remain only a research tool for many years because of the expense involved, not to say the technical challenges in warmer parts of the world. In-situ testing techniques will remain the backbone of our approach to evaluating the state of a tailings deposit, for many years to come. Although the Standard Penetra-tion Test (SPT) is still used by some practitioners because of its ease and low cost, the method is extremely limited and will not be discussed here. A far more valuable tool is the cone penetration test, particularly when a pore pressure sensor is included with the cone, (so-called CPTu tests). Some promising results have also been obtained with the seismic cone device, but this seems to suffer from producing a ‘smeared’ result of a volume that is too large to identify problematic zones in a TSF, and the database of field data versus laboratory data is very limited. It is therefore suggested that the CPTu technique remains our best option for characterising the state of tailings in-situ. Techniques for undertaking an evaluation of this type are given by Been and Jeffries (2006) amongst others, and in future we will hopefully see more comparisons of this approach versus other interpretations, as well as comparisons with laboratory test data.

In doing comparisons of this type, however, we would be well advised to acknowledge (once again) our lack of knowledge. The offshore oil industry is probably far more advanced than the tailings management industry and useful lessons can be learned from their experiences. An exam-ple of this is the development of ‘full-flow’ penetrometers, such as those described by Chung and Randolph (2004) and DeJong et al., (2004). These penetrometers take forms such as a sphere (the ball penetrometer), or a T-bar. As these penetrometers are pushed into the ground, material is able to flow around them (analogous to water flow around a bridge pier) and flow of this type is then amenable to the derivation of virtually exact analytical solutions. Contrast this with the cone penetrometer, which requires the use of a bearing capacity factor, N

k, which typically varies

between 11 and 20 for tailings, when calculating in-situ undrained shear strength. If we are going to improve our predictions of in-situ state, we surely need to be using instruments that minimise uncertainties in fundamental interpretation, before we try to incorporate factors such as anisot-ropy, layering, etc.

Another factor that is conveniently ignored is the potential for partial drainage to occur during penetration, and the distortion this induces into calculation of parameters such as the coefficient of consolidation. Teh and Houlsby (1991) produced techniques for interpreting dissipation curves from penetrometer test results, and we should be incorporating approaches such as these at every opportunity.

5 GEOTECHNICS OF TAILINGS WITH HIGH CLAY CONTENT

Many large mining operations produce tailings with little or no clayey tailings. Although some tailings streams may have clay-sized particles (less than 2 μm), the particles are often merely very finely ground rock flour, as opposed to clay minerals. This is generally the case when intact, unweathered rock is mined, whether it is underground or openpit. However, when oxide material is mined, true clay minerals are usually present in the tailings. It is thus not unusual for gold tail-ings from places like West Africa, Indonesia and Papua New Guinea to have a high proportion of clay minerals. Other industries where clay minerals occur in the tailings include mineral sands, diamonds, phosphate and oil sands. In all these cases the problems associated with the presence


9

of clay minerals have presented enormous challenges for the geotechnical engineers involved in the management of facilities storing this material. There have been cases when the TSF has filled up much more rapidly than expected because of poor settling rates, where capping has been virtu-ally impossible (witness the recently released draft directive governing new requirements for the regulation of oil sands fluid fine tailings (ERCB, 2008)) or where the risks of instabilities have been a major concern.

Probably our greatest deficiency in dealing with these materials is sufficient understanding of the complexities of clay mineralogy, and the associated interaction of clayey tailings with various process waters and additives such as flocculants. A good example of this is the work by Boger and co-workers (e.g. Nguyen and Boger, 1998) who investigated the rheological behaviour of bauxite residue over a number of years. The effect of changing the pH, or otherwise altering the surface chemistry of the material had a dramatic effect on the yield stress (which is effectively the undrained shear strength, but is a strength measure in the very low strength region, typically below 1 kPa) of various tailings and other suspensions. Although studies of this type are numer-ous, and the effects reasonably well recognised, it is surprising how often laboratory testing pro-grammes take no cognisance whatsoever of these effects. Samples of dry tailings are sent for laboratory testing with instructions to prepare samples for testing using tap water. Although this may be reasonable in some circumstances, the importance of the process water on the strength and compressibility characteristics of some clayey tailings can render tests using tap water completely meaningless. It is high time we developed strategies for screening tailings material (e.g. establish-ing an appropriate database) in order to determine when aspects such as water chemistry, pH, etc can be ignored and when they need to reproduce site conditions accurately. Otherwise we are sometimes simply throwing money away on meaningless tests, and worse still, producing data that is incorrect and misleading.

6 ALTERNATIVE AND EMERGING TECHNOLOGIES

It is now more than three decades since Robinsky (1975) touted the idea of improving tailings management by thickening it to a consistency such that no free water (or ‘bleed’ water) separated from the tailings upon placement. The tailings would be non-segregating upon placement and no retaining embankments or dykes would be required. For almost two decades after this initial suggestion, very little development of the concept occurred, although the Kidd Creek operation in Canada persevered with Robinsky’s concept. Gradually more and more operations decided to experiment with high density thickened tailings, such as the bauxite industry (see Cooling (2007) as an example) and some of the benefits began to become apparent. An annual series of seminars on the topic was initiated by the Australian Centre for Geomechanics, and the 11th in the series was recently held in Botswana, Africa. Many operations have reported positive outcomes from implementing thickening schemes, although some difficulties have also emerged, such as the need in some cases to cover (or cap) a greater surface area of tailings when the Central Thickened Discharge (CTD) technique is used. The technique has been applied to a complete range of types of tailings, including diamond, gold, bauxite, mineral sands, coal and copper (see Williams et al., 2008), although the more fine-grained materials appear to be better suited to management using this approach than very coarse grained tailings.

A significant aspect to emerge from the use of high density tailings (the term ‘paste’ is avoided here, as it is usually more relevant to the material that is increasingly being used for underground backfilling, where cement is added to full-plant tailings, which is reticulated underground at solids content as high as 80%) is the need to better integrate the preparation and transport components of the process with the deposition component. Unlike conventional tailings disposal, where the TSF operator has to accept that the solids content of the material arriving at the TSF will be variable and develop operational techniques for dealing with this variability, with high-density tailings the control of solids content (and more relevant, the control of rheological parameters, particularly the yield stress) is critical to the operation of a TSF. Relatively small changes in yield stress can


10

result in very significant changes in beach slope, with severe effects on the impoundment opera-tion. Management of high-density TSF facilities require a greater understanding of preparation and transportation methods, and thus adds to the skills base required of a designer and operator of these facilities.

Perhaps it is inevitable with any new and relatively revolutionary technology that both pro-ponents and opponents tend to take rather extreme viewpoints, and this is also unfortunately the case with high-density tailings. Proponents, such as some vendors, tend to advocate high-density (or indeed paste) tailings for almost all new and retrofit operations, irrespective of the real merits of the technology in each application. On the other hand, there are some who see the technol-ogy as just another fad and proceed to develop destructive counter-arguments to the use of the technology, based on unrepresentative hypothetical situations. Thankfully the majority of tailings practitioners have taken the pragmatic approach of dealing with each potential application on its individual merits, and many of the papers presented at the annual seminars on Paste and Thick-ened Tailings referred to earlier bear testimony to this assertion. This surely is the responsible approach to take, particularly as the use of high-density tailings certainly does provide signifi-cant water savings compared with conventional approaches (see McPhail and Brent, 2007) and in the current climate of water scarcity in many mining areas of the world, this can only be a good thing.

As with many of the issues raised in this paper, the only truly rational way to deal with the viability of high-density tailings in a particular application is through appropriate and on-going training and education of those tasked with making these decisions.

7 DESIGNING FOR PERPETUITY

When a designer of a closure facility for a TSF signs off the drawings for the design, there is one certainty. The designer expects the design to outlast his/her remaining working life. No doubt many geotechnical engineers have reviewed closure designs that are based on claims the design has a lifetime of 1000 years. Modeling results are presented for erosion profiles that can be expected in the year 3008, based on a technique that utilises two or three days of rainfall simulation, coupled with a modeling technique that has not been tested over even a single decade successfully. Clearly a case of some good science mixed with a lot of wishful thinking.

The requirement of closure designs for TSFs is to provide security that the facility will remain stable and non-polluting for a very long time after closure, indeed in perpetuity. However, we should not be lulled into a pretence of knowledge when it is not founded on solid science. Rather, we should acknowledge the limitations of our current methods and endeavor to improve them to a point where we can make predictions with much improved degrees of certainty. In the meantime we should include in our predictions of the long-term performance of closed facilities some indication of the uncertainty of our predictions. Only by quantifying our lack of certainty can we get to a point where advances in our predictive capability become assured, rather than accidental.

Advances in our ability to predict the long-term performance of closed facilities will only be possible if we are able to incorporate the likely variability of future climatic scenarios. Predictions of this kind are subject to great uncertainty, but it is nevertheless possible to quantify the extent of this uncertainty and to undertake sensitivity analyses of the effect of these uncertainties on the likely performance of closure designs. Of course it may not provide clients with the answers they would like to hear, but at least it should provide them with answers that regulators will not view with skepticism or disbelief. Despite the difficulties in making predictions of future climatic conditions, great strides are being made in these fields. As an example, the work of the CSIRO Marine and Atmospheric Research group is producing information that is helpful in this regard (see Bryson, 2007).

Once again, it is suggested that through taking the high road, acknowledging the extent of our predictive capability and the degree of uncertainty that is currently inevitable, that our designs


11

will be more credible and easier to support when subjected to public scrutiny. There is inevitably a risk associated with this approach; a risk that the degree of uncertainty is considered unacceptable by stakeholders, and a potential new project is rejected. However, engineers pride themselves on dealing with uncertainty and incomplete information, so there is no reason this should not be the case with long-term climate variability and closure designs. What is needed, however, is certainty of the extent of the uncertainty that must be dealt with.

8 DEALING WITH THE OGRE: ACID DRAINAGE MANAGEMENT

The issue of acidic and metallic drainage generation and seepage is without doubt the greatest negative image that has plagued the mining industry over the past two or three decades. No mining or geoenvironmental conference is complete without pictures of bright-coloured, contaminated seepage, and many new projects have to have a prime focus on this troublesome issue. Techniques for predicting likely, and thus preventing acid drainage, have improved significantly in recent times, and organisations such as CANMET, ACMER and others have all played useful roles in this development. The issue remains a thorny one. Particularly when insufficient non-acid forming (NAF) waste rock or tailings is available to buffer the potentially acid-forming (PAF) material. In cases such as this, the choice of cover system becomes a critical one, with a key differential being whether a ‘dry’ system or a ‘wet’ system is chosen. These choices may appear to be relatively clear cut, with a dry system being used in arid and semi-arid climates, and a wet system being preferred in more temperate, and tropical regions of the world. However, the implications of the choice can be enormous. If we consider a wet cover system in a tropical environment, the concept is to main-tain a lake or pond over the tailings, which by virtue of the exclusion of oxygen remains non-acid generating. The problem is that the barrier that ensures the tailings remains inundated, i.e. the retaining embankment (or embankments) has to perform this function virtually in perpetuity. This means that active management of the facility will be required post-closure, with regular monitor-ing and inspections being required. While this may not appear to be an insurmountable problem, it requires that adequate funds be allocated to fund this kind of activity once the mine closes and there is no income stream from the mining operation. A workable solution to this dilemma would appear to be to incorporate some form of income generating scheme as part of the closure design, so that reliable and ongoing funding is assured.

The same argument applies to a dry cover system, although the consequences of a deteriora-tion of such a system are unlikely to be as severe as with a wet system. Whereas deterioration of a wet system could result in a catastrophic breach and flowslide failure (because of the saturated nature of the tailings plus the presence of a large body of water on the TSF), a dry system will generally degrade due to processes of wind and water erosion. These processes are certainly not high hazard processes, but have the potential to severely degrade the surrounding environ-ment. Once again, allowance will need to be made for ongoing monitoring and maintenance, and funding will need to be provided for these functions. Factors that can negatively impact on the performance of a dry cover system are many and varied (Campbell, 2004, Fourie and Tibbett, 2007) and cognizance must be taken of these external effects when choosing a suitable closure system. Cover designs that rely on the exclusion of vegetation, for example, because of the potential damage that can occur due to root development, are likely doomed to fail-ure because this (vegetation exclusion) cannot be guaranteed in perpetuity. It makes more sense to work together with experts in various fields, such as botany and zoology, to develop a cover system that is robust enough to withstand the range of environmental forces it will inevitably be subjected to. An excellent example of how the performance of a cover system can (and will) change with time is provided by the work of Benson et al., (2007) who instrumented a number of experimental landfill cover systems at fourteen sites across the United States. They showed changes of more than two orders of magnitude in the hydraulic conductivity of some of these covers, in as little as five years. One can only speculate on the changes that might be expected over 1000 years.


12

9 OPERATIONAL ASPECTS OF TSFs: THINGS THAT MAY BE OUTSIDE YOUR ZONE OF INFLUENCE

TSFs are unlike most other geotechnical structures, in that they continue to get larger with time, and the factor of safety generally tends to decrease as a consequence. This is unlike most other geotechnical structures, where processes such as consolidation of the underlying foundation tend to improve conditions with time. TSFs thus require constant and on-going monitoring. Fortunately this is easily achieved, as operational personnel are on site on a continual basis. What is required is appropriate training of these personnel, so that they are able to recognize and react to aspects of unacceptable performance before these issues become critical. One only has to read the literature on failures such as those that occurred at Stava in Italy and Merriespruit in South Africa, to appre-ciate how timely response to visible signs of poor facility performance could have prevented the terrible accidents that occurred subsequently.

It may well be argued that it is not the job of the TSF designer to ensure that proper training of operational personnel occurs, but it certainly makes sense to do so from the simple viewpoint of stewardship. No one would like to see a design that they were responsible for deteriorate, or worse still, fail, because something avoidable had not been attended to. Along with the need for appro-priate training, is the industry-wide problem of change management. During the commissioning and early years of operations on a TSF, many problems usually occur and are dealt with at the time by operational personnel. Too often, however, these solutions are not properly documented and the details reside in the memories of the site supervisor and other staff. When these staff retire or resign, the historical aspects of the TSF are probably lost for ever, with potentially damaging con-sequences. For example, knowledge of how a vertical tower decant system had become damaged and subsequently sealed with a concrete plug can have very important implications with regard to decisions such as whether the TSF can be safely raised beyond its initial design height. Without full knowledge of all the important events that occurred on a particular TSF, decisions such as this are sometimes (unknowingly) made in ignorance.

The issues raised in this section are best dealt with in a formal manner, and the trend is to develop and maintain an appropriate TSF Operations Manual, which needs to be comprehensive (but not verbose) and easily accessible. Most operations today tend to have a system of some type in place, but recent experience has highlighted the problems that occur when one key person decides to leave the operation, and the knowledge that resides with that person has not been adequately recorded and documented.

10 CONCLUDING COMMENTS

Enormous advances have been made in the management and stewardship of tailings storage facili-ties over the past two decades or so. The days when a TSF was operated on a part-time basis by the most junior metallurgical cadet on site are (hopefully) behind us. Many operations now have full-time, on-site geotechnical engineers managing these facilities, providing much needed guid-ance and development of continual improvement processes.

However, there is no justification for complacency. Failures of TSFs still happen with monoto-nous regularity, with the centre of gravity of the worst affected parts of the industry constantly shifting around the world, as well as shifting from one type of resource to another. In the past two or three years there have been a number of failures in China that have resulted in fatalities, but who knows where the next catastrophic failure might occur? A number of aspects of tailings man-agement, and ways to improve this vital component of the mining life cycle have been discussed, albeit briefly, in this paper. Hopefully the adoption of some of these ideas will ensure that no catastrophic failures occur in the future, plus impacts to the receiving environment are minimized, particularly once the operation closes and those responsible for the concentration of a potential pollutant into a relatively small footprint are long gone.


13

REFERENCES

Benson, C., Sawangsuriya, A., Trzebiatowski, B. and Albright. W. (2007). Postconstruction changes in the hydraulic properties of water balance cover soils. ASCE Journal of Geotechnical and Geoenvironmental Engineering, 133(4), pp. 349–359.

Blight G.E. 1998. Destructive mudflows as a consequence of tailings dyke failures. Proceedings Institution of Civil Engineers. Vol.125, pp. 9–18.

Bryson, B. Latest projection of climate change & impacts on planning systems. http://www.planning.org.au/index.php?option=com_docman&task=docclick&Itemid=0&bid=1356&limitstart=0&limit=10. Accessed on 14th July 2008.

Campbell, G. (2004). Store/release covers in the Australian outback: a review. Proceedings of Mine closure seminar—towards sustainable outcomes. Australian Centre for Geomechanics, August 2004, 66 pp.

Chung, S.F. and Randolph, M.F. (2004). Penetration resistance in soft clay for different shaped penetro meters. Proc. ISC-2 on Geotechnical and Geophysical Site Characterisation. Ed: da Fonseca and Mayne, Millpress, Rotterdam, pp. 671–677.

Cooling, D.J. (2007). Improving the sustainability of residue management practices—Alcoa World Alumina Australia. Keynote address: Proc. Tenth International Seminar on Paste and Thickened Tailings, March 2007, Perth, Australia, pp. 3–15.

Coussot, P. (1997). Mudflow rheology and dynamics. A.A. Balkema, Rotterdam.Coussot, P., Laigle, D., Arattano, M, Deganutti, A. and Marchi, L. (1998). Direct determination of rheological

characteristics of debris flow. Journal of Hydrological Engineering, 124(8), pp. 865–868.DeJong, J.T., Yafrate, N.J., DeGroot, D.J. and Jakubowski, J. (2004). Evaluation of the undrained shear

strength profile in soft layered clay using full-flow probes. Proc. ISC-2 on Geotechnical and Geophysical Site Characterisation. Ed: da Fonseca and Mayne, Millpress, Rotterdam, pp. 679–686.

Energy Resources Conservation Board, Draft Directive, ‘Tailings Performance Criteria and Requirements for Oil Sands Mining Schemes’, March 2008.

Fourie, A.B. and Papageorgiou, G. (2001). Defining an appropriate steady state line for Merriespruit gold tailings. Canadian Geotechnical Journal. 38(4), pp. 695–706.

Fourie, A.B. and Tibbett, M. (2007). Post-mining landforms: engineering a biological system. Keynote Address. 2nd International Seminar on Mine Closure, October 2007, Santiago, Chile, pp. 3–12.

Hofmann, B., Sego, D.C. and Robertson, P.K. (2000). In situ ground freezing to obtain undisturbed samples of loose sand. ASCE Journal of geotechnical and geoenvironmental engineering, 126(11), pp. 979–989.

Hungr, O. (2000). Analysis of debris flow surges using the theory of uniformly progressive flow. Earth Sur-face Processes and Landforms, 25, pp. 483–495.

Hungr, O. (2006). Rock avalanche occurrence, process and modelling. Keynote paper in: NATO Advanced Workshop on Landslides from Massive Rock Slope Failure, Celano, Italy. Ed: S.G. Evans, G. Mugnozza, A. Strom and R. Hermanns. Springer, Heidelberg, pp. 243–266.

Iverson, R.M. (1997). The physics of debris flows. Reviews of Geophysics, 35(3), pp. 245–296.Jefferies, M. and Been, K. (2006). Soil liquefaction: a Critical State approach, Taylor and Francis.Lade, P.V. and Yamamuro, J.A. (1997). Effects of nonplastic fines on static liquefaction of sands. Canadian

Geotechnical Journal, 34, 918–928.McPhail, G.I. and Brent, C. (2007). Osborne high density discharge—an update from 2004. Proc. Tenth Inter-

national Seminar on Paste and Thickened Tailings, March 2007, Perth, Australia, pp. 339–350.Naef, D., Rickenmann, D., Rutschmann, P. and McArdell, B.W. (2006). Comparison of flow resistance rela-

tions for debris flow using a one-dimensional finite element model. Natural Hazards and Earth System Sciences, 6, pp. 155–165.

Nguyen, Q.D. and Boger, D.V. (1998). Application of rheology to solving tailings disposal problems. Interna-tional Journal of Mineral Processing, 54(3–4), pp. 217–233.

Robinsky, E.I. 1975. Thickened discharge—A new approach to tailings disposal, Bulletin, The Canadian Institute of Mining and Metallurgy, 1975, pp. 47–59.

Teh, C.I. and Houlsby, G.T. (1991). An analytical study of the cone penetration test in clay. Geotechnique 41, No. 1, 17–34.

Wagener, F. 1997. The Merriespruit slimes dam failure: overview and lessons learned. Journal of the South African Institution of Civil Engineering, 39(3): 11–15.

Williams, M.P.A., Seddon, K.D. and Fitton, T.G. (2008). Surface disposal of paste and thickened tailings—a brief history and current confronting issues. Keynote address: Proc. Eleventh International Seminar on Paste and Thickened Tailings, May 2008, Kasane, Botswana, pp. 143–164.



Design, operation, and disposal



17


Simple mine waste management planning for successful operations in the Canadian low arctic

P.M. BedellGolder Associates Ltd., Burnaby, British Columbia, Canada

K.S. WillisDe Beers Canada Inc., Snap Lake Mine, Northwest Territories, Canada

ABSTRACT: Operations started at De Beers Canada Inc.’s Snap Lake Mine, about 220 km northeast of Yellowknife, Northwest Territories, Canada, in late August 2007. The Snap Lake Mine is located in the Canadian Low Arctic in a region of continuous permafrost. A number of “firsts” are associated with the Snap Lake Mine:

• The mine is De Beers’ first mining operation outside of Africa;• The mine is Canada’s first completely underground diamond mine; and• The mine will be the first mine located in the low arctic, Canadian or otherwise, to deposit mine

waste on surface as a paste.

The surface disposal facility at the Snap Lake Mine is referred to as the North Pile. The North Pile will be sequentially developed in three phases: the Starter Cell, the East Cell and the West Cell. The design objective of the North Pile is that neither the operation of the mine nor the process plant will be impacted by activities at the North Pile. This requires that the North Pile must be capable to receive all waste materials produced at all times.

The paper presents a general discussion of the deposition and development management plan for the North Pile. Emphasis will be placed on the elements of the plan which result from consider-ation of the physical setting of the mine: the Canadian Low Arctic. It will be seen that many of the elements of the deposition and development plan seemingly contradict each other; the trade-offs and impacts of these will challenge the optimization of the deposition and development plan.

It is important to note that the quantities, tonnages, data and management concepts presented in the paper are accurate for the time of writing and are subject to change and modification as operations progress.

1 INTRODUCTION

De Beers Canada Inc. (De Beers) owns and operates the Snap Lake Mine located about 220 km northeast of Yellowknife, Northwest Territories, Canada, at a latitude of 63° 36’19”N and a longi-tude of 110° 52’00” as shown on Figure 1. The mine is located in the Canadian low arctic in the region of continuous permafrost. Operations at the Snap Lake Mine started in late August 2007.

The mine is an underground diamond operation following a kimberlite dyke that dips beneath Snap Lake. The nominal daily processing rate is 3,150 tonnes per day. The material remaining following the removal of the diamonds is referred to as processed kimberlite (PK). The operation will generate about 22.8 million tonnes of PK over the 22 year mine life.

The PK and mine waste rock form the mine waste. The mine waste stored on surface is disposed of into the facility referred to as the North Pile. Backfilling of the mine will start in about Year 3 of operations. Backfill will comprise cemented PK paste. About 50% of the PK generated during the life of the mine will be used as backfill; the remainder will be placed in the North Pile.


18

A number of “firsts” are associated with the Snap Lake Mine:

• The mine is De Beers’ first mining operation outside of Africa;• The mine is Canada’s first completely underground diamond mine; and• The mine will be the first located in the low arctic, Canadian or otherwise, to deposit mine

waste on surface as a paste.

The paper presents a general discussion of the deposition and development plan for the North Pile. Emphasis will be placed on the elements of the plan which result from consideration of the physical setting of the mine: the Canadian Low Arctic. It will be seen that many of the elements of the deposition and development plan seemingly contradict each other; the trade-offs and impacts of these will challenge the optimization of the deposition and development plan.

It is important to note that the quantities, tonnages, data and management concepts presented in the paper are accurate for the time of writing and are subject to change and modification as operations progress.

2 SITE CONDITIONS

2.1 General

The layout of the Snap Lake Mine is shown on Figure 2. The Plant Site is located on the northwest peninsula on the western shore of Snap Lake. The site is an isolated area with no permanent surface access. Access to the site is by a winter road spur or by aircraft.

The topography of the site is gently sloping with occasional knolls referred to as “tundra.” The surface elevation varies between about 445 m to 484 m. The site is generally barren of vegetation

Figure 1. Site location.


19

with the exception of some isolated small trees and dwarf shrubs. Bedrock outcrops are common with a veneer of overlying till of varying thickness. Deposits of organic material are present in most of the low-lying areas. Surface drainage does not follow a defined pattern as there are no major water courses on the site.

The underground mine dips beneath Snap Lake from the Plant Site to the east at about 15 degrees from the horizontal to follow the kimberlite dyke. The North Pile facility is located about 500 m west of the Plant Site.

2.2 Climate

The site is located in the Canadian low arctic. Long, cold winters with short, cool summers are experienced. Typical maximum and minimum average monthly temperatures are −30 °C and 15 °C, respectively. Annual rainfall and snowfall totals are about 148 mm and 225 mm, respectively. Wind is a common occurrence at the site; speeds in excess of 30 km/h are commonly experienced.

2.3 Permafrost

The site is located just north of the diffuse boundary between the discontinuous and continuous permafrost zones. Based on the results of monitoring performed on-site, the average thickness of the active layer is about 6 m. The permafrost thickness is expected to be at least 100 m based on the available literature; no site-specific investigation to determine the permafrost thickness has been performed to date.

3 MINE WASTE MANAGEMENT STRATEGY

3.1 General

The surface disposal facility at the Snap Lake Mine is referred to as the North Pile and is shown on Figure 3. Located to the west of the Plant Site (refer to Figure 2), the North Pile is planned

Figure 2. Project site.


20

to be about 90 hectares in area with a maximum thickness of about 40 m. The North Pile will be sequentially developed in three phases: the Starter Cell, the East Cell and the West Cell. The development strategy of the North Pile includes the use of progressive closure of the facility; as each cell is developed and filled, it will be covered with non-reactive rock. The closure surface for each cell ties into that for the entire North Pile. Progressive closure allows for the monitoring of closure conditions during operations for assessment and application of learning.

The design objective of the North Pile is that neither the operation of the mine nor the process plant will be impacted by activities at the North Pile. This requires that the North Pile must be capable to receive all waste materials produced at all times.

The North Pile is not designed as a water retaining facility. The facility is operated to promote drainage of water through the perimeter embankments for routing and collection by the perimeter water control structures. Water collected in the perimeter sumps is transferred to the water manage-ment pond on an on-going basis. Should water pond inside the North Pile, it will be transferred to the water management pond. A permanent pond is not maintained within the North Pile facility.

The North Pile will be developed using the PK materials and mine waste rock. The deposition and development plan for the North Pile have been developed with the recognition that the PK mate-rials will be variable in properties during operations; especially during start-up. The plan is flexible due to the realization that the PK materials to be pumped to the North Pile will be paste and slurry. A primary goal of the deposition plan is to limit ice entrapment to increase the storage capacity for PK and mine waste rock. An important objective of the development of the Starter Cell will be to develop a methodology to construct stable embankments throughout the year using paste. The plan includes monitoring activities to enable the optimization of the waste management operations.

3.2 Mine waste materials

3.2.1 Processed kimberliteThe process plant produces PK materials in three size fractions: coarse, grits and fines. The approximate average proportions of the three fractions and the solids content (by mass) at which

Figure 3. General arrangement of the North Pile.


21

they be produced will depend on the actual process. The design proportions and solids contents of the three fractions are presented in Table 1; some day-to-day variations may occur.

To date, the fines fraction is deposited as slurry. The coarse and grits are dewatered to a solids content of about 84% (by mass) at the process plant and are hauled and placed in the surface disposal facility using conventional earth moving equipment. Following start-up activities, the process plant will produce paste from the PK materials, comprising fines, coarse and grits frac-tions, for disposal on surface.

Full mix paste, comprising coarse, grits and fines fractions, will be pumped to the North Pile. It is expected that the process plant will be able to produce a “good quality” paste during the devel-opment of the Starter Cell. A “good quality” paste, from a depositional point-of-view, is defined as a material which is non-segregating, exhibits little bleed and has the ability to remain static in a pipe for a period of time followed by the restarting of flow.

Placement of underground backfill is expected to commence in September 2010 using an aver-age of about 50% of the PK paste quantity produced. It is anticipated that this will involve under-ground backfilling operations requiring 100% of the paste for periods of time and 0% for others.

3.2.2 Mine waste rockMine waste rock will be placed within the North Pile. Depending upon the schedule of material, the mine waste rock may be used to construct the embankments of the North Pile. Rock type and geochemistry (i.e., non-acid generating or potentially acid generating) will dictate where within the North Pile the mine waste rock will be placed.

3.3 Perimeter water control structures

The North Pile is not designed as a water retaining facility. The facility is operated to promote drainage of water through the perimeter embankments for routing and collection by the perimeter water control structures. Water collected in the perimeter sumps is transferred to the water management pond on an on-going basis.

Each cell of the North Pile will be surrounded by perimeter water control structures comprising ditches and sumps. To date, the only perimeter water control structures constructed are the temporary sumps (TS1, TS2, TS3 and TS4), perimeter sumps (SP1 and SP2) and ditches of the Starter Cell, as shown on Figure 3. Water collected in the sumps is pumped to the water management pond prior to treatment and release to the environment.

3.4 Perimeter embankments

An important objective during the development of the Starter Cell will be to develop a meth-odology to construct stable embankments using paste, especially throughout the year. Perime-ter embankments will be constructed for each cell. At present, only the Starter Cell perimeter embankments have been constructed as shown on Figure 3. A typical cross-section of the North Pile perimeter embankment is shown on Figure 4. The initial embankments will be constructed on a prepared foundation surface to provide stable conditions in frozen and thawed conditions. The embankments will typically be raised upstream over the deposited materials; embankments will

Table 1. Processed kimberlite size distribution.

Processed kimberlite Particle size Proportion of processed Solids contentfraction (millimetre) kimberlite (by mass)

Coarse 1.5 to 6.0 45% 90%Grits 0.125 to 1.5 35% 78%Fines Less than 0.125 20% 47% to 55% Total (Full Mix) 100% 72% to 76%


22

be raised downstream in the areas of deposited slurry to facilitate construction. Mine waste rock and PK materials will be used to construct the embankments.

The downstream slope of the initial embankment of the Starter Cell was constructed using mine waste rock at a slope of 2 horizontal to 1 vertical, as shown on Figure 4. Embankments constructed using PK materials will have downstream slopes of 3 horizontal to 1 vertical. For ease of embank-ment construction, all downstream slopes of the East Cell and West Cell embankments will be constructed at 3 horizontal to 1 vertical.

3.5 Deposition considerations

3.5.1 ObjectivesThe objectives of the PK deposition plan are to:

• Conform to the design objective of the North Pile: the waste management system can never interrupt or cause shutdown of the mine or process plant operations. The surface facilities must be, therefore, capable of receiving all materials that potentially may be produced by the mining operations.

• Limit ice entrapment to increase the PK and mine waste rock storage capacity of the North Pile. During the deposition planning work, a 10% allowance (by volume) for ice entrapment in deposited PK materials has been made during the winter season.

• Provide a flexible plan with the realization that PK materials will be variable in properties dur-ing the operations. Further, the actual processing schedule and the date which underground backfilling commences will influence development and the deposition sequencing. The opera-tional experienced gained during the Starter Cell development and deposition will be used to optimize those of the East Cell and West Cell.

• Provide a flexible plan with the realization that the PK materials pumped to the North Pile will be paste and slurry. It is expected that the majority of the material deposited into the North Pile will be paste; however, the deposition plan includes provisions for the deposition of slurry.

• Use PK materials and mine waste rock to develop containment and deposition structures for the North Pile. An important objective during the development of the Starter Cell will be to develop a methodology to construct stable embankments using paste, especially throughout the year.

3.5.2 Processed kimberlite piping distribution system.Slurry and paste distribution lines are installed on the crests of perimeter embankments. Slurry is deposited from a 150 mm diameter steel pipeline. Paste will be deposited from 200 mm diameter steel pipelines. End discharge will used for both slurry and paste materials. High density

Figure 4. Typical cross-section of North Pile perimeter embankment.


23

polyethylene (HDPE) extensions will be required at some paste discharge locations for use during deposition. The extensions will be elevated on trestles and/or deposition berms to the midpoints of each cell.

3.5.3 Operational inputsThe following constraints were used as inputs into the deposition planning work:

• During winter months, deposition would use a “pull back” methodology (i.e., use deposition points as far from the Valve House Building as possible and retreat along the pipeline to the Valve House Building).

• Limit switching between PK deposition lines so to reduce line flushing requirements.• Changing between deposition points on a given line will, in general, consist of stopping PK flow

in the line, flushing the line with water for cleaning, depositing slurry from the slurry line or the emergency slurry discharge point and producing the coarse and grits fractions at the process plant while the line is flushed and the deposition point is relocated followed by the reinstatement of PK flow to the deposition line.

• The emergency discharge point is not to be used as a planned PK deposition point.• All pipelines should be flushed of PK materials and fully dewatered prior to periods of non-use

to reduce the likelihood of materials and/or water freezing within the pipelines.

3.5.4 Slurry managementIt is expected that the majority of the material deposited into the North Pile will be paste; however, the deposition plan includes provisions for the deposition of slurry. Slurry will consist of the fines frac-tion of the PK at the solids content presented in Table 1. To promote drainage of the deposited slurry, it will be contained in the topographic low areas of each cell of the North Pile. Drainage of the slurry will be trained to flow through the embankment and report to the perimeter water control structures. Provisions for pumping of water should it pond within the North Pile will be made by operations.

During slurry deposition, the coarse and grits fractions will be hauled to the North Pile.

3.5.5 Paste managementThe paste to be deposited in the North Pile will consist of the full mix of PK fractions at the solids content presented in Table 1. The general paste management considerations include:

• An important objective during the development of the Starter Cell will be to develop a methodology to construct stable embankments throughout the year using paste. It is recognized that the development of this methodology will be required for the development of the remainder of the North Pile.

• Steep deposition slope angles are expected for the paste. Elevated pipe extensions from embankment crests will be required. Extensions will be elevated to at least the elevation of the corresponding embankment crest through the use of trestles, berms and/or geosynthetic-contained material. The use of conventional earthmoving equipment for movement of deposited materials will also be required.

3.5.6 Water management and surface drainageThe North Pile is designed to promote drainage of deposited materials to the perimeter drainage control structures to limit ice entrapment. Prior to deposition, all ponded water will be pumped out of low-lying areas. During operations, any ponded water in low-lying areas or on deposited materials will be pumped out of the North Pile. Grading of the surface of the deposited materials to promote drainage and/or limit water ponding will be required.

Geotechnical field monitoring and investigations, including excavations to determine ice content and depth of freezing and thawing will be performed. Thermistor strings will be installed to monitor temperature changes during deposition.


24

4 COMMENTARY

The development of the North Pile challenged by the climatic conditions similar to other mine sites in the Canadian low arctic. Operations will be challenged due to the initial variability of the PK materials produced by the process plant, material management and handling of water.

The extreme winter conditions, primarily low temperatures and high winds, impact the ability of both equipment and crews to function efficiently. Infrastructure care, assembly and operation, such as the PK distribution pipelines and spigots, will be challenging during the winter.

From the above discussion, the some of the trade-offs to be considered during the development of the North Pile include:

• Pumping PK at lower solids contents to lengthen deposition time from a given point versus installation of pipe extensions and/or mechanical movement of deposited PK. Less water pumped into the North Pile results in a lower possibility of ice entrapment and less water management activities. Deposited PK materials at higher solids contents are expected to have steeper deposition slope angles and will require additional infrastructure, machinery and/or crew support.

• Possible loss of available PK storage capacity to ice accumulation versus water management. Pumping less water into the North Pile with the PK materials puts more burdens on the deposition activities. Deposition of PK materials at lower solids contents puts additional burdens on water management activities within the North Pile itself and/or the perimeter water control structures.

• Material management for deposition versus embankment development. As the PK material is used to construct the containment embankments for the North Pile, it is necessary to separate appropriate streams and quantities to prepare for deposition requirements.

• North Pile embankment construction versus underground backfilling operations. Deposition activities require that embankments be constructed in advance. The scheduling of embankment construction may conflict with those for underground backfilling.

The successful operation of the North Pile will be optimized through a monitoring program of many of the items discussed above. The monitoring program will permit De Beers to determine the appropriate drivers of the trade-off items above and allow for informed and proactive decisions to be made. The waste management plan was developed to enable flexibility and revisions to apply the experience gained during operations.


25


James Creek Tailings Facility relocation for the gold quarry open pit expansion

A. Boye & J. YoungNewmont Mining Corporation, Carlin, NV, USA

ABSTRACT: The Gold Quarry open pit mine design includes a layback that encompasses a portion of the James Creek Tailings Storage Facility (JC-TSF). In preparation for the layback the removal of approximately 6.71 million tonnes of tailings and local alluvium was required. The JC-TSF was constructed in 1985 as a clay lined basin with an earth and rock filled main dam and two secondary embankments. There have been no deposits within the facility since the early 1990s. Mining of the JC-TSF was completed in two phases. The first phase involved cutting a trench to maintain confinement of tails and to eliminate potential interruptions to mining in the pit below caused by slope instability issues. The second phase was to mine the pit layback. The exposed tailings material was then sloped and covered. The mining conditions were challenging, however, with proper planning and operation, the relocation of the tailings was a success.

1 INTRODUCTION

1.1 Location

Newmont Mining Corporation’s Carlin Trend Operations consists of two geographically distinct areas—the South Area, where the Gold Quarry pit is located, is 11-km north of the town of Carlin, Nevada; the North Area, where the Carlin pit is located, is approximately 40-km northwest of Carlin, Nevada, as shown in Figure 1.

1.2 History

Exploration along the Carlin Trend began in the 1870’s and intermittent mining, consisting of small scale gold, copper, lead, barite and turquoise mines, continued for the next 90 years. During the 1960’s Newmont Mining Corporation, as the Carlin Mining Company, began exploration in the area. In the summer of 1963 a deposit consisting of 3 million ounces was discovered. New-mont’s first oxide mill began production in 1965 and exploration along the Carlin Trend yielded enough gold to allow production to continue to the present day.

Exploration around the Gold Quarry pit began in the 1960’s, production began in 1981 and continues to this day. Gold Quarry’s current dimensions are approximately 2.4-km southeast to northwest, 1.6-km east to west, and 470 m in depth. Approximately 1.5 billion tonnes of materials have been removed from the Gold Quarry pit with 0.5 billion tonnes of materials remaining in the current life of mine plan.

The James Creek Tailings Storage Facility (JC-TSF) was constructed in 1985 and contains approximately 46 million tonnes of material. The JC-TSF was constructed as a clay lined basin with an earth and rock filled main dam, and two secondary embankments. The material directly underlying the JC-TSF is Tertiary Carlin Formation which consists of interbedded tuffaceous clays and silts. The Carlin Formation is between 60 m and 180 m thick in the area directly underlying the JC-TSF. The JC-TSF has not been used since the early 1990’s except when the Mill 5-6 tailings impoundment (MTI) was down for emergency maintenance.


26

2 DESIGN WORK

2.1 Golder associates

Continued exploration around Gold Quarry resulted in a potential layback to the Gold Quarry pit. The new lay back, designed in 1994 and 1995 at AU$400, would only be feasible if a portion of the JC-TSF could be removed, as shown in Figure 2. In 1995 Newmont Mining Corporation had

Figure 1. Regional map of carlin trend open pit operations.


27

Golder Associates evaluate the feasibility of the pit expansion. Golder Associates proposed the re-mining of approximately 27 million tonnes of the total 46 million tonnes located in the JC-TSF. Based on the material properties that were determined during a study conducted in 1995, Golder Associates suggested a dredging operation to remove the tailings material. It was believed conven-tional mining equipment, including low ground penetrating equipment, would not be able to float due to the wet condition of the tailings.

The dredging operation would require very close monitoring of the pond used to create the slurry at a 25% solids concentration. The tailings material would be relocated to the MTI. It was Golder Associates estimate that it would take approximately 2.5 years to complete the dredging operation. At the end of the operation, approximately 73,000 cubic meters of water would have to be evacuated to the MTI and re-circulated back into the reclaim water system.

Golder Associates also proposed the final slope of the remaining tailings material be no greater than 7:1 for geotechnical reasons. The 7:1 slope would protect against any seismic activity due to blasting or short term shock. The report also required the construction of a safety embank-ment to protect against liquefaction of the remaining tailings material caused by larger seismic shock activities. The slope parameters of the project were derived from several tests performed by Golder Associates, including Cone Penetrometer Testing (CPT), borehole drilling, surface sam-pling, laboratory testing, and vane shear testing.

Shortly after Golder Associates completed their study the Gold Quarry layback was shelved due to an unfavorable gold price.

2.2 Smith Williams consultants, inc.

In 2004 the rising gold price made the Gold Quarry layback economically feasible, once again. In 2005 Newmont commissioned Smith Williams Consultants, Inc. (SWC) to re-evaluate the relo-cation of JC-TSF material. With approximately a 10 year difference between Golder Associates report and SWC report the material properties of the tailings had changed. With no deposit of tailings on the impoundment since the early 1990’s the moisture content of the tailings had been reduced significantly causing the tailings to consolidate. SWC proposed conventional mining equipment be used to excavate the tailings in two phases.

The first phase would mine the tailings material and Carlin Formation in addition to develop-ing a barrier of Carlin material between the tails mining and the pit. This barrier was required by representatives of state regulatory agencies to contain the tailings material until slope monitoring could be established and prove the tails slope would be stable. The first phase would complete 90% of tails mining associated with the project.

After the removal of the tails in the first phase the remaining tailings would be sloped to a final slope of 7:1. The slope would then be covered with a 46 cm sand and gravel drainage blanket and a 91 cm thick growth media cover (137 cm total cover) in order to prevent erosion and surface

Figure 2. Cross section of JC-TSF with existing pit edge and proposed new pit edge (SWC 2006).


28

water infiltration. Once this was completed representatives of the state regulatory agencies were satisfied that the slope was stable the second phase of mining could begin.

The second phase was to remove the Carlin Formation that formed the barrier between the tailings and the pit. Once the barrier was removed an 18.2 m catch bench would be established between the toe of the tailings material and the crest of the pit edge.

The complete excavation required the relocation of approximately 6.71 million tonnes of tailing and Carlin material. The material was relocated directly on top of the existing tailings facility and offset 152 m from the final crest of the tailings slope, as shown in Figure 3. The relocated material would be placed in two 7.62 m lifts starting on the southwestern part of the JC-TSF and continuing to the northeastern section of the JC-TSF tailings dump.

SWC geotechnical investigation included the data from Golder Associates along with addi-tional (CPT) testing on the more consolidated tailings material shown in Figure 4. The SWC geo-technical investigation also indicated a final slope of 7:1 to ensure slope stability. SWC required that piezometers, inclinometers, and prisms would be installed at 152 m increments to monitor the slope stability throughout excavation and the life of the JC-TSF.

3 EXECUTION

3.1 Contractor mining

Newmont Mining Corporation contracted American Asphalt and Grading Mining Division (AA) to begin the relocation of the JC-TSF material in September of 2006 for what was originally thought to be a 14 month project. AA was contracted for the project due to their flexibility and smaller mining equipment. The smaller mining equipment had a better chance of floating on the tailings

Figure 3. Designed mining and dump for relocation of JC-TSF (SWC 2006).


29

Figure 4. Geotechnical drilling locations throughout the JC-TSF (SWC 2006).

Figure 5. AA’s Hitachi 1800 and 777 Haul truck working in the JC-TSF face (SWC 2006).


30

material then the larger equipment Newmont was currently running. The AA fleet consisted of a Hitachi 1800 shovel and a CAT 992 loader, and approximately eight CAT 777 haul trucks, see Figure 5. The support equipment included a 5.5 m grader, a water truck, and four or five CAT D6 wide track dozers. AA worked straight 10 hour days 5 days a week throughout the project.

Newmont and AA decided the safest way to mine the cut was in two benches due to the total depth of the cut (25 m) and concern about the stability of the tails face. The first cut was com-pleted with a loader using a bench height of approximately 10 m, the bench height of the second cut varied from 7 m to 15 m depending on the depth of the tailings. The first cut was accessed from an existing haul road leading from the pit to a dump.

The first cut in the trench was the most challenging as all mining occurred on tailings. Waste material from the Gold Quarry pit was used to plate the bench with 1.5–2.5 m of rocky material as shown in Figure 6. This allowed AA to keep equipment afloat. As the first cut advanced to the north, dozers pushed the tailings face to a 7:1 slope.

Mining progressed faster than the 7:1 slope could be pushed in. The solution to this problem was the split bench elevations again, with smaller mining benches the dozer power required to put in the 7:1 slope was nearly cut in half. More of the material was able to be mined with the loading units reducing the amount of material needing to be dozed and the overall push distance.

The first cut was pushed from the south to the north until a secondary road could be constructed to tie into the tails dump and the main haul roads. The second access road reduced the haul cycle significantly which improved overall efficiency. This allowed the two cuts to proceed simultane-ously. The floor of the second cut was on Carlin Formation, which was more stable then the tail-ings and required less plating.

Figure 6. Plating material required to keep equipment a float (SWC 2006).


31

As the second cut advanced to the north, the 7:1 slope was completed. With the more stabile floor conditions the main concerns for the second cut were stabilization of the trench wall to the pit side and controlling surface water.

Survey control for the tails mining was intense and required the survey crew stake the floor grade as well put out cut and slope stakes on a daily basis. Daily progress was surveyed along both the cut and the tailings slope to ensure the mining progressed as designed.

3.2 Dump construction

The SWC dump designs were created for 6.71 million tonnes of material, which allowed for no additional capacity for over-mining or the road plating that was necessary to keep equipment floating. During the first month of operations AA used approximately four tonnes of road plating for every ten tonnes of tails mined. This meant that the amount of material going to the tails dump was increased by 40–50% which was not accounted for in the original design. Due to the oversight and the extra material added to the dump the available space was insufficient. To address dump space availability additional dump space was designed into the southern dump. Much effort was also made to reduce the amount of road rock necessary and through diligent efforts the plating requirements were reduced to 20–30% of the total tons mined. This was found to be a minimum requirement to allow for safe and efficient operation of both the dig and dump surfaces.

Managing the tailings dump face proved to be one of the greatest challenges in the tails mining project, even with the smaller trucks. The SWC nominal dump height of 7.62 m was established, however, maintaining the dump height was difficult because low material strength resulted in slope stability issues. Toe heave combined with wet tailings adhering to the trucks beds created the potential for trucks to overturn. The lift height was reduced to 4.5 m in order to mitigate the slope stability and truck overturning issues, however the toe continued to heave and the road plat-ing requirements did not change for the smaller lift heights resulting in the overall road plating requirements for the dump nearly doubling in volume. This was unacceptable due to the limited dump capacity.

The 7.62 m dump face was reestablished and procedural controls were implemented to miti-gate the safety concerns. The first of these procedures was to offset the location the trucks would dump away from the crest of the dump. The trucks would dump between three and six meters from the crest of the dump depending on the conditions of the dump slope and the tailings they were hauling. A dozer would then push the material out over the dump face which increased the dozer power required but reduced the opportunity for a haul truck to over turn on the crest of the dump. The second procedure was to conduct visual inspection of the truck bed every cycle to ensure that there was not excessive build up of tails in the truck beds. During rain or when the tailings were unusually wet dump rates were also decreased.

Sections of the dump face were buttressed prior to dumping to reduce heaving of the toe. Buttressing was used on the perimeter of the dump and in sensitive areas, specifically to protect dewatering wells and monitoring points as well as light vehicle access roads. Sloughing inside the dump perimeter continued to occur frequently. Sloping the dump face to a slope of 3:1 when the material was relatively dry and a 6:1 when the material was wet reduced the amount of intermedi-ate sloughing.

3.3 American asphalt schedule and rates

The mining tonnage rates during the first cut were planned at 6,500 tpd, AA achieved rates of 12,000 tpd. Once the second cut began the planned tonnage rates were increased to 15,000 tpd, AA achieved an average rate of 28,000 tpd. By the end of April 2007 the project was six months ahead of schedule. AA had moved 6.8 million tonnes, 5.7 million of which were inside the design for average cost of $1.41/ton. At the achieved rates it was estimated the project would be finished in July of 2007. However due to budget overruns in March of 2007 Newmont decided to put the project on hold, starting at the end of April, until January 2008.


32

3.4 Newmont mining

In the early morning on July 19, 2007 prism monitoring indicated a section of highwall directly below the unfinished portion of the JC-TSF cut would fail. The area was evacuated and bermed off. Latter that afternoon a section of highwall measuring approximately 60 m tall by 80 m long consisting of Carlin material failed onto one of the main pit ramps. Prism monitoring indicated that movement in the surrounding area was slow but ongoing making it unsafe to work below. It was determined that a de-weighting cut would be necessary in order to reestablish access to the pit and stabilize the east highwall. The de-weighting cut would require completion of the tails mining cut. JC-TSF mining began, with Newmont’s equipment in August of 2007. Equipment consisted of a Hitachi 5500 and Cat 793D trucks.

Due to concerns over the larger equipment’s ability to float on tails, the decision was made to use only the lower cut to ensure that the shovel was always operating on the Carlin formation. For safety, dozers would push the 24 m face down to the shovel. This strategy proved to be dozer intensive and required two to three dozers attending to the shovel at all times. When it became apparent that mining the tails in a single cut would require more dozers than the site could provide, it was decided to break the project into two benches with dig faces of approximately 12 m each. The 12 m face could be better managed by the shovel with dozer support to keep the floor in good condition. The upper cut required between 1.5 m and 2.0 m of road plating to keep the shovel and trucks floating. Grade control in the floor was difficult to maintain due to the amount of plating required. Constant attention from the shovel operator and high precision GPS were used to help the shovel maintain floor grade.

The lower cut began as soon as the upper cut was completed. A trench was developed in the lower cut to maintain confinement of the tails, while the upper cut was day lighted into the pit. The lower cut was able to proceed with the shovel running almost exclusively on the Carlin formation. This greatly reduced the amount of road rock required to keep equipment running. Mining on the lower level progressed quickly with sloping of the 7:1 slope falling behind the dig face.

Dump construction was also problematic for Newmont’s equipment. Newmont’s larger trucks required more plating on the dumps and had a greater chance of overturning at the dump crest. Newmont implemented all of the procedures AA had used for dump management. Trucks would dump 3–6 m from the crest of the dump with dozers pushing the material down the face to prevent the toe of the dump from heaving. Two dozers were required on the dump to keep up with the production coming from the shovel.

The expectation was for the shovel to achieve an average rate of 27,000 tpd, actual tonnage rates were 50,000 tpd. While the shovel preformed better then expected the project also required more dozer support then expected. The mining portion of the project was expected to last until mid-November but finished nearly six weeks ahead of schedule.

Figure 7. Overview of 7:1 slope on JC-TSF with the new pit wall set.


33

3.5 Final sloping and reclamation

The remaining tailings were sloped to a 7:1 final slope angle. At the toe of the tails slope a ditch was constructed and perforated pipe placed to collect any water running off the tailings slope. All water from the tailings slope is collected and pumped to the MTI. The tailings slope was then cov-ered with material sourced from the Carlin Formation inside the Gold Quarry Pit. The material to be used for the sand and gravel drainage blanket had specific requirements for porosity and com-paction, the material that was used was tested and approved by SWC. Appropriate material was found and the 46 cm drainage blanket was placed. Once the drainage blanket was placed a 91 cm growth media cover was placed in order to prevent erosion and surface water from infiltrating the tailings material as shown in Figure 7.

4 CONCLUSION

Overall the relocation of the JC-TSF has been a success. None of the problems encountered were insurmountable. It was found that mining of the tailings required more dozer power then expected for dig face management, final sloping and dump supervision. This was problematic given the finite nature of the equipment available and in some cases was the limiting factor in the projects progression. Overall the mining rates achieved far exceeded the expected rates. Even with delays the project was completed well ahead of schedule. This, in part, was due to the amount of road plating used to keep the shovels, loaders and trucks floating, however, it was also found that road plating would need to be used judiciously as the tailings dumps were under designed at the start of the project. Managing the amount of road plating used was a major concern as the available dump area was limited, even after the dump was redesigned. Management of the road plating, specifically limiting its use allowed all materials to fit within the redesigned capacity. Monitoring of the 7:1 slope has not indicated any movement and representatives of the state regulatory agen-cies have been convinced of the 7:1 slopes stability, allowing mining of the Gold Quarry layback to proceed.

REFERENCES

Golder Associates Inc. 1996. Preliminary Report Design of Tailings Remining and Relocation Operation for the Proposed $400. Prepared for: Newmont Gold Company. Prepared by: Golder Associates Inc. Lakewood, CO.

Smith Williams Consultants, Inc. 2006. James Creek TSF Tailings Relocation for the Gold Quarry Pit Expan-sion Final Design Report. Prepared for: Newmont Mining Corporation. Prepared by: Smith Williams Consultants, Inc. Englewood, CO.



35


Heat and mass balance modeling of a subaqueous tailings disposal facility

G.R. Eykholt & J.B. ManchesterFoth Infrastructure & Environment, LLC, Madison, WI, USA

S.V. DonohueFoth Infrastructure & Environment, LLC, Green Bay, WI, USA

J.C. CherryKennecott Eagle Minerals Corporation, Marquette, MI, USA

ABSTRACT: A subaqueous Tailings Disposal Facility (TDF), located in the Michigan Upper Peninsula, is a former iron mine excavation adjacent to a milling facility. Plans are being con-sidered to rehabilitate the mill for nickel and copper ore processing. Pyritic tailings are proposed to be placed as a slurry and managed in the subaqueous TDF. Tailings from a nearby mine were placed in the TDF in the 1980’s. Using the recovery data since the 1980’s and current chemical and limnological conditions, a multi-compartment heat and mass balance model was constructed to evaluate potential water quality impacts from additional loadings. A companion paper addresses current chemical and limnological conditions. The model was developed to keep track of tailings loadings, TDF volume, water chemistry, oxidation/reduction reactions, oxygen transfer dynamics, and pH-alkalinity relationships. Modeling outcomes generally show long-term stability of the TDF and manageable water quality impacts.

1 INTRODUCTION

1.1 Background

A subaqueous Tailings Disposal Facility (TDF), located in the Michigan Upper Peninsula, is a former iron mine excavation adjacent to a milling facility. The surface area of the TDF is 67 acres, with a maximum depth, length, and width of approximately 192 feet, 3500 feet, and 900 feet, respectively. Plans are being considered to rehabilitate the mill facilities for processing of nickel and copper ore. Processed, tailings are proposed to be placed as a dense slurry and managed in the subaqueous TDF.

Tailings from the nearby mine were placed in the TDF from 1985 until 1989. Monitoring data and the dynamics of recovery of the water quality since the former operation serve as a major data source for projecting the TDF behavior upon additional loadings.

1.2 Subaqueous tailings management and modeling

Tailings disposal in former mining pit lakes falls within a general category of subaqueous dis-posal. Coverage of several inches of water has been shown to greatly reduce the rate of oxida-tion of minerals, especially sulfide-bearing tailings and waste rock (Moses and Herman, 1989; Robertson, 1991; Morin, 1993; Pedersen et al. 1993; Fraser and Robertson, 1994; Peacey et al. 2002). Oxygen solubility and solubility of mineral oxides is limited in water, and processes leading to a passivation crust layer have been shown to greatly reduce oxidation rate and subsequent acid rock drainage (ARD). A recent literature review of water covers is provided by Peinerud (2003).


36

Also, the Mine Environment Neutral Drainage (MEND) program of Natural Resources Canada directed research on subaqueous disposal of sulfide tailings between 1988 and 2000 and has pub-lished many reports and reviews describing this work (i.e., MEND, 1992; and MEND, 1996).

Modeling of deep, subaqueous tailings disposal facilities is challenging. Many physical and chemical processes may be involved (Castendyk and Webster-Brown, 2007ab). These include mass transfer of oxygen from the surface, the internal mixing dynamics of deep and stratified water bodies, alkalinity changes from oxidation of minerals, and other geochemical interactions. Limnological modeling typically requires consideration of the temperature-depth profile and water densities, which depend on water temperature and dissolved solids. Although geological and other evidence has indicated that groundwater interactions with the TDF are limited to shal-low depths, the dynamics of other deep subaqueous tailings facilities may be significantly affected by groundwater interactions.

1.3 Purpose and scope

The purpose of this paper is to summarize the development, calibration, and simulations from a multi-compartment heat and mass balance model of the TDF. A companion paper addresses geochemical considerations and limnological conditions. The model was developed to keep track of tailings loadings, TDF volume, water chemistry associated with loadings, oxidation/reduction reactions, oxygen transfer dynamics, and pH-alkalinity relationships. Modeling of mixing and thermal stability of the TDF was coupled with the heat and chemical mass balance modeling, and model predictions were made for periods during and after tailings loadings to the TDF.

Many of the modeling steps taken to simulate the TDF are generalized here, so the scope of this paper is limited to the TDF and general applicability is not claimed. It is also noted that, at the time of this writing, the TDF modeling and tailings loading concept at the site is still under development and not yet fully formalized.

1.4 Description of TDF bathymetry and watershed basin characteristics

Bathymetric characteristics of the TDF, gathered from acoustic survey, are reported in Table 1. A relatively shallow and isolated area of the TDF is referred to as the Northern TDF Area, and the

Table 1. Bathymetric characteristics for the TDF.

Full TDF area Main TDF area

Water depth ElevationBed area(m2)

Lowervolume(m3)

Bed area(m2)

Lowervolume(m3)(feet) (m) (feet) (m)

0828486888108128148168188192

02.448.5314.6320.7326.8232.9239.0145.1151.2157.3058.52

153815301510149014701450143014101390137013501346

468.78466.34460.25454.15448.06441.96435.86429.77423.67417.58411.48410.26

269400257504220408184789159005138674115561 97813 84843 62281 23047 0

79972167354816589468946812433632225272184919558921306477 750580 292602 14050 0

246831237529210189180144159005138674115561 97813 84843 62281 23047 0

77974347206804583431746647633632225272184919558921306477 750580 292602 14050 0


37

remaining area is referred to as the Main TDF Area. The tailings loadings and other operations are only expected in the Main TDF, but some of the natural watershed drainage is expected in the Northern TDF. The watershed drainage area is 158 acres, plus the 67 acre surface area of the TDF to yield a total area of 224 acres. The Northern TDF contains roughly 13% of the 158-acre drain-age area, and 8.4% of the surface area of the TDF. Further details regarding compartmentalization are discussed below.

2 METHODS

2.1 Modeled tailings loadings to TDF

The milling process is assumed to be continuous, with operation 24 hours per day, 7 days a week, 320 days a year, for approximately 7.1 years. The first calendar year of operation was set at 700 dry tonnes per day for 220 days beginning in mid-April. For calendar years 2 through 7, loading was set at 1159 dry tonnes per day for 320 days beginning in mid-January. The last calendar year of loading was set at 900 dry tonnes per day beginning in mid-January. A total of 2.502 million dry tonnes of tailings are expected to be placed.

The specific gravity of the tailings is estimated as 3.03, the dry density of the tailings slurry (60% solids) is expected to be 1001 kg/m3, and the dry density of the tailings formed in the bed (72% solids) is expected to be 1387 kg/m3. With these characteristics, the total proposed tailings loading is expected to reduce the volume of the TDF by approximately 1.8 million m3.

For the dry processing rate of tailings at 1159 tonnes/d and 60% solids, the expected volumetric flow for the tailings slurry is 1158 m3/d (212.5 gpm). Additional water discharges, from miscel-laneous sources, are expected to add 54.5 m3/d (10 gpm) to the flow, so the total mill discharge is expected to be 1213 m3/d (222.5 gpm). Flow rates for calendar years 1 and 8 are proportional with respect to the dry processing rates.

The tailings are expected to be loaded to the TDF as a slurry, forming a flat surface at depth. Recycle water will be reclaimed from mid-depth of the TDF and a control structure and discharge to a wastewater treatment plant (wwtp) will regulate the discharge flow from the TDF. Additional inputs from precipitation, snow melt, groundwater, and losses from evaporation also are consid-ered in the water balance. The general loading strategy is shown in Figure 1. Details for natural inputs and compartmentalization are discussed below.

Figure 1. TDF loading strategy and general orientation of model compartments.


38

2.2 Natural flows to TDF

A summary of monthly climate is presented in Table 2. Climate data are drawn from Champion—Van Riper State Park (NCDC site 201439), a weather station in the Michigan Upper Peninsula. Normal annual precipitation is 33.3 inches (846 mm), with extremes of 20.17 inches (low) and 46.94 inches (high). Annual evaporation is estimated as 14.8 inches. Monthly mean temperatures range from 11.3°F (−11.5°C) in January to 64.3°F (17.9°C) in July. Mean annual snowfall is 132.8 inches, with snow depth highest in February and March.

Area attributes of the watershed and TDF were used with the monthly net precipitation infor-mation to estimate flows for all time periods. Without flows from the mill operation, the average annual flow into the TDF is estimated as 246 gpm (1341 m3/d). Of this, 43 gpm is the expectation from a groundwater inflow model along the southern TDF boundary and 203 gpm is from precipi-tation minus evaporation over the watershed.

Maximum annual precipitation is roughly 1.4 times that of the normal annual precipitation of 33.3 in. (845 mm) and the maximum annual net precipitation (precipitation minus average evapo-ration) is 1.7 times that of normal. Although some simulations were conducted at high flows, all simulations presented here apply to average flow conditions. This is reasonable, given that the hydraulic residence time in the TDF ranges from 10 to 16 years.

2.3 Monthly flows with and without TDF loading

Monthly varying natural inflow follows from the monthly net precipitation and melt record , as combined with basin areas within the model. The flow record, shown in Figure 2, also demon-strates the flows from the mill discharge and the return (reclaim water) flow. The net precipita-tion to the main TDF is the same for every year, but the outlet flow (to WWTP) is held constant at 1816 m3/d (333 gpm) during the period of mill discharge and 1340 m3/d (246 gpm) after the tailings loading is completed. Water elevations during the loading period range from 468.5 m (1537.2 ft) to 469.1 m (1539.1 ft). This range is maintained for periods after loading.

2.4 Compartmentalization and changes during tailings loadings.

The volumes of compartments in the main TDF are shown in Figure 3. The compartmentalization of TDF was based on isolation of the northern TDF as a separate compartment, and division of the

Table 2. Net precipitation characteristics for the TDF and average monthly conditions.

Precipitation -

Precipitation Evaporation evaporation Percent Storage Net

Month (inches) (inches) (inches) storage (inches) (inches) (mm)

January 1.82 0.15 1.67 100% 4.09 0 0February 1.32 0.30 1.02 100% 5.12 0 0March 2.32 0.89 1.43 80% 5.24 1.31 33.3April 2.42 1.33 1.09 30% 1.90 4.43 112.6May 3.10 1.63 1.47 0% 0.00 3.37 85.7June 3.35 2.07 1.28 0% 0 1.28 32.5July 3.80 2.81 0.99 0% 0 0.99 25.2August 3.74 2.66 1.08 – 0 1.08 27.4September 3.88 1.77 2.11 0% 0 2.11 53.5October 3.29 0.74 2.55 0% 0 2.55 64.8November 2.44 0.30 2.14 35% 0.75 1.39 35.4December 1.82 0.15 1.67 100% 2.42 0 0Annual 33.3 14.78 18.52 18.52 470.3


39

Figure 2. Simulated inflows and regulated outflows for TDF simulations.

Figure 3. Trends in modeled compartment volumes from tailings loading.


40

main TDF into five layered, completely-mixed water compartments. The upper two compartments, defined primarily by an observed thermocline to a water depth of roughly 24 feet, vary in vol-ume due to changes in storage in the TDF. The lower compartments (3, 4, and 5) are beneath the observed thermocline and vary in volume due to the tailings loadings. The interface of compart-ments 3 and 4 represents an observed oxic-anoxic boundary and the lower compartment 5 repre-sents the near bed condition and the compartment in most intimate contact with the new tailings added to the TDF. The upper 0.5 m of the tails bed is considered an active model compartment, and diffusive mass and heat transfer between the active bed and compartment 5 is considered. Tailings loadings over a period of 7.1 years leads to a gradual reduction of lower compartment volume, but the lower compartments maintain the same proportion of the remaining, lower TDF volume.

Bathymetric relationships are used to correlate the interfaces of TDF compartment volumes with compartment interface water depths and compartment thicknesses. These trends are shown in Figure 4.

The mass and heat balance model for the TDF, including simulation of loading, bathymetric, and flow processes, was fully implemented using ModelMakerTM (Version 4.0, Cherwell Scientific, Oxford, UK). In addition to the processes discussed above, the model also addresses dynamic simulation of the following processes:

♦ Flow, mass, and heat balances with varying volumes.♦ Density-driven inter-compartment mixing and discharge plume distribution.

Figure 4. Trends in modeled compartment layer thicknesses from tailings loading and bathymetric relationships.


41

♦ Heat balances at surface (radiation, conduction, convection, ice formation, and other processes), discharge points, compartments, and tails bed.

♦ Modeling of multi-component and reactive mass balances, including interactions of dissolved oxygen, oxidized and reduced iron, dissolved solids, oxygen mass transfer at the water surface, alkalinity changes from oxidation and reduction reactions, and metals scavenging by complexa-tion and settling of oxidized iron.

The interactions between various model components (chemical constituents and variables) are summarized in Figure 5, and a diagram of chemical interactions of chemical constituents in a given compartment (compartment 3) is summarized in Figure 6.

2.5 Thermal aspects, density stratification, and mixing of TDF compartments

The tendency for thermal stability of a body of water may be expressed by the densimetric Froude number, N

DF (Tchobanoglous and Schroeder, 1987), a dimensionless ratio comparing inertial and

gravitational forces acting on the body of water:

N

Q bd

g dDF = /( )

( / )Δρ ρ (1)

where Q = flow rate; b = average width; d = average depth; Δρ = difference between top and bot-tom water density; ρ = depth-average water density; and g = gravitational constant. Values of N

DF

greater than one indicate that turbulent mixing prevents thermal stratification, while values less than one indicate that thermal stratification will remain despite small levels of mixing. The TDF

Figure 5. Diagram of TDF model interactions.


42

has a NDF

less than 0.00001 and is considered very strongly stratified. Although mixing in the upper two compartments is considered regular, due to winds and seasonal temperature swings, it is unlikely that the TDF will mix completely.

It is important to note that the model assumes complete mixing within a given compartment and a minimum amount of mixing flow between the compartments, providing some conservatism to the mixing model. The mass and heat balances affect the estimates for the dissolved solids concen-tration and temperatures of the compartments, and these variables are used to estimate the water density of the compartments. The level of density-dependent mixing flow between compartments (in the following case, between compartments 1 and 2) is set by proportioning of a maximum mix-ing flow (Q

mix, typically set to 15,000 m3/d),

Q MF Qmix12 12= ( ) (2)

where the expression for the mixing function (MF) is:

MF MF12 = min for =2 1ρ ρ (3)

MFDF

MF MF MF12

1 2

1 21=

−+

⎧

⎨⎪⎪

⎩⎪⎪

⎫

⎬⎪⎪

⎭⎪⎪

− +min , ( )max min min

ρ ρρ ρ

for <2 1ρ ρ (4)

where DF = density factor (typically < 0.03) which scales the sensitivity of mixing function to the difference in densities; and MF

max and MF

min = maximum and minimum portions of the maximum

mixing flow (typically 0.98 to 0.02). The mixing parameters were found through a calibration exercise. A wind component to mixing flows in the top three compartments is also assigned during the ice-free season.

Figure 6. Diagram of TDF model interactions for compartment 3, relating to oxidation of reduced iron (P3), reduction of oxidized iron (OxFe3), and dissolved oxygen (DO3) and alkalinity (Alk3).


43

2.6 Discussion of methods for chemical reactions and metals scavenging

Chemical constituent mass balances applied for the mass balance model are complex, and not addressed in detail here. Oxidation and reduction kinetics are handled with pseudo first-order and temperature-dependent reaction rate expressions. Oxidized iron is expected to settle and, if the alkalinity is suitable, scavenge dissolved metals. Simplified complexation expressions and coeffi-cients are set for each metal (nickel, copper, and mercury), as calibrated according to observation of trends from the recovery period after the 1980’s tailings operation in the TDF. A maximum rate of iron oxide settling is selected (typically 0.03 m/d), but the actual settling rate is scaled from zero to 100% of this rate by an empirical alkalinity pivot function. The alkalinity pivot function applied for the model is shown in Figure 7. At low alkalinities, scavenging is greatly reduced. Also, as oxidized iron settles into a compartment dominated by reduced conditions, it will be con-verted by the model to the reduced form and the scavenging effect is greatly reduced. Alkalinities are affected by the discharge, oxidation and reduction reactions (Fig. 6).

3 RESULTS

3.1 Temperatures, densities, and TDF stratification

Outcomes of the modeling for compartment temperatures are shown in Figure 8. The upper com-partment temperatures are affected by convection, conduction, and radiation gains and losses, and cycle seasonally. A calibration exercise was also undertaken to estimate surface heat transfer parameters so that the ice season and ice thickness was reasonable.

For the lower compartments, the temperature of the discharge was higher than that of compart-ment 5, and generally leads to the warming of the lower compartments over time. Implications of mixing and thermal stratification are not severe, however, because the higher dissolved solids

Figure 7. Empirical alkalinity pivot function used to regulate potential metals removal.


44

concentrations at depth tend to offset the decrease in density that would be expected from increas-ing the temperature. Model outcomes for density trends are shown in Figure 9. For the upper compartments, the large, seasonal temperature swings at the surface lead to significant instability and mixing of the upper two compartments. However, compartment 3 maintains a higher density than the upper two compartments, and is expected to remain somewhat stratified. For the lower three compartments, the discharge plume density remains above or near the density of compart-ment 5, and so it is distributed primarily to compartment 5 throughout the loading period. For the last 3 years about half of the discharge plume is directed to compartment 4. Despite significant warming of the lower compartments, the density stratification of the lower compartments is main-tained, primarily because of the higher dissolved solids concentration in the discharge. Modeled stratification remains for at least seven years after the loading.

It is important to note that the model considered efficient heat transfer between compartment 5 and the active tailings bed, during formation of the discharge plume and by heat transfer from the active tailings bed. Also important was that the heat and mass transfer models considered compression of the lower compartment volumes due to the tailings loadings.

3.2 Trends of profiles in physical parameters, alkalinity and dissolved oxygen

Model results for alkalinity, dissolved oxygen, and the physical water quality parameters of water temperature and density are shown in Figure 10. Due to a high alkalinity of the discharge (900 mg/L

Figure 8. TDF model outcomes of water temperature trends for TDF during and after tailings loading, for upper three compartments and air temperature (top graph) and lower three compartments and discharge plume (bottom graph).


45

as CaCO3) and stable stratification, alkalinity increases significantly at depth during the period

of loading, then decreases slightly due to mixing and oxygen consumption. Dissolved oxygen remains low at the lower compartments, but some rebound is observed for compartment 4 at year 14. The dissolved oxygen concentrations at the surface do not change appreciably. Temperature and density profiles also indicate that stratification should remain.

The trends in compartmental metal concentration depth profiles are shown in Figure 11. Before loading, the nickel concentration is above the preliminary effluent limit (PEL) in compartments 4 and 5. However, as the loading increases, the nickel concentration actually drops, even though the nickel concentration in the mill discharge was input as 2203 μg/L. Dilution and scavenging of nickel and other metals cause the concentration decrease, especially after loading is complete (see profiles for years 10 and 14). The removal by scavenging is expected to be particularly effective in the lower compartments, due to stratification and higher alkalinities. A similar result is obtained for copper (mill discharge set to 2760 μg/L). However, the extent of complexation and scaveng-ing was set to a lower level for copper, and the loading effects were more pronounced. Still, water quality at the surface and the recovery of water quality at the lower compartments is expected to be excellent relevant to the discharge quality.

Modeled trends in the water quality profile for mercury indicate a more significant impact from the loading, primarily because of the low initial concentrations in the TDF relative to the modeled mill discharge concentration (200 ng/L). Initial mercury concentrations were actually non-detect in the TDF, but were set to the 0.5 ng/L as a conservative measure. The loading led to an

Figure 9. TDF model outcomes of water density trends for TDF during and after tailings loading, for top three compartments (top graph) and lower three compartments and discharge plume (bottom graph).


46

increase in mercury concentrations at depth, but stratification and scavenging was also expected to be effective for mercury, leading to low surface concentrations and recovering concentrations at depth after loading.

An additional TDF model run was made to simulate a nonreactive tracer. The mill discharge concentration was set to 1000 μg/L and the initial and background concentrations were set to

Figure 10. Trends in TDF model outcome depth profiles for alkalinity (top set), dissolved oxygen, water temperature, and water density (bottom set). Profiles are drawn for the month of April for all years.


47

zero. The simulation time was extended to roughly 27 years. Concentrations in compartment 5 increased steadily from roughly 110 μg/L after 1 year of loading to a peak of 530 μg/L at the end of loading. After loading the concentration drops steadily to roughly 200 μg/L at year 27. At the surface, the concentrations are 0.0 μg/L initially, 15–33 μg/L from years 7 through 14, and then declining to roughly 20 μg/L at 27 years. During loading the surface outlet concentrations are gen-erally less than 1.5% of the discharge concentration, and increase to roughly 3% of the discharge at 14 years. After 14 years, there is little decay of the tracer in the TDF outlet, due to the significant inventory in at depth and no internal removal within the TDF.

4 CONCLUSIONS

The multi-component, multi-compartment mass and heat balance model for the TDF is complex, involving several important coupling processes. Further details of monitored conditions and chemical processes have been addressed in a companion paper. While full details are not presented here, the main modeling strategies and outcomes have been presented. The model outcomes show that the density stratification of the TDF should remain and that surface concentrations of nickel,

Figure 11. Trends in TDF model outcome water depth profiles for nickel (top set), copper (middle set), and mercury (bottom set).


48

copper, and mercury are expected to be significantly below the preliminary effluent limits for the whole period of TDF operation and recovery.

REFERENCES

Castendyk, D.N. & Webster-Brown, J.G. 2007a. Sensitivity analysis in pit lake prediction, Martha Mine, New Zealand 1: Relationship between turnover and input water density, Chemical Geology, 244:42–55.

Castendyk, D.N. & Webster-Brown, J.G. 2007b. Sensitivity analysis in pit lake prediction, Martha Mine, New Zealand 2: Geochemistry, water-rock reactions, and surface adsorption, Chemical Geology, 244:56–73.

Fraser, W.W. & Robertson, J.D. 1994. Subaqueous disposal of reactive mine waste: An overview and update of case studies-MEND/CANADA, Bureau of Mines Special Publication SP 06 A-94, pp. 250–259.

MEND 1992. MEND Project 2.11.1d, A critical review of MEND studies conducted to 1991 on subaqueous disposal of tailings. CANMET, Natural Resources Canada, Ottawa.

MEND 1996. MEND Project 2.11.1e, Review of MEND studies on the subaqueous disposal of tailings (1993–95). CANMET, Natural Resources Canada, Ottawa.

Morin, K.A. 1993. Rates of sulfide oxidation in submerged environments: Implications for subaqueous dis-posal, Proceedings of the Seventeenth Annual Mine Reclamation, Port Hardy, British Columbia, May 4–7, pp. 235–237.

Moses, C.O. & Herman, J.S. 1989. Pyrite oxidation at circumneutral pH, Geochimica et Cosmochimica Acta, 55(2): 471–482.

Peacey, V.; Yanful, E.K. & Payne, R. 2002. Field study of geochemistry and solute fluxes in flooded uranium mine tailings, Canadian Geotechnical Journal, 39(2): 357–376.

Pedersen, T.F.; Mueller, B.; McNee, J.J. & Pelletier, C.A. 1993. The early diagenesis of submerged sulphide-rich mine tailings in Anderson Lake, Manitoba, Canadian Journal of Earth Sciences, 30(6): 1099–1109.

Peinerud, E. 2003. A literature review on subaqueous tailings disposal. MiMi 2003:5, The MISTRA-programme MiMi, Mitigation of the environmental impact from mining waste program, Stockholm, Sweden.

Tchobanoglous, G. & Schroeder, E.D.,1987. Water Quality: Characteristics-Modeling-Modification. Addison-Wesley.


49

Water chemistry and metal cycling in a subaqueous tailings disposal facility

J.B. Manchester & G.R. EykholtFoth Infrastructure & Environment, LLC, Madison, WI, USA

S.V. DonohueFoth Infrastructure & Environment, LLC, Green Bay, WI, USA

J.C. CherryKennecott Eagle Minerals Corporation, Marquette, MI, USA

ABSTRACT: The flooded pit created after closure of a surface mine was subsequently used dur-ing the 1980s to contain pyritic tailings from an adjacent mill, thus becoming a Tailings Disposal Facility (TDF). A mining company proposes to place new tailings from sulfide ore processing in the TDF. Investigation of the water chemistry in the TDF reveals a water column stratified with respect to temperature, dissolved oxygen, oxidation-reduction potential and the concentrations of many metals. Further research strongly indicates that upward diffusion of high concentrations of metals in bottom waters is limited by an active scavenging process mediated by iron hydroxides. Iron cycling maintains the scavenging process and is possible due to reducing conditions at the bottom of the TDF maintained by the sulfide tailings.

1 INTRODUCTION

Disposal of mill tailings after the milling of sulfide ores presents challenges because of the poten-tial for oxidation of sulfide remaining in the tailings. Exposure of sulfide tailings to atmospheric oxygen may begin a cascade of chemical reactions that result in Acid Rock Drainage (ARD). The flow of acidic water over tailings can then increase the potential for metal leaching from tailings and surrounding rock. It is now widely accepted that placing sulfidic tailings under a water cover will reduce exposure to atmospheric oxygen, and so essentially eliminate ARD and greatly reduce leaching of metals. Where conditions permit, subaqueous disposal is now the preferred method for long-term entombment of sulfidic tailings.

An example of the successful use of subaqueous disposal of sulfidic tailings is provided by a Tailings Disposal Facility (TDF) that has operated for over a decade. The TDF is a former iron mine excavation (pit). The pit flooded after mining ceased; water depth was originally about three hundred feet. Sulfidic tailings from an adjacent mill were placed in the pit during the 1980s, converting it to a TDF and reducing water depth to about two hundred feet. A mining company proposes to place new tailings under the water cover of the TDF. A study conducted in preparation for tailings placement shows that the water chemistry of the TDF actively limits transport of dis-solved metals from deep water to surface water.

Recent measurements show that the TDF is vertically stratified with respect to temperature, dissolved oxygen concentration, oxidation-reduction potential (ORP) and the concentration of a number of dissolved metals. Low concentrations of dissolved organic carbon indicate that deep water anoxia is maintained by previously placed sulfidic tailings. Anoxic bottom waters chemi-cally reduce available dissolved iron and create an iron redox cycle that effectively limits metal transport away from TDF bottom water.



50

This paper presents a summary of field and laboratory measurements of water chemistry, as well as a conceptual model of water stratification and metal cycling in the TDF. Information generated as a result of the measurements and conceptual model was used to create a numerical model of the TDF. The numerical model is discussed in a companion paper and references to parameterization of that model are included here.

Methods and materials used to obtain the data presented herein conform to recognized standards for environmental data collection. Field parameters were measured using electronic instrumenta-tion. Water quality parameters were measured by a NELAP certified laboratory using traceable field sampling methodology.

2 FIELD MEASUREMENTS

Measurements of temperature, pH, dissolved oxygen (DO), and specific conductance as a func-tion of TDF water depth were collected in May, 2007. Parameters were measured with three-foot depth resolution at seven locations on the TDF water surface.

Measurements of each of the four parameters show little variability at a given depth across all seven stations, indicating that TDF water is well mixed laterally, and values presented here are the average of each parameter at each depth across all seven locations. However, all four parameters varied with water depth and indicate that the TDF water column is vertically stratified. The vari-ous elements of the vertical structure present in the TDF water column are more apparent when all four profiles are displayed in a single figure; Figure 1 shows this comparison. Because units and

Figure 1. Depth profiles of pH, temperature, specific conductance, and dissolved oxygen in the TDF during May, 2007. Profiles are plotted relative to the maximum value in each profile.


51

scaling differ among the profiles, each profile is plotted in Figure 1 relative to its maximum value. Thus, the water temperature falls deep in the TDF to about 40% (value of 0.4) of its maximum value in surface water.

2.1 Water temperature

A compartmentalization of temperate water bodies can occur when a thermocline is created. As springtime air temperatures increase and the water body warms from the surface downward, the surface water will be warmer than the deep water. Depending on specifics of thermal and mass transfer in a particular water body, the warm surface water may become thermally separated from deep cold water, creating two stable compartments. The stability of the two compartments is a result of the density differences between the water in each compartment, with lower density warm water floating on top of denser cold water. The boundary between these compartments is known as the thermocline and is characterized by an extreme vertical temperature gradient relative to temperature gradients in each compartment. Mixing between compartments is reduced when a thermocline is present.

The temperature measurements shown in Figure 1 display a distinct thermocline (surface water temperatures were about 20°C; bottom water, 5°C). Temperature stratification begins sharply at approximately 20 feet below the surface and is complete by the 50 foot depth. The presence of a thermocline indicates that the TDF must be modeled with at least two vertical compartments.

2.1.1 Seasonal persistence of the TDF thermoclineTemperature depth profiles observed in April and October, 2006 show that the TDF was thermally stratified in the spring and fall. Thermal stability of a body of water may be quantified by the den-simetric Froude number (Tchobanoglous and Schroeder, 1987), a dimensionless ratio comparing inertial and gravitational forces acting on the body of water:

N

Q b d

g dDF = ×

× ×/( )

( / )Δρ ρ (1)

where NDF

= densimetric Froude number; Q = volume flow; b = average width; d = average depth; Δρ = top and bottom water density difference; ρ = depth-average water density; g = gravity constant.

Values of NDF

greater than one indicate that turbulent mixing prevents thermal stratification, while values less than one indicate that thermal stratification will remain despite small levels of mixing. Representative values used to calculate N

DF for the TDF were obtained from the above field

data. From equation 1, the TDF has a NDF

= 0.0000054 and is very strongly stratified. Moreover, because of the very small flow out of the facility compared to the cross-sectional area, the differ-ence in density between the upper and lower zones must be less than about 10–14 before the Froude number approaches unity. This suggests that it is unlikely that the TDF will mix completely.

2.2 Dissolved oxygen

A second important compartmentalization of a water body can occur when dissolved substances enter from sources located at either the surface or the bottom of the water body. A decreasing chemical concentration gradient away from the source will then be present. As with temperature, a single gradient may form, or a very sharp gradient may develop between two compartments, each with their own chemical gradient, and each with a unique concentration of the substance. When two chemically distinct compartments form, they are separated by a chemocline.

The dissolved oxygen profile in Figure 1 exhibits a well-defined oxygen chemocline and divides the TDF into an oxygenated upper zone (DO of about 9 mg/L) and anoxic bottom waters (oxygen below detection limits). It is important to note that the oxygen chemocline is roughly sixty feet below the thermocline. Thus, considering both temperature and dissolved oxygen, the TDF is vertically stratified into a system characterized by at least three compartments.


52

2.3 pH

Hydrogen ion concentration (pH) does not change greatly in the TDF (Fig. 1), remaining relatively constant both laterally and vertically (slightly alkaline with pH between 7 and 8). However, there is a distinct vertical structure: maximum pH at the water surface, minimum values at approximately 100 feet deep, and intermediate pH values in the deep water. Also, the pH minimum occurs at the same depth as the oxygen chemocline.

2.4 Specific conductance

Specific conductance (Fig. 1) provides a general measurement of the concentration of dissolved salts in water. The TDF may be divided into two distinct vertical compartments based on specific conductance measurements. While not as pronounced as temperature and oxygen, specific con-ductance (and thus, dissolved ions) changes rapidly from lower values (about 500 μS/cm at 25°C) in the upper waters to higher values (750 μS/cm) in the bottom waters. The specific conductance profile can be seen in Figure 1 to mirror the dissolved oxygen profile, suggesting that the con-centrations of dissolved salts increase as oxygen fall to zero. And, the compartmentalization sug-gested by specific conductance is equivalent to that indicated by dissolved oxygen.

2.5 TDF model compartmentalization

Field measurements of the TDF suggest that the water column may be viewed and modeled as comprising three vertical compartments. The air-water and the water-sediment interfaces also have unique characteristics. Thus, a complete conceptual (and numerical) model will be structured with five vertical compartments, as shown in Table 1.

3 DISSOLVED METALS STRATIFICATION

Water quality samples were collected in May, 2007 from five depths centered in the five com-partments shown in Table 1. Sampling at each depth was done at two locations to assess horizon-tal variability in the TDF. All water samples were analyzed for a suite of chemicals. Twenty-two chemicals (including related parameters such as hardness, turbidity and alkalinity) were success-fully quantified (reported values were greater than the Limit of Quantitation) at all five depths at each sampling station. In addition, ammonia was quantified at the two deepest depths (120 and 175 feet). As with field parameters, water quality data from a given depth varied little between the two locations. The results presented here are from a station near the horizontal center of the TDF.

The constituents measured in TDF water vary in concentration over five orders of magnitude. Despite this range, three general patterns with respect to depth are observed in the data and are visible in the depth profiles for nickel, copper and calcium shown in Figure 2. Some constituents, for example nickel, are present at low concentrations in surface water, but increase to high levels with increasing depth; other constituents, like copper, increase from intermediate to high values

Table 1. Compartmentalization of the TDF.

Compartment description Compartment depth range (feet)

Surface water 0–6Start of the thermocline 6–24Below the thermocline, above the oxygen chemocline 24–96Below the oxygen chemocline 96–144Deep water 144–192


53

with increasing depth; finally, another set of constituents, represented here by calcium, are nearly constant with depth.

The fourth depth profile shown in Figure 2 displays measured concentrations of iron. The shape of this profile is unique and points to the importance of iron in the TDF as a mediator of surface concentrations of a number of metals.

Almost all of the chemical constituents in the TDF are present at highest concentration at the very bottom of the TDF. This result strongly suggests that the sediments (old tailings) are the pre-dominant source of these constituents. As noted above, almost all of these chemicals then decrease to varying degree at shallower depths. The greatest decrease occurs for nickel, manganese, molyb-denum, cobalt, and antimony; all of these fall to less than 20% of their maximum concentrations. Potassium, sodium, chloride, copper, and boron also decrease, but by only about one-half their maximum levels. Finally, magnesium and calcium change very little with depth, suggesting that the native rock around and forming the TDF is the source of these cations.

The varying shapes of the metal depth profiles indicate that metal concentrations are not control-led solely by diffusion and mixing. Instead, the range of surface concentrations relative to initial

Figure 2. Depth profiles of four metals measured in the TDF demonstrating varying degree of concentration change with depth. Iron exhibits unique behavior.


54

bottom water concentrations suggests that internal water chemistry is actively removing metals from TDF surface water, and the extent of this process depends on the chemistry of each dissolved metal. The rock forming the TDF contains large amounts of iron, so the fact that iron does not behave like calcium and magnesium strongly suggests that iron participates in reactions within the TDF and that the decrease in other metal concentrations is likely due to iron chemistry in the TDF.

4 TDF CHEMICAL LIMNOLOGY

Understanding the complexity represented by the varying shapes of the metal depth profiles requires a more thorough examination of the chemical limnology of the TDF.

4.1 Background

All surface water bodies share many common chemical characteristics because of a common bulk solvent (water), and because almost all exist under an atmosphere rich in oxygen with a small, but relatively constant amount of carbon dioxide. Also, a small set of twelve elements, present at rela-tively high concentrations, dominate the chemistry of surface waters. These major aquatic elements are boron, carbon, nitrogen, oxygen, sodium, magnesium, aluminum, silicon, sulfur, chlorine, potas-sium and calcium. All other elements are considered minor elements (Buffle and De Vitre, 1993).

The chemistry of surface waters is regulated by four types of reactions: acid-base, oxidation-reduction (redox), complexation, and precipitation. The state of a particular body of water with respect to these four types of reactions is assessed with intensity parameters; pH (acid-base reac-tions) and pE (redox reactions) are common intensity parameters. The chemical state of a body of water, as described by intensity parameters, determines which chemical reactions are most prevalent. The extent of a particular type of reaction required to move a body of water to a new chemical state is quantified by capacity parameters. Alkalinity is the capacity parameter associ-ated with acid-base reactions; similar capacity parameters exist for the other three reaction types, though they are more difficult to quantify.

The most prevalent major aquatic elements (and chemical species formed from these elements) within a body of water control the acid-base and redox chemistry within that body of water, and the intensity and extent of these two types of reactions then regulate the concentration and specia-tion of both the major and minor elements in the body of water.

In many surface bodies of water, acid-base chemistry (and, therefore, pH) is dominated by the carbonate system, comprised of atmospheric carbon dioxide and calcium carbonate. In bodies of water in contact with calcareous rock, the fixed atmospheric concentration of carbon dioxide and solubility of calcium carbonate set the pH at about eight.

At the air-water interface of a body of water, the redox chemistry (and, therefore, pE) is most influenced by atmospheric oxygen. Oxygen is one of the most potent oxidants, and conditions near the air-water interface are generally very oxidizing. However, the extent to which oxygen influences redox conditions in deeper waters varies greatly, and is affected by both physical and biological processes.

Surface concentrations of oxygen are set, in part, by its solubility in water, which varies inversely with water temperature (oxygen saturation is about 8–9 mg/L at 20°C and increases to 12–13 mg/L at 5°C). Also, oxygen concentrations at increasing depths are affected by the relatively slow dif-fusion of oxygen downward from the surface. Given the constant atmospheric supply, diffusion of oxygen to deep water is continuous. However, depletion of oxygen with increasing water depth is often observed in bodies of water, as internal chemical processes consume oxygen more rapidly than diffusion can replace it.

The presence of life capable of photosynthesis (predominantly algal cells, and referred to as primary production) near the air-water interface will also affect oxygen concentrations, as well as acid-base (pH) and redox (pE) chemistry in these waters. During daylight hours, ongoing pho-tosynthesis will produce oxygen and consume carbon dioxide, increasing dissolved oxygen, pH,


55

and pE. Conversely, algal respiration during dark hours will consume oxygen and release carbon dioxide, and so, often lower pH and pE.

Oxygen concentrations in deep waters (greater than about thirty feet) depend on the rate at which oxygen is consumed by various chemical processes. In waters with high primary produc-tivity (eutrophic waters), most oxygen in waters below the surface is rapidly consumed by the decomposition of settling detritus resulting from primary productivity. Although the detrital mate-rial may consist of a number of compounds, it is predominantly organic, or reduced, carbon. Often in eutrophic systems, the amount of organic carbon produced greatly exceeds that needed to remove essentially all oxygen from deep waters. In these systems, an oxygen chemocline will develop, the bottom waters will become anoxic, and excess organic carbon will accumulate in bottom sediments.

Organic carbon is a potent reductant, and in a system like that described above, a redox gradient will be present, with oxidizing conditions in surface waters and reducing conditions in deep water. The system represented by oxygen reduction and organic carbon oxidation sets the possible redox (pE) range for most bodies of water. With acidic surface water, pE values in these systems range between almost 20 at the water surface to −10 at (or just under) the sediment surface.

If primary production is very low in a body of water (oligotrophic waters), the extent to which oxygen is depleted in deep water will depend on the amount of reductants present from sources other than the settling detritus of primary production. These may still include organic carbon which enters the water body from outside (allochthonous carbon), but also include reduced nitro-gen, manganese, iron, and, importantly, sulfur.

In water bodies where organic carbon is in limited supply, sulfur compounds may become the dominant reductants in sediments. Sulfur compounds are common in most natural systems, and are particularly prevalent in those systems in contact with sulfide minerals.

Once the large-scale chemistry of a water body is established by the major aquatic elements through acid-base and redox reactions, the fate of the minor aquatic elements is largely deter-mined by those complexation and precipitation reactions that are possible under prevailing pH and pE conditions. Minor elements are most often removed from the water column through complexation reactions with the surface of aquatic particles, and subsequent sedimentation of these particles.

Settling particles with suitable complexation sites for minor element binding may come from a variety of sources, including allochthonous organic matter, clay particles, and iron, aluminum and manganese oxide particles. In some water bodies, autochthonous (formed within the water body) colloids form from precipitation of major and minor elements within the water column. The colloid surfaces often contain a large number of complexation sites and are in close proximity to dissolved metal ions in the water body. This combination results in effective scavenging of the dis-solved metals by the colloids. Subsequent coagulation of these colloids produces settling particles. This completely internal mechanism effectively reduces surface concentrations of a number of metals by transporting them to deep waters.

Dissolved oxygen, primary production and reduced carbon, oxidation-reduction (redox) chem-istry, and iron cycling are discussed below with respect to metal concentrations in the TDF.

4.2 Dissolved oxygen and primary production

Graph A in Figure 3 displays measurements of dissolved oxygen measured in the TDF during July, 2007. The oxygen chemocline observed in the May profile is also present in the July profile. The loss of oxygen occurs at the same depths in both profiles, with anoxic water below 100 feet in both cases. Thus, oxygen dynamics are very stable over at least months of time, and show no seasonal-ity between spring and summer.

Also shown in Graph A, Figure 3 is a depth profile of measurements of dissolved oxygen in pit water collected in 1984. In 1984, before tailings were placed, the pit was about 300 feet deep, and the historical measurements show that the entire water column was oxygenated, with dissolved oxygen concentrations at the bottom equal to about 75% of surface values.


56

Comparison of the historical and contemporary oxygen profiles shows that, in the surface waters of the TDF, little has changed with oxygen dynamics over about twenty three years. The profiles show very similar surface concentrations, and each profile captures the increased oxygen due to decreasing water temperature.

However, below about 75 feet the two profiles are very different. In 1984, the entire pit was oxy-genated, whereas currently the TDF is anoxic below about 100 feet. Given that atmospheric oxygen levels and rates of oxygen diffusion through water have not changed over the past twenty years, the difference in bottom water oxygen over this time is most likely due to an increase in oxygen demand, that is, an increase in chemical reductants, in the bottom of the TDF during this time.

An increase in reductants at the bottom of a body of water most often occurs as a result of increased primary productivity in the surface waters. Also, an increase in input of allochthonous carbon might occur. In either case, more organic (reduced) carbon would reach the sediments, leading to an increase in oxygen consumption, and oxygen depletion in bottom waters.

Figure 3. Water column profiles of dissolved oxygen in the TDF. Graph A shows historical and recent dis-solved oxygen concentration. Graph B shows early morning and late afternoon dissolved oxygen, and water temperature.


57

Aerial photography shows that land usage has not changed greatly around the TDF in the last several decades, so allochthonous carbon loading to the TDF has likely been relatively constant during this time. Organic carbon measurements were obtained as a part of the May, 2007 field sampling. Most samples were below detection limits. Quantified samples had an average value of 2.39 mg/L. Based on measured organic carbon, the TDF would be classified as, at most, an oligotrophic system (Wetzel, 2001).

The extent of primary production may also be assessed by examining dissolved oxygen. Two sets of dissolved oxygen measurements are shown in Graph B, Figure 3. These measurements were taken on the same July day as the measurements shown in Graph A, but at different times. The first set shown in Graph B was obtained at about 5:00 AM, while the second set was measured at about 4:00 PM.

As explained above, photosynthesizing algal cells would release a pulse of oxygen during the day and consume oxygen at night, resulting in very different dissolved oxygen levels early and late in the day. The fact that the oxygen profiles in Graph B are almost identical strongly indicates that photosynthesis, and therefore, primary production, is essentially not occurring in the TDF.

Measurements of organic carbon and dissolved oxygen show no evidence of high levels of organic carbon production in the water column of the TDF, and so little organic carbon is expected in the sediments. The depletion of oxygen in the bottom waters must then be due, in part, to con-sumption of oxygen by other reductants in the sediments.

The change in oxygen dynamics between 1984 and present, visible in Graph A, Figure 3, and the fact that the tailings were placed during this interval suggests that the tailings are helping to maintain anoxic conditions in the TDF. In this scenario, oxygen continually diffuses into the waters of the TDF, past depths where primary production would otherwise begin to consume it. At deeper depths, oxygen encounters in situ reductants, including small amounts of predominantly allochthonous organic carbon, and is at least partially consumed by these reductants. Any oxygen that remains at deeper depths eventually reacts with reduced sulfur from mine tailings. Thus, the large mass of sulfidic mine tailings is self-stabilizing, in that variations in amounts of other reduct-ants, principally organic carbon, cannot lead to a build up of oxygen near the tailings.

4.3 Oxidation-reduction chemistry

The chemistry of the TDF was further investigated by measuring the oxidation-reduction poten-tial (ORP or redox potential) in July, 2007. Measurements were taken at three-foot intervals and converted to Standard Hydrogen Potentials (E

H) using a calibration measurement obtained in the

field. Measured values compared favorably with values obtained through a Nernst calculation using measured concentrations of the ammonium and nitrate redox couple. The redox potential profile for the TDF, expressed as E

H and pE is shown, along with temperature, dissolved oxygen

and calculated values, in Figure 4.The redox potential profile is useful for evaluating the redox conditions within a body of water,

as indicated by the associated intensity parameter, pE. Because bodies of water are complex, containing many chemical species and often several redox couples, redox profiles should be interpreted with care and cannot be expected to predict individual chemistries. However, field measurements are very useful in understanding relative conditions within a body of water, and may provide chemical information when compared to literature values of redox potentials.

It is also important, when interpreting redox profiles, to recognize that the presence of a spatial gradient implies that the system is not at equilibrium; kinetics is as important as thermodynamics in determining the type and amount of chemical species present at any given time.

As with most bodies of water, the redox profile for the TDF displays a reduction gradient with increasing water depth; waters are most oxidizing at the surface and are most reducing in the bot-tom waters. The profile is erratic in the oxygenated top waters, but much more stable in the anoxic bottom waters. Although a gradient is present, the range of pE values is small, varying between about 7 at the surface and 5.5 near the bottom; as mentioned previously, bodies of water may vary by as much as +20 at the surface and -10 at bottom sediments. The first value in the redox profile


58

is at a depth of three feet below the surface. It is likely that water just below the surface would have higher pE values, perhaps approaching the theoretical maximum for oxygen of 13.75 at pH = 7.

Common redox reactions are shown in Figure 4. Each reaction has a pE°(W) value. This is the standard pE° when pH = 7. These values allow direct comparison of redox half-reactions under conditions similar to those often found in aquatic environments (Stumm and Morgan, 1996). The reactions are in order of decreasing pE°(W). It is thermodynamically possible for the oxidant in a reaction with a higher pE°(W) to oxidize a reductant at a lower pE°(W). Whether or not this occurs depends on the concentration of each species in the reaction.

While specific chemistries cannot be known based on the redox profile alone, several infer-ences can be made. A simple linear gradient connects the top two pE values with values below 100 feet. This indicates that oxygen (the most potent oxidant in the reactions show in Figure 4) is controlling the redox conditions in the water above about 100 feet but is encountering several reductants at various depths in the top waters. These reductants are oxidized by the oxygen and pull pE values lower until they are consumed. Narrow bands where this occurs are referred to as redox boundaries, and several major and minor boundaries are visible in the profile.

One of the reductants present in the top waters of the TDF is organic carbon, particularly from allochthonous sources. The carbon redox couple is shown as the last reaction listed in Figure 4

Figure 4. Oxidation-Reduction (redox) potential measured in the TDF. Common aquatic redox reactions are shown with associated environmentally relevant pE values.


59

(the six-carbon sugar glucose is a model organic carbon). The large difference in pE°(W) values between the oxygen and carbon redox reactions implies that oxidation of reduced carbon by oxygen will essentially go to completion, such that the reactant present in excess will completely con-sume the other reactant. As mention previously, in eutrophic systems excess carbon eventually consumes all oxygen deep in the water column and accumulates in bottom sediments. If this were the case in the TDF, the bottom waters would be expected to have a lower pE than that observed in the profile, given the negative pE°(W) associated with the carbon couple.

Other possible oxygen-consuming reductants are shown in the reactions listed in Figure 4. Thus, oxidation of nitrogen compounds, and the formation of oxidized solid phases of manganese and iron from their reduced, dissolved metal ions can all contribute to oxygen consumption within the TDF.

In the TDF, sulfate concentrations are high, particularly in the bottom waters, and pH is about 7. Thus, under existing conditions, the bottom two reactions shown in Figure 4 may couple such that sulfate is reduced and any remaining organic carbon present at the bottom of the TDF is con-sumed. In the process, bisulfide (HS-) is produced. Bisulfide is also a product of any dissolution of metal sulfide tailings. Production of bisulfide by oxidation of organic carbon would therefore be expected to limit dissolution of tailings and as the case with dissolved oxygen, the sulfide tailings would be self-stabilizing under the redox conditions in the TDF.

As with oxygen at the surface, the presence of bisulfide at or just in the sediments (tailings) suggests that the actual pE at the very bottom of the TDF is lower than measured values indicate, and may approach the theoretical maximum for sulfur of around −3 at pH = 7. Thus, the actual shape of the redox profile may be sigmoid, with a central linear portion resembling the measured profile and long tails at each end very near the air-water and water-sediment interfaces.

4.4 Iron cycling and metal scavenging

The iron cycle has long been recognized as an important mechanism in water bodies for the regu-lation of a number of species, including trace nutrients (Stumm and Morgan, 1996). The relation-ship between reduced and oxidized iron is of primary importance in the iron cycle, as is the fact that iron changes oxidation state when transported between oxic and anoxic portions of the water column. The main chemical reaction describing this relationship is shown below

Fe

1

4O

5

2H O Fe(OH) (amorph,s) 2H2

2 2 3+ ++ + = +

(2)

It is important to note that Reaction 2 affects alkalinity. In oxic waters, iron is oxidized and hydrogen ions are released, thereby reducing alkalinity.

The above reaction is a redox reaction and involves release of hydrogen ions. Therefore, the direction of the reaction not only depends on the presence of oxygen, but also on pE and pH. Also, iron undergoes a phase change from reduced dissolved to oxidized solid. Several other reactions involving the iron couple in water, and various dissolved and solid phases, are possible. To understand the iron cycle, and if and when it may occur in a water body, it is necessary to describe these various states of iron and to know which state will be dominant under different pE and pH conditions. A useful tool for this type of analysis is the pE-pH diagram.

A pE-pH diagram for the iron-carbon dioxide-water system is shown in Figure 5. The diagram displays regions bounded by lines within which a particular form of iron is the dominant species. The axes show the values of pE and pH so that for any set of pE-pH values, the dominant species can be found. The dashed lines on the diagram indicate the stability zone for water.

The dominant iron species in the TDF can be found by reading the diagram. pH in the TDF ranges between 7 and 8, and pE values over the entire water column, including estimates near the air and sediment interfaces, range between −3 and about 12. By examining the area on the dia-gram enclosed by these pE-pH ranges, it is found that amorphous iron (III) hydroxide solid and dissolved iron (II) are the dominant iron species in the TDF. These two iron species form the iron cycle in water bodies, so it is very likely that this cycle is operating in the TDF.


60

The iron cycle is displayed in the lower diagram in Figure 5. Dissolved iron (II) enters the water column by diffusion upward from the sediments. This is common in water bodies with anoxic bottom waters, and very likely in the TDF, given the pyritic tailings that form the sedi-ments. As the upward diffusing iron (II) encounters oxygen diffusing downward from above, the iron is oxidized, producing amorphous solid iron (III) hydroxide, initially as colloids. The colloid surfaces provide many complexation sites, and various dissolved minor aquatic metals react with these sites to form surface-bound complex ions (this process is also known as adsorption). As the oxidized iron oxide colloids and associated metals coagulate, larger particles form and begin to settle back to low pE, anoxic conditions. Some of the particles begin to undergo reductive dis-solution, while other, larger particles deposit in the sediments. As the cycle continues, some of the iron (III) oxides on the sediment surface are reductively dissolved. This provides more iron (II) for the cycle, while also releasing minor metals. The released metals may diffuse upward, but they are prevented from moving beyond the bottom waters by the iron cycle. In this way, high concentrations of dissolved minor metals may appear in the deep waters, but these metals cannot reach the surface water.

Figure 5. pE-pH diagram for the iron-carbon dioxide-water system and a diagram of the iron cycle com-monly found in bodies of water with anoxic bottom waters.


61

The cycle is strengthened by allochthonous iron entering from above and settling toward the bottom, adding to the pool of available iron. Measurements not discussed here indicate that ground water near the south end of the TDF has high dissolved iron. Additional iron, both reduced and oxidized may enter the TDF at the south end and help sustain the iron cycle, and so help retain dissolved metals deep in the TDF.

5 CONCLUSIONS

Measurements in the TDF of field and water quality parameters demonstrate that the facility is operating as an active treatment system. Oxidation of sulfidic tailings placed under the water cover of the TDF is greatly limited by anoxia in the bottom waters. Anoxia is maintained by the tailings and does not depend on the presence of organic carbon. Upward diffusion of dissolved metals associated with the tailings is halted by an ongoing iron cycle that effectively scavenges metals from the water column and returns them to the bottom waters. The controlling water chemistry is maintained by internal processes and is expected to continue to operate if additional tailings are added. A numerical model designed to estimate TDF water quality trends based on the conceptual model developed here is presented in a companion paper.

REFERENCES

Buffle, J. & De Vitre, R.R. 1993. Chemical and biological regulations of aquatic systems. Boca Raton: Lewis Publishers.

Stumm, W. & Morgan, J.J. 1996. Aquatic chemistry: Chemical equilibria and rates in natural waters. Third edition. New York: John Wiley & Sons, Inc.

Tchobanoglous, G. & Schroeder, E.D. 1987. Water Quality: Characteristics-Modeling-Modification. Addison-Wesley.

Wetzel, R.G. 2001. Limnology: Lake and river ecosystems. Third edition. San Diego: Academic Press, Elsevier.



63


Heap leach lixiviant flow—myth versus reality

James R. KunkelKnight Piésold and Co., Denver, CO, USA

ABSTRACT: Flow of lixiviant through heap leach materials should occur as unsaturated flow under the rate of lixiviant application chosen by the operator. Often, the unsaturated hydraulic characteristics of the ore being leached are not well known or understood. Recovery of the metal resource from the ore is more about the solution flow characteristics than the material itself; whereas, unsaturated flow characteristics in heaps are more about the material than the fluids (air and lixiviant). Opposing driving forces and non-linear flow mechanisms result in consequences which may be non-intuitive to the operator.

This paper takes the generally accepted assumptions by operators about lixiviant flow in the unsaturated heap leach ore and exposes the myths and corrects the misunderstandings. Some of the myths and realities discussed include the following: liquid fluid velocities based on lixiviant application rates, spread of lixiviant based on lixiviant application rates, impacts of fines on leaching, and impacts of air and liquid flow interactions on resource recovery. The conclusions reached in the paper are supported by observed responses from operating heap leach systems for copper, silver, and gold.

1 INTRODUCTION

The design and operation of heap leach facilities typically requires a broad range of technical expertise including exploration geologists, metallurgists, geotechnical engineers, mining engi-neers, geochemists, hydrologists, and geo-hydro-environmental engineer/scientists. In many cases, the design and operation of heap leaching in relation to lixiviant application rate, drain down time, and fluid management is left to experience rather than science. Usually, unsaturated material hydraulic and geotechnical properties, flow modeling, and other unsaturated material behaviors are not included in the heap leach design (van Zyl 2008).

Heap leaching of copper, gold, silver, zinc and other metals is done under unsaturated liquid flow conditions, often with upward flow of air concurrently with downward flow of lixiviant. Throughout the mining industry, heap material particle sizes range from boulders (greater than 300 mm median diameter) to fines (less than 0.074 mm median diameter) consisting of silt and clay-sized particles. Often the heap materials contain a wide range of particle sizes from run-of-mine through primary to tertiary- or quaternary-crushed and agglomerated ores.

Lixiviant application rates are selected for each heap leach project such that unsaturated flow conditions are maintained and, if air is required to help the leaching process, upward air flow is maintained during downward lixiviant flow. Lixiviant application rates for metals recovery during heap leaching may range from 1 to 10 liters per hour per square meter (L/hr/m2). Drain gravel and drain pipes overlying a plastic liner at the bottom of the heap are used to capture pregnant leachate and conduct it to metal recovery processes. Because liquid will not enter a pipe if the surrounding gravel or other porous medium is unsaturated, a zone of saturation is required at the drain pipe at the bottom of the heap.

This paper examines five aspects of generally accepted fluid flow and fines transport in heaps and shows that typically these concepts are misunderstood. The reality of the “myths” about fluid


64

flow in heap leaching operations are presented along with a discussion of the reality physics asso-ciated with fluid flows in heaps.

2 MYTHS AND REALITIES

2.1 Myth 1 and reality 1: Lixiviant flow characteristics in unsaturated porous media

Myth 1: In heap leach operations, flow through unsaturated porous media is via a uniform wetting front.

Reality 1: Depending on the hydraulic characteristics of the ore and the lixiviant application rate, the flow may occur as (1) a uniform wetting front, (2) capillary fingering, or (3) viscous fingering.

Discussion: Based on linear hydrodynamic stability analysis, Philip (1975) generalized the criteria for flow instability as: flow instability occurs whenever the pressure gradient, G, at the wetting front opposes the flow. Similarly, Raats (1973) indicated that flow instability occurs when-ever the pressure head at the soil surface is smaller than at the wetting front. All three of the above stability criteria (Hillel and Baker 1988, Philip 1975, and Raats 1973) defined flow instability in porous media. These concepts are shown schematically on Figure 1.

Examples of these behaviours in heap leach operations include (1) air compression ahead of the lixiviant wetting front, (2) lixiviant application rate smaller than the saturated hydraulic conductivity, K

s, of the ore, (3) layered heap materials (fine over coarse), and (4) certain lixiviant

redistribution conditions in the heap. Therefore, because these circumstances are difficult to avoid

Figure 1. Schematic of stability criteria for flow in unsaturated porous media.


65

in actual operations, it appears that unstable, and perhaps fingering flow, may be the norm in heap leaching and the locations, sizes, and durations of the fingers cannot yet be predicted mathemati-cally. The literature also indicates that once preferential flow occurs, it is unlikely that wetting front flow can be re-established without mechanically re-working the ore (Jury, 2006).

Or (2006) has proposed using the capillary and Bond numbers (defined below) as measures of unstable flow conditions in unsaturated porous media. The capillary number, Ca, is the ratio of vis-cous to capillary forces, and the Bond number, Bo, is the ratio of gravitational to capillary forces.

The capillary number is defined mathematically as

CaV= μσ

where μ is the dynamic viscosity of the invading fluid (lixiviant or air), V is a characteristic veloc-ity, sometimes taken as the average Darcian velocity and/or the average pore velocity for steady state lixiviant applications, and σ is the surface or interfacial tension between the air and lixiviant phases. For low capillary numbers (a rule of thumb says less than 10−5), flow in the heap is domi-nated by capillary forces.

The Bond number is defined mathematically as

Bogr= Δρσ

2

where Δρ is the density or the density difference between air and lixiviant, g the acceleration due to gravity, r the “characteristic length scale”, e.g. average radius of a pore or the radius of a capil-lary tube, and σ is the surface tension of the air-liquid interface. The Bond number is a measure of the importance of surface tension forces compared to gravity forces. A high Bond number indi-cates that the system is relatively unaffected by surface tension effects; a low Bond number (typi-cally less than one is the requirement) indicates that capillary forces (surface tension) dominate. Intermediate numbers indicate a non-trivial balance between the two effects. Or (2006) concludes that a Bo of approximately 0.05 is an estimate of the threshold for gravity-driven unstable flows. That is, for Bo > 0.05, unsaturated zone flows are stable and Richards’ equation applies. For Bo < 0.05, unsaturated zone flows are unstable and capillary or viscous fingering (preferential flow) may occur as demonstrated schematically on Figure 2.

Or (2006) indicates that a generalized Bond number Bo* = Bo – Ca; can be used which incor-porates porous medium characteristics through mean pore size and width of pore (throat) size dis-tribution of the medium. The interpretation of this generalized Bond and capillary numbers related to stable displacement of liquid (a wetting front) is Bo* > 0 or Bo > Ca. For capillary fingering the generalized Bond and capillary numbers interpretation is Bo* < 0 and Ca << 1. For more dominant viscous forces (Bo* more negative) front morphology transitions to viscous fingering where the interpretation of the generalized Bond and capillary numbers is given by Bo* < 0 and Ca >> 1.

The stability criteria of Raats (1973) and later Philip (1975) are still applicable. However, the criterion Bo* > 0 is more general and equivalent to other stability criteria. A roadmap for unsatu-rated flow behaviour expressed by the above theory and equations has been summarized by Or (2006) and is presented on Figure 2.

Another method of assessing the stability of unsaturated flow behaviour in porous media is to use the Lenormand et al. (1988) phase diagram shown on Figure 3. Lenormand et al. (1988) introduced the concept of a “phase-diagram” for drainage displacements where vari-ous experiments and simulation were plotted in a plane with the logarithm of the capillary number, Ca along the y-axis and the logarithm of the viscosity ratio, M, along the x-axis (see below for a definition of viscosity ratio). The plot, reproduced on Figure 3, clearly shows that the different structures they obtained divide into the major flow regimes whose region of valid-ity in Ca and M space is given by the plot. The boundaries of the regions were qualitatively


66

discussed and they concluded that the drainage displacements were fully characterized by Ca and M. However, Lenormand et al. (1988) mentioned that changing the pore size distribution of the simulation and the experiments resulted in translations of the boundaries but that the general shape should remain unchanged. Because the capillary number does not take into account the pore size distribution, a careful analysis is required to understand better this effect before any complete “phase-diagram” can be drawn.

The three regions bounded by thick (black) lines on Figure 3 correspond to the three major flow regimes: viscous fingering, stable displacement and capillary fingering obtained in the simula-tions and experiments performed by Lenormand et al. (1988). Note that the capillary number in the “phase-diagram” is defined by always inserting the viscosity of the invading fluid (in this case lixiviant), even if this is the lower viscous one.

The viscosity ratio, M, is defined as the ratio of the viscosity of the defending fluid (air) to the viscosity of the invading fluid (lixiviant), and is given mathematically by

M a

w

= μμ

where μa is the viscosity of the defending air in the heap ore and μ

w is the viscosity of the invading

lixiviant in the heap ore.Given the hydraulic properties of the ore, the above generalized Bond and capillary numbers

and the viscosity ratio, including the stability criterion of Bo* > 0, were used as a powerful tool to assess the behaviour of various copper/zinc ore gradations related to fingering and/or flow insta-bilities at a lixiviant application rate of 10 L/hr/m2.

The question arises as to what to use for the “characteristic” velocity, V in the capillary number. It must be recognized that in unsaturated porous media, such as heap leach ore, the flow velocity is governed by the degree of saturation (percentage of the pores filled with lixiviant), the unsatu-rated hydraulic conductivity of the ore, and a unit gradient. In unsaturated flow the flux rate, q, is numerically equal to the unsaturated hydraulic conductivity, K(θ), (Darcian flow velocity), and given by Darcy’s law as

Figure 2. Roadmap for unsaturated flow behavior.


67

q V K= θ= ( )

where q is the net or effective lixiviant application rate and K is unsaturated hydraulic conductivity as a function of volumetric moisture content (θ). Note that because the liquid gradient in an unsaturated medium is always 1.0, Darcy’s law is simplified. The pore fluid velocity also could be used as the velocity term in the capillary number. In this case V is velocity in an individual average pore rather than the Darcian (average) velocity and is given as a function of the moisture content, θ, of the ore as

VK q= =( )θ

θ θ

which was used to calculate pore velocity for the capillary number in this paper.Results of stability analyses for several copper/zinc ore types with fines (less than 0.074 mm

median diameter) content ranging from 4 to 11 percent and a maximum grain size of 12.5 mm also are shown on Figure 3. In each case these gradations indicate that the unsaturated flow behaviour is capillary fingering (not a bad thing) at a lixiviant application rate of 10 L/hr/m3.

Figure 3. Stability results for 10 L/hr/m2 for 20 and 40°C.


68

The conclusions from these analyses is that capillary fingering rather than either a uniform wetting front or viscous fingering will occur in these particular copper/zinc ores due to their rather coarse gradations and low pore velocities, and that heap lixiviant flow can be modelled using Richards’ equation (see Figure 2).

2.2 Myth 2 and reality 2: Ore gradation characteristics

Myth 2: In heap leach operations a coarser gradation is better than a finer gradation and leaching occurs better in ores with few fines. Fines are not necessary to have an effective heap leaching operation.

Reality 2: Uniformly coarse ore particles that are poorly graded result in poor lixiviant spread-ing and poor ore wetting, resulting in poor metal recoveries.

Discussion: The “bag of marbles” myth that leaching is best done with a minimum of fines to avoid “plugging” of the ore also has been shown to be un-economical and reduce resource recov-ery. This case was seen in a gold heap leach operation where the ore was crushed to a well sorted gravel with zero fines. The operator drilled the heap and discovered that substantial portions of the heap were dry and had never seen lixiviant.

A theoretical analysis of a copper/zinc heap was done using a three dimensional analytical model of the spread of a point source (Tindall and Kunkel 1999) to show that the spread of the lixiviant from individual drippers (point sources) spread about 60 cm as it migrated through the coarse, crushed ore. The drippers were spaced at about 100 cm, leaving a large portion of the heap un-irrigated or poorly irrigated as shown on Figure 4. Whereas the percent fines in the heap was approximately 1.5 percent and flow rate was excellent through the heap, much of the ore was not wetted and the ore that did see solution had quite large velocities. For these reasons the actual heap performance would not match the laboratory column expectations for either leach rate or recovery.

The conclusions from this analysis is that for this ore the lateral lixiviant spread under steady-state flow conditions of 7 L/hr/m2 at a depth of 45 m (assumed typical lift thickness) is approximately 60 cm; approximately 7 percent of the ore would remain unleached, approximately 86 percent of the ore may be “ineffectively” or partially leached, and only 7 percent may be “effectively” leached. Thus, gradation and fines content is important in effectively leaching most ores.

2.3 Myth 3 and reality 3: Fines in heap operations

Myth 3: Fines, that is, materials with median particle diameters less the 0.074 mm (200 mesh), impede leach rates and migrate into the larger pores, plugging them.

Reality 3: Migrating fines, while a possibility in high lixiviant application rates are a practical rarity. Unless the fines are dispersive clays there is rarely adequate solution flow or velocity to mobilize fines in a heap.

Discussion: Fines separation during heap material placement (trucks or stackers) may cause low conductivity zones within the heap which leach slowly or not at all. Often the ore with high-est potential resource concentrations are the finer-grained materials. Recent techniques utilizing Hydro-Jex® technology or a similar “frac” processes, to force lixiviant into these fine ore zones is one method to leach these poor hydraulic conductivity zones (Ulrich 2008).

Some operators blame the migration of fines on the leaching process, claiming that fines are moved en mass within the heap; blocking or impeding lixiviant flow. A simple calculation exam-ple can clarify this myth. Table 1 shows hypothetical saturated and unsaturated characteristics of a copper ore.

Based on Table 1, a pore velocity of about 1 × 10–3 cm/s would be a reasonable value during “typical” leaching at 10 L/hr/m2. If the leaching rate were to be increased to 20 L/hr/m2, the pore velocity would be approximately 2 × 10–3 cm/s at 68 percent of saturation. The question is “What size particle could be moved by these pore velocities?”

To begin incipient movement of a particle at rest in a fluid flow field, the flow must overcome the friction and other forces holding the particle in place. Recent literature indicates that for laminar flows (creeping motion) the critical shear stress, τ

c (M/L2), for initiation of particle motion ranges


69

from 0.12 ≤ τc ≤ 0.30. Additionally, the particle will be suspended and moved only if the critical

shear stress divided by the fluid density is greater than the particle settling velocity, or

τ

ρ

ρ ρ

μc

f

s fgd2

≥−( )

18

Table 1. Hypothetical saturated and unsaturated characteristics of a copper ore.

Characteristic Units Value

Dry bulk density g/cm3 1.65Particle density g/cm3 2.76Porosity percent 40.2K

sat cm/s 2.3 × 10–2

−200 Sieve fraction percent 16Leaching moisture (θ) cm3/cm3 0.268 @ 10 L/hr/m2

Leaching K(θ) cm/s 2.8 × 10–4

Leaching % saturation percent 66.7Leaching pore velocity cm/s 1.04 × 10–3

Figure 4. Calculated lixiviant spread for a copper-zinc ore.


70

where: ρs and ρ

f are the particle and fluid densities, respectively; g is gravitational acceleration, d

is particle diameter, and μ is dynamic viscosity. Solving for grain diameter, d, for reasonable fluid values and a particle density of 2.76 g/cm3 gives a particle diameter range of between 0.001 and 0.002 mm. These are the particle diameters, or less than these diameters, which will be transported by a pore velocity in the heap on the order of 1 × 10−3 cm/s.

These particle diameters are in the range of fine silt and clay. It is concluded that that portion of the leached ore with particle sizes less than or equal to the fine silt fraction could be transported within the heap as a result of a leaching rate of 10 L/hr/m2. Many caveats are required for this conclusion including how reactive the clay particles may be to the lixiviant, how much cohesion the clay has, and the fact that the analyses ignore the electrical forces which typically bind clays to other particles. While it may be possible to have “fines migration” due to ore degradation dur-ing leaching, this appears to be uncommon and unproven. Most likely, if fines are an issue in a given heap, they were most likely placed in the heap during loading and not as a result of massive fines migration during leaching. Documented cases of dispersive clay migrations, especially on the surface of a heap are known but usually these dispersions are localized, although they can be problematical, impeding flow and creating perched solution zones or impermeable lenses.

2.4 Myth 4 and reality 4: Any unsaturated heap will permit upward air flow for leaching

Myth 4: If air (oxygen) flow for leaching is a requirement, any degree of saturation less than fully saturated will allow upward air flow through the heap.

Reality 4: Heaps do not need to be saturated to restrict air movement. As little as 65 to 70 percent liquid saturation can eliminate air flow on both micro and macro scales.

Discussion: Many gold and copper heap leach processes are done under unsaturated conditions, because air is an important part of the metal recovery process. Often the operator does not inten-tionally plan or design for a specific air flow for a given leach rate. Sometimes air flow is limited by increased lixiviant flows which the operator believes will provide more recovery.

The flow of air and water in heaps is linked via the portion of the voids which are filled with each of these fluids. In fact, the heap does not have to be saturated to effectively eliminate air flow. Only enough of the pores need be filled with lixiviant to eliminate air pathways through the heap. Our experience based on copper and gold ores has been that if more than about 70 percent of the available void volume is filled with lixiviant, air flow is essentially ineffective. Thus, careful management of lixiviant and air is necessary to successfully leach ores where air is important in metal resource recovery.

A typical example of air/lixiviant flux rates versus liquid saturation is given for a copper ore as shown on Figure 5. Laboratory air and lixiviant flux rates (L3/T/L2) for that copper ore were a min-imum desired lixiviant flux of 3.0 × 10–3 and a minimum desired air flux desired of 1.2 × 10–2.

The lixiviant and air flux values shown on Figure 5 give an operating range of between approximately 60 and 75 percent of liquid saturation, which is within the desired lixiviant leach-ing rate. The conclusion here is that there exists a range of lixiviant application rates that allow for the desired air flux for this copper ore; indicating that copper recovery will proceed as anticipated.

If lixiviant application rates are outside the acceptable operating range, metal recovery will be reduced. This problem can occur on a localized level or through very large areas of a heap leach pad. Regardless of size, once an oxygen barrier is created the operator is at the mercy of oxygen solubility in the lixiviant for leaching kinetics. Again, this rarely matches the controlled conditions in the laboratory leach column, so the metallurgist is left wondering why things worked so well in the lab, but those results are not matched in the field.

2.5 Myth 5 and reality 5: A phreatic surface in the heap

Myth 5: Phreatic (water table) surfaces in heaps mean that the plastic liner has a high hydraulic head which could exacerbate leakage to the environment.


71

Figure 6. Schematic of actual capillary pressure with drains working and hydrostatic pressure.

Figure 5. Laboratory air and lixiviant flux versus degree of liquid saturation for a copper ore.


72

Reality 5: If the pipe drains are working and the pipes are not flowing full, the head on top of the plastic liner is approximately the same as the flow depth in the underdrain pipes.

Discussion: Just because there is a water table in a heap leach facility does not mean that there are unusually high heads or pressures on the plastic liner. In fact, as shown on Figure 6, if the drains are working and not flowing full, the head at the top of the drain pipe or in the gravel drain-age layer must be at atmospheric pressure in order to allow fluid to enter the pipe. The corollary to this is that there must be a small saturated zone around any pipe in an otherwise unsaturated media in order for fluid to enter the pipe if it is not flowing full. Therefore, hydrostatic heads typically do not occur or propagate from the phreatic surface in a heap to the underlying plastic liner.

3 CONCLUSIONS

In most heap leach pads the lixiviant flow velocity almost never occurs as saturated flow under normal operations and often is as film flow over the individual particles or as preferential, finger-ing flow through voids created by the solids.

Fines, that is, materials with median particle diameters less the 0.074 mm, are typically needed to have an effective heap leaching operation. Even then, heterogeneities in the leap materials may cause fingering flow and uneven wetting of the ore resulting in poor resource recovery.

Fines are almost never transported within a heap causing plugging due to low velocities.Air movement may be important in gold/silver and copper leach operations. Smaller lixiviant

leach rates may be required for adequate upward air movement in the heap in order to maintain acceptable resource recoveries.

Water tables in heaps do not indicate unusually high pressures on underlying plastic liners if the drain pipes are operating and not flowing full.

There are so many non-linear forces at work in a leach pad that rarely does classical linear mathematics describe actual practice.

REFERENCES

Hillel, D. & Baker, R.S. 1988. A descriptive theory of fingering during infiltration into layered soils. Soil Science vol 102: 135–140.

Jury, William. 2006. Pore and Darcy scale physics of preferential flow arising from fluid instability in homo-geneous soil. Workshop on Preferential Flow and Transport Processes in Soil. November 4–9, 2006. Ascona, Switzerland. http://www.ito.ethz.ch/conferences/preferential-flow/.

Lenormand, R., Touboul, E. & Zarcone, C. 1988. Numerical models and experiments on immiscible displace-ments in porous media. Journal of Fluid Mechanics vol. 189: 165–187.

Or, Dani. 2006. Scaling of capillary, gravity, and viscous forces affecting flow front morphology in unsatu-rated porous media. Workshop on Preferential Flow and Transport Processes in Soil. November 4–9, 2006, Ascona, Switzerland. http://www.ito.ethz.ch/conferences/preferential-flow/.

Philip, J.R. 1975 Stability analysis of infiltration. Soil Science Society of America Proceedings vol. 39: 1042–1049.

Raats, P.A.C. 1973. Unstable wetting fronts in uniform and nonuniform soils. Soil Science Society of America Proceedings vol. 37: 681–685.

Tindall, J.A. & Kunkel, J.R. 1999. Unsaturated zone hydrology for scientists and engineers. Upper Saddle River, NJ: Prentice Hall, Inc. 624p.

Ulrich, B.F. 2008. Geotechnical aspects of the hydro-jex operation. In Rock Dumps 2008 [A. Fourie (ed)]. Proceedings of the First International Seminar on the Management of Rock Dumps, Stockpiles and Heap Leach Pads, 5–6 March 2008. Australian Centre for Geomechanics. Perth, Australia, 47–55.

van Zyl, Dirk. 2008. Integrated heap leach design—incorporating unsaturated material considerations. In Rock Dumps 2008 [A. Fourie (ed)]. Proceedings of the First International Seminar on the Management of Rock Dumps, Stockpiles and Heap Leach Pads, 5–6 March 2008. Australian Centre for Geomechanics. Perth, Australia, 153–166.


73

Challenges in heap leach pad design: Consideration of thermal conditions

A.L. HudsonTetra Tech, Inc., Littleton, MA, USA

T. MeyerTetra Tech, Inc., Grand Junction, CO, USA

ABSTRACT: Heap leach pad design can be affected by many factors including material properties, ore grade, and climate. Often the climate component can be overlooked, but in the case of an extreme environment, the climate can create serious problems with pad operation. In an environment that experiences extreme temperature changes from summer to winter, the design must consider the pad thermal conditions to allow heap operation throughout the entire year. A heap leach pad in northern Mongolia will experience temperatures as low as −40ºC, which would cause the leach solution to freeze within the heap material and freezing of the pregnant leach solution pond. Therefore, year-round operation will require hearing the leach solution with a boiler and maximizing the heating capacity of solar radiation. Through a com-bination of variably saturated flow modeling, thermal solution modeling within the heap, and thermal modeling of the solution pond, mitigation measures were evaluated as part of the heap leach pad design. The boiler proved to be effective, but excessive heat loss from the pond will require the use of plastic Bird BallsTM on the pond surface to insulate the solution and to maxi-mize solar heating. This paper details the modeling activities and thermal considerations of this heap leach pad design project.

1 INTRODUCTION

This project was completed for a mine that is located in the Selenge Province of northern Mon-golia. The project is an operating gold mine that is in the process of expanding their operation to include heap leaching. The location of the mine offers many challenges for designing the facilities due to the extreme climate of the mine. The temperature fluctuations from summer to winter can be from 40ºC to −40ºC. The goal of the mine is to be able to operate throughout the entire year, utilizing a heap leach pad to extract the gold.

1.1 Project scope and objectives

The objective of this portion of the heap leach pad design was to provide design parameters for the heating of the leach solution to maintain leaching operations throughout the year. The design parameters were determined through a series of models of the heap leach facility and the pregnant leach solution (PLS) pond. The models utilized were:

− Variably saturated flow modeling of solution application;− Thermal modeling of the solution application; and− Thermal modeling of the PLS pond.



74

2 MODEL CONSTRUCTION—HEAP LEACH FACILITY

In VADOSE/W (GEO-SLOPE, 2004), a finite-element model grid was built that represents the constructed and loaded heap leach pad. A combined variably saturated and thermal model was constructed for two cross-sections of the heap leach pad. Two cross-sections were used because of the geometry of the heap. The heap will be placed on a slope, so there are areas that will be thinner and could be more susceptible to freezing than the thicker central portions of the heap. The two cross sections have been designated Cross-section A and Cross-section B. Figure 1 present the two cross-sections as modeled.

2.1 Conceptual model

Based on the design developed for this heap leach pad, the conceptual model is similar to other heap leach facilities. The water balance of the system consists of precipitation, evaporation, run-off, infiltration, and application of the leach solution. The one unique aspect of this facility and the associated variable saturated model is that the solution emitters are placed approximately 2.5 meters (m) below the top of the ore pile to prevent freezing during the winter. Figure 2 shows the conceptual model of the heap leach pad.

2.2 Modeling assumptions

As with any modeling of complex systems, some simplifying assumptions were necessary to com-plete the project. For this modeling effort, one of the key input parameters of the thermal modeling was the starting ore temperature of the heap. It was assumed that no ore would be placed on the heap during the winter months; however, leaching will occur during the winter months. For this reason, the starting ore temperature is assumed to be equal to the average air temperature for the end of September or early October (the end of ore placement for the year). It is also assumed that no ice lenses will form within the heap material because the material will not be loaded on the pad during the winter months when layers of ore might freeze during placement or when snow could be trapped within the heap.

Figure 1. Facility cross-sections and model construction.


75

The heating of the leach solution is assumed to be 5ºC above the temperature at which it enters the boiler. This corresponds to the heating capacity of the boiler, and provides a target for deter-mining if this will be a sufficient temperature gain. Two heating scenarios were considered in this modeling effort, a worst case and a typical operating case. As a worst-case scenario, it was also assumed that the solution starting temperature would be approximately 0ºC and heated to 5ºC. For a typical operating scenario, it was assumed that the solution was heated from approximately room temperature (20ºC) to 25ºC.

2.3 Modeling technique

The modeling software VADOSE/W (GEO-SLOPE International Ltd.[GEO-SLOPE], 2004), a two-dimensional saturated/unsaturated zone flow model that is part of the GeoTudio suite of programs, was used to simulate the flow of water and thermal characteristics throughout the heap leach facility. This is a commercially available piece of software that specializes in the simulation of infiltration along the surface of the model. The modeling program also has the following key advantages:

− The model can be used to simulate both water and thermal properties using a single software platform;

− The model can run under both steady state and transient conditions; and− The model can be run with a user define climate file with varying temperature, wind speed,

evaporation, and humidity.

2.3.1 Steady-state modelingSteady-state modeling is always challenging when analyzing mining facilities because operations are dynamic in nature and conditions do not typically reach “steady-state” conditions until after mine closure. To mitigate this problem, the modeling was only run under steady-state conditions to get the starting values for moisture content and ore temperature within the model. The results of the steady-state modeling are not designed to replicate operational conditions, just to offer non-zero starting values for the subsequent transient modeling simulations.

2.3.2 Transient modelingTransient modeling is the true simulation of flow and thermal conditions within the heap leach facility. The model starts with the surface region (see Section 2.4.2.1). It is in this part of the model

Figure 2. Conceptual model schematic.


76

that air conditions, precipitation, and soil come in contact; it is also the layer of the model that drives the water balance of the system. Next, the water moves according to the rules of unsaturated flow physics through the heap material until it reaches the liner (base of the heap leach pad struc-ture). At each element in the grid, moisture, water flux, and temperature are calculated.

2.4 Model input parameters

The following input parameters were incorporated into the VADOSE/W (GEO-SLOPE, 2004) modeling:

− Site climate data;− Solution application rate;− Current heap leach facility design plans; and− Unsaturated flow parameters for the ore material on the heap.

2.4.1 Material propertiesThe most important input parameters are the physical properties of the ore placed on the heap and the ground surface under the heap leach pad facility. These parameters control the flow of water, air, and heat through the heap leach pad. The following sections discuss the specific properties used to simulate the heap and the underlying foundation soils.

2.4.1.1 Foundation soilsThe foundation soils were modeled using a saturated hydraulic conductivity of 10−13 centimeters per second (cm/sec). This simulates the plastic liner that will be placed under the heap leach ore pile. Even though the liner is a very thin component of the entire facility, the unit was simulated as if it were the entire ground material unit. This allowed the heap to be simulated accurately as a lined facility, but also improved the overall model stability.

2.4.1.2 Ore materialThe ore material was determined to have a saturated hydraulic conductivity of 10−2 cm/sec or 8.46 meters per day (m/day). This is equivalent to a uniform sand material which is comparable to the expected grind of the material prior to placement on the heap. The saturated volumetric water content of material will be 35%. In addition, the thermal conductivity of the ore material was determined to be 9.83 kilojoules per day per meter per degree centigrade (kJ/day/m/C) with a specific heat equal to 1.32 × 103 kilojoules per cubic meter (kJ/m3).

2.4.2 Boundary conditionsThe next most significant input for the model simulations is the application of the boundary con-ditions. The boundary conditions necessary for this modeling were limited to the application of leach solution and the application of the climate data. The heap leaching operations involves the application of a combination of solutions to the heap surface for controlled infiltration and leach-ing of the ore. The application rate that will be used is 0.005 gallons per minute per square foot (gpm/ft2). The solution will be applied to the heap using a 60 day leaching cycle (45 days of solu-tion application and 15 days of drain-down). A boundary condition function was developed within the model to simulate this leaching cycle. Because the leach solution emitters will be placed 2.5 m below the surface of the heap, the boundary condition representing the solution application was also applied at that depth below the model surface. The climate boundary condition was applied to the surface of the model and is discussed in more detail in the following sections.

2.4.2.1 Surface regionThe surface of a mining facility, such as a heap leach pad, is a critical area to consider when simu-lating how the facility will interact and be impacted by the surrounding environment. VADOSE/W (GEO-SLOPE, 2004) rigorously simulates the dynamics of the heap surface by considering cli-mate and soil interactions through the use of a defined surface region. Precipitation is simulated


77

by dividing each day of the transient model into increments with a maximum step size of two hours. The daily climate data is applied according to a sinusoidal function that peaks at noon.

Actual evaporation of the system water balance is calculated by the model based on the climate parameters:

− Air temperature;− Soil temperature;− Relative humidity;− Solar intensity (from latitude);− Soil moisture content;− Wind speed; and− Measured pan evaporation.

The combination of all of the factors listed above gives a very accurate estimate of actual water evaporation from soil. Infiltration is based on the unsaturated hydraulic conductivity and the mois-ture content of the material at a given time. Excess precipitation that is not evaporated or does not infiltrate is tabulated as runoff or surface snowpack (depending on average air temperature).

2.4.2.2 ClimatologyOne of the advantages to using the program VADOSE/W (GEO-SLOPE, 2004) over other similar programs is the application of site-specific climate data to the model. The following parameters were included as part of the climate data file used and applied to the surface region of the model:

− Minimum and maximum daily temperature;− Daily precipitation;− Minimum and maximum daily humidity;− Daily measured evaporation; and− Average daily wind speed.

Climate data from the Baruunkharaa meteorological station was used in the modeling. The data spans a record of over 30 years. This station is located approximately 19 kilometers (km) north of the site at an elevation of 810 m.

In general, the climate at the mine is characterized by long cold winters and short hot summers. Winter air temperatures can reach −40ºC and summer temperatures can reach 40ºC. The average monthly temperatures range from −24.5ºC in January to 18.3ºC in July. Temperature data from the Baruunkharaa meteorological station for the period of 1961 to 1990 is presented in Table 1.

Table 1. Monthly temperature data from the Baruunkharaa meteorological station.

Average Maximum Minimum temperature temperature temperatureMonth ºC ºC ºC

January −24.5 1.9 −45.7February −21.2 7.8 −43.7March −8.3 20.6 −37.7April 2.5 29.8 −23.1May 10.8 36.0 −9.8June 16.5 37.0 −6.7July 18.3 38.8 −8.7August 16.3 36.4 −2.3September 9.2 30.1 −10.2October 0.5 27.0 −24.5November −11.1 14.1 −37.5December −20.8 8.4 −42.8


78

3 MODEL CONSTRUCTION—PLS POND

A thermal model of the PLS pond was completed to determine how fast the water in the pond would freeze if the pond was stagnant. The pond may become thermally stagnant if there is a problem with the boilers or the pumping system. Under these conditions, the pond could lose heat rapidly and freeze, severely limiting operations for the remainder of the winter.

3.1 Conceptual model

It is assumed that the conditions of the pond are steady relating to the volume and the flow in and out of the pond. The thermal modeling of the pond was developed as a heat budget for the system. The types of parameters that were considered in this modeling were:

− Net radiation effects;− Heat loss due to evaporation; and− Heat transfer to the atmosphere.

The heat budget equation used to model the PLS pond is presented as Equation 1:

dH/dt = ΘR + Θ

E + Θ

L + Θ

adv + Θ

B (1)

where dH/dt = rate of heat change; ΘR = net radiation; Θ

E = latent heat of exchange; Θ

L = sensible

heat exchange; Θadv

= net advective exchange; and ΘB = conduction through sediments.

By modeling the PLS pond using a heat budget, the sources of heat loss and increase could easily be considered, and engineering controls considered to control the greatest sources of heat loss.

3.2 Modeling assumptions

For the modeling of the PLS pond, it was assumed that the primary engineering control that would be used to prevent heat loss from the pond would be plastic Bird BallsTM. Bird BallsTM are a proven tech-nology that has been apply for similar purposes, and will be the most cost effective solution to year round operations. The Bird BallsTM are assumed to be a complete, single layer of coverage for the pond surface. This cover layer has significant advantages in such an extreme climate. The single layer will cover approximately 91% of the total water surface. In addition, the surface evaporation is decreased by 90% and the freezing point of the solution will be lowered by 10ºC. The decreased evaporation is also an operational advantage for this mine during the summer. With potential summer tempers in excess of 30ºC, the solution will not be subject to the same rate of evaporation as it would be without the cover. For comparison and to justify the added expense of the engineering control, a model was also complete for the water surface without the Bird BallsTM. (Nelson Environmental, Inc., 2008)

4 COMBINED HEAP AND PLS POND THERMAL MODELS

The variably saturated/thermal model of the heap leach pad and the heat budget model of the PLS pond were used in combination to optimize the design of the heap leach pad and to define the required operational conditions. The variably saturated/thermal model defined the heat loss that is expected to occur within the heap material during leaching and provided a starting temperature for the PLS pond. The heat budget model defines the expected heat loss while the solution is exposed to the extreme climatic conditions.

4.1 Steady-state model

As mentioned above, state-state modeling of mining facilities is always challenging due to the rapid changes in the facility conditions. To mitigate this problem, the steady-state models were only run


79

to get non-zero starting values for the heap moisture content and ore temperature. Because the final stacking scenario of the heap was modeled, the conditions targeted were an ore temperature of approximately 3ºC, and a moisture content in the range of 6% to 12%. These values correspond to the average air temperature during the late September and early October time period, and the drain-down and the active leaching moisture contents, respectively.

4.2 Transient model

Transient modeling is the true simulation of flow and thermal conditions within the facilities. The top of the model is a surface region. It is the part of the model that air, precipitation, and soil come in contact. At a depth of 2.5 m below the surface of the heap are the emitters for the solu-tions application, which are the primary drivers for the system’s water and heat balance. Next, the water moves according to the rules of unsaturated flow physics through the heap material. At each element in the grid, the water flux and temperature are calculated. The resulting water temperature modeled was then used in the heat budget calculations for the pond to determine an overall heat loss from the system.

4.3 Modeling results

The results of the variably saturated and coupled thermal modeling showed that the heating of the solution helped to maintain the heat near the emitters, even with near freezing starting ore temper-atures and an average winter climate. The surface of the heap is frozen during the winter months, but at a depth of 2.5 m there appears to be sufficient heat to prevent the emitters from being affected by the harsh climate. It should be noted that the heat added to the solution is lost within the ore. Under both the typical conditions (heating from 20ºC to 25ºC) and the worst case condi-tions (heating from 0ºC to 5ºC), the solution lost between 3ºC and 5ºC within the heap material.

Under the typical conditions, the solution temperature exiting the heap is sufficient to prevent freezing in the PLS pond. For this model, it was assumed that the solution exiting the heap will be at approximately 20ºC (the maximum heat loss resulting within the heap for the typical conditions

Table 2. Results of typical conditions model for PLS pond with Bird BallsTM.

WaterAmbient air surface Theta Theta Theta dH Time totemperature temperature R E L dt lose 2ºCKelvin Kelvin W/m2 W/m2 W/m2 cal/m2 · sec minutes

253.3 293.15 −80 5.1 3965 930 35.8253.3 291.15 −77 5.1 3766 883 37.8253.3 289.15 −74 5.1 3567 836 39.9253.3 287.15 −71 5.1 3368 789 42.2253.3 285.15 −69 5.1 3169 742 44.9253.3 283.15 −66 5.1 2970 695 47.9253.3 281.15 −64 5.1 2771 648 51.4253.3 279.15 −61 5.1 2572 601 55.4253.3 277.15 −59 5.1 2373 554 60.1253.3 275.15 −56 5.1 2174 507 65.7253.3 273.15 −54 5.1 1975 460 72.4253.3 271.15 −52 5.1 1776 413 80.6253.3 269.15 −49 5.1 1577 366 91.0253.3 267.15 −47 5.1 1378 319 104.4253.3 265.15 −45 5.1 1179 272 122.4253.3 263.15 −43 5.1 980 225 148.0


80

simulation). Based on the results of the heat budget model, if the pumps were to be turned off, the upper meter of the pond will begin to freeze in approximately 18 hours. This is based on the pond having a single layer of Bird BallsTM on the surface. If the Bird BallsTM are not used, the first meter of the pond will freeze in approximately 1.5 hours. The results of this modeling are presented in Table 2 (PLS pond with Bird BallsTM) and Table 3 (PLS pond without Bird BallsTM).

Under the worst case conditions, the solution temperature exiting the heap is sufficient to pre-vent freezing in the PLS pond. For this model it was assumed that the solution exiting the heap will be approximately 2ºC. Based on the results of the heat budget model, if the pumps were to be turned off, the upper meter of the pond will begin to freeze in approximately 11 hours. This is

Table 3. Results of typical conditions model for PLS pond without Bird BallsTM.


253.3 293.15 −157 51 23790 5661 5.9253.3 291.15 −146 51 22596 5378 6.2253.3 289.15 −135 51 21402 5095 6.5253.3 287.15 −124 51 20208 4812 6.9253.3 285.15 −114 51 19014 4529 7.4253.3 283.15 −104 51 17820 4246 7.8253.3 281.15 −94 51 16626 3963 8.4253.3 279.15 −84 51 15432 3680 9.1253.3 277.15 −75 51 14238 3397 9.8253.3 275.15 −66 51 13044 3114 10.7253.3 273.15 −57 51 11850 2831 11.8

Table 4. Results of worst case conditions model for PLS pond with Bird BallsTM.


253.3 275.15 −56 5.1 2174 507 65.7253.3 273.15 −54 5.1 1975 460 72.4253.3 271.15 −52 5.1 1776 413 80.6253.3 269.15 −49 5.1 1577 366 91.0253.3 267.15 −47 5.1 1378 319 104.4253.3 265.15 −45 5.1 1179 272 122.4253.3 263.15 −43 5.1 980 225 148.0

Table 5. Results of worst case conditions model for PLS pond without Bird BallsTM.

WaterAmbient air Surface Theta Theta Theta dH Time totemperature temperature R E L dt lose 2ºCKelvin Kelvin W/m2 W/m2 W/m2 cal/m2 · sec minutes

253.3 275.15 −66 51 13044 3114 10.7253.3 273.15 −57 51 11850 2831 11.8


81

based on the pond having a single layer of Bird BallsTM on the surface. If the Bird BallsTM are not used, the first meter of the pond will freeze in approximately half an hour. Even with these less than ideal conditions, it appears that there will be sufficient heat to maintain operations if the pond does not remain stagnant for a long period of time. The results of this modeling are presented in Table 4 (PLS pond with Bird BallsTM) and Table 5 (PLS pond without Bird BallsTM).

5 CONCLUSIONS

It is critical that the solution be as warm as possible when applied to the heap, but the heat that is added by the boiler is expected to be completely lost before the solution exits the heap. If the solution exiting the heap is 20ºC, then the pond will not freeze unless the pumps are turned off for a period of 18 hours with an air temperature of −20ºC. If the solution temperature is approxi-mately 2ºC when it leaves the heap and the air temperature is −20ºC, then the time before freezing is reduced to 11 hours. The pond will be susceptible to freezing should a problem occur with the pumps. For this reason, the pumps and a backup system are the most critical components of the leaching system for successful winter operations.

As discussed, the Bird BallsTM are a critical component of the system. The Bird BallsTM decrease the freezing point of the solution and protect the water surface from the wind and heat loss affects of evaporation. All of these factors combine to increase the time that the water can be stagnant and still not freeze.

This modeling only considered a single cycle through the system. Cumulative cooling impacts were not considered, but could impact the long term operation of the system. If too much heat is lost throughout the system, and it cannot be recovered through the use of a boiler, the time before the pond begins to freeze will be decreased. This is particularly important for the worst case conditions.

REFERENCES

GEO-SLOPE International Ltd. 2004. Vadose Zone Modeling with VADOSE/W: An Engineering Methodol-ogy. Alberta:GEO-SLOPE.

Nelson Environmental, Inc. 2008. Bird BallTM Cover System. http://www.nelsonenvironmental.com/TechProd6_BirdBalls/. Information verified 14 July 2008.



83


Innovative mine waste disposal in two distinctly different settings

I. Wislesky & A. LiGolder Associates Ltd., Mississauga, Ontario, Canada

ABSTRACT: Innovation within the mining industry has always been important as a result of worldwide pressure to improve their practice. The tarnished image, primarily a result of environ-mental legacies and some more recent events, has resulted in a concentrated world focus on all mining ventures. Financial institutions have developed the Equator Principles which mines must follow to obtain project financing. In today’s heavily scrutinized mining environment, project fund-ing will not get off the ground unless both environmental and sustainability issues are considered and accommodated. Mining companies are currently becoming proactive in their pursuit of new approaches to develop and achieve sustainable goals in line with the goals of the community.

Innovation in management of mine waste materials is one of the most important aspects to consider for a mining project to advance and, in fact, this has come a long way since the days of finding a nearby lake or river to dump their wastes into. Important factors include physical and chemical stability over the short and long term as well as the size of the area of influence.

This paper discusses the use of careful understanding and planning for the disposal and man-agement of mine wastes in two distinctly different mine settings. By utilizing this methodology and current technology, innovative, unique designs were developed to optimize the disposal of mine waste materials in stable and cost effective systems which minimize the environmental effects footprint. Waste disposal at both of these mines involves co-disposal of mine waste rock and tailings to reduce the space required for disposal, as well as to limit the effects of potentially acid generating materials.

Co-disposal is being considered more often these days since it can enhance flexibility of a waste disposal system and provide both economically and environmentally viable solutions, as well as improved social benefits. At a mine site, several streams of wastes are produced including tailings, waste rock, slag, etc. Co-disposal allows for the optimization of disposal of the various materials by taking advantage of the individual properties of each of the materials, whether it is low perme-ability, buffering, strength, etc., and combining these to produce desired overall characteristics.

The use of co-disposal and thickening technology, in consideration of the overall characteristics of the mine sites and the waste material characteristics has created the opportunity for optimiza-tion which provides benefit to both the mining company and the local communities.

1 INTRODUCTION

This paper discusses the development of two mines in very different areas of the world where innovative approaches to site development, in particular, waste management, were required to ena-ble mine development to proceed. The use of thickened tailings, co-disposal and co-mingling are the common links considered in the design of the waste management systems at these two mines. Consideration of the individual characteristics of each site was necessary to optimize design and allow development. One of the mines (Shakespeare, owned by URSA Major Minerals (URSA)) is located in a rugged region of Canada subjected to seasonal extremes in temperature and the other (Cerro De Maimon, owned by GlobeStar Mining Corporation (GlobeStar)) is located in a tropical region of the Dominican Republic. The locations of these mine sites are shown on Figure 1.


84

New and innovative approaches to the design of mine waste management systems are neces-sary in today’s mining climate for a number of reasons. Mining companies recognize the need for change. As a result, the mining industry is actively seeking measures to achieve sustainability and social responsibility primarily through sound environmental practices. There are many drivers for change within the mining industry including the following:

− Financial—Mine developments must follow the Equator Principals to obtain adequate project financing.

− Regulatory—There are laws and regulations in place with regards to land use, closure, etc.− Risk—System failures can lead to extensive environmental, human and financial losses.− Reputation—The reputation of the mining industry as a whole requires some mending as a

result of legacies as well as more recent events.− Public Awareness—Lobbying by NGOs and media scrutiny increases the exposure of the

industry.

The management of mine wastes is considered by many to be the most important area where improvement can be made to enhance the viability for the development of new mine sites. In effect, mining companies can be considered to be primarily in the waste management business (generally

Figure 1. Location map of the Shakespeare and Cerro de Maimon projects.


85

only a small fraction of the material moved becomes product and the rest becomes waste). It is impor-tant to develop a complete understanding of the site and the characteristics of the waste materials including: climate, topography, natural hazards, environmental/social constraints, regulatory require-ments, tailings characteristics, waste rock characteristics, quantity and schedule of wastes, etc.

The key design drivers for mine waste management are:

− Chemical Stability—Acid generation and leachable contaminants (metals).− Contaminant Transport—Groundwater, surface water, air.− Surface and Groundwater Protection—Downstream use.− Water Management—Dry/wet/winter climatic conditions.− Erosion Stability—wind and water.− Physical Stability—static and dynamic.− Aesthetics—Primarily at closure.− Minimize Footprint—Area of impact.

There are several waste disposal options that have been used or are currently being developed and these include:

− Slurry deposition,− Stacks,− Co-disposal and/or co-mingling;− Underwater disposal, and− Disposal in mined out open pits or as underground backfill.

The actual methodology used at an individual mine site can be any of the above but the choice should be made based on sound engineering principles in consideration of all aspects of the site conditions, surrounding area and waste characteristics. The methodologies used for waste disposal at the two sites considered in this paper involve a combination of thickened tailings, co-disposal and co-mingling.

2 THICKENED TAILINGS

Thickening technology has come a long way over the past several years with deep cone thick-eners and improved distribution systems. This technology is currently proven and acceptable. Thickening produces a dewatered, non-segregating material that can still be pumped and piped to a disposal site. Thickening technology was primarily developed for underground mine backfill operations and is more recently being used for surface disposal of tailings. Dr. Eli Robinsky pio-neered the thickening process in 1968 and designed the first surface disposal site for tailings in 1973 (Robinsky, 1999). This site is still being used for thickened tailings deposition today. Thick-ening provides the following benefits over conventional slurry deposition:

− Greatly reduces risk by limiting or eliminating retained water ponds,− Particle segregation does not occur, producing a denser, less permeable tailings mass,− Greater chemical stability by inhibiting the ingress of water and oxygen,− Conserves water (very important in dry climates) and reduces water management

requirements,− Accelerated consolidation providing accessibility to foot and equipment faster than slurry,− Facilitates progressive closure,− Better control of wind and water erosion,− Can use tailings for underground backfill,− Reduces the need for large dams,− Smaller basin footprint,− Less seepage and groundwater contamination, and− Reduces or eliminates the need for a liner.


86

Not all of these benefits may apply to a particular facility, however, if a sufficient number of these benefits do apply, it may be worthwhile to pursue this methodology.

3 CO-DISPOSAL

Co-disposal consists of placing different waste materials into a common facility. Although co-disposal on surface is not commonly used in the mining industry, there are many examples where both tailings and mine rock are disposed of together. There are tailings areas where mine rock has been used as a primary material for construction of tailings containment dams. In addition, mine rock is often used for the construction of internal access roads or berms within tailings facilities. There are also numerous examples of mine sites where other materials are stored within the tail-ings facility and/or used for internal construction purposes.

Co-disposal takes these examples and extends them into an engineered facility where tail-ings and mine rock are placed together in an efficiently engineered, environmentally sound and economic system. Some advantages of co-disposal include:

− Easier to manage the disposal in one facility;− Smaller footprint and area of disturbance;− Reduced dam construction;− Eliminates/reduces the need for a liner;− Extends the life of an existing tailings basin (stacking, placement angle, in-situ density);− Reduced issues with respect to seepage and evaporation losses;− Reduced wind and water erosion, reduced infrastructure requirements (roads, pipelines,

pumps, etc.);− Better control of acid generation through efficient mixing (tighter matrix);− Reduced water management issues (single point of discharge to the environment);− No ponded water on top of the co-disposed materials (hence less risk);− Facilitates progressive closure and reduced closure issues (smaller area, one water management

system and one treatment plant, easier to monitor, etc.); and− More accessible to foot traffic and equipment.

4 CO-MINGLING

Waste rock generally has quite a high void ratio which translates into space availability for other materials. In conventional waste dumps, this space is filled with water and air which creates perfect conditions for acid generation and metal leaching. Co-mingling utilizes tailings to fill the voids within the waste rock matrix. The benefits of this symbiotic system are:

− Reduced storage volume requirements (reduced footprint);− Reduced potential for water and oxygen ingress and contact with potentially acid generating

materials (chemical stability); and− A strong, physically stable structure can be produced.

To provide adequate mixing of the waste rock and tailings and to optimize filling of the voids, thickening of the tailings to a non-segregating material is required. Mr. B. Wickland and Dr. W. Wilson, have studied the effects of co-mingling tailings and waste rock (Wickland, 2006) and further studies are necessary to develop procedures to optimize mixing.

5 SHAKESPEARE PROJECT

The Shakespeare site is located in Canada, west of Sudbury, Ontario, on the northern shore of Agnew Lake (Figure 1). The project is an 11.3 Mt, nickel ore body that will be mined in two


87

open pits. The ore will be processed on site at a nominal rate of 1,642,500 t/y, or approximately 4,500 t/d, with a mine life of about seven years. In general the site consists of the two open pits, a mill, a thickener plant, the co-disposal site and several ponds for water management. The general site layout is shown on Figure 2.

Previous to Golder’s involvement, a preliminary design was developed for conventional slurry tailings disposal at a location about seven kilometres north east from the mine. However, in order to reduce the area of disturbance, environmental impact and cost, URSA retained Golder to design a cost effective, fully functional facility which included the possibility of co-disposing thickened tailings and mine rock in an area close to the mine with good topographic containment.

The location of the mine on a prominent, rocky ridge, extending as a peninsula into Agnew Lake, limits the potential sites for tailings and mine rock disposal close to the mine. The site that was originally selected for mine rock disposal was the only potential site close to the mine with good topographic containment that was not an existing lake. Though not big enough for slurry tailings disposal, a co-disposal strategy was considered to increase the containment volume without major dam construction. The site is valley shaped with the north and south sides rising up to 35 m above the base. This topographic containment enhances stability of the dams and the existing relatively impervious clayey foundation soils will inhibit seepage. Space is also available downstream for water management (settling and polishing ponds). The use of the co-disposal technique reduces water management infrastructure by confining the mine wastes within one catchment area. The footprint of the waste management area has been reduced from the initially planned 150 ha for tailings alone to approximately 90 ha for both mine rock and tailings. There will be four primary streams of waste material produced as a result of mining and milling that require disposal at the Shakespeare Project Site. These materials are described as follows:

− Acid generating mine rock (1.6 M-m3);− Non-Acid generating mine rock (27.73 M-m3);− Acid generating pyrrhotite tailings (high sulphide content) (0.57 M-m3); and− Thickened tailings (very low sulphide content) (5.95 M-m3).

Figure 2. Topography and site layout.


88

The design methodology consists of a single co-disposal facility where all the above materials will be disposed for several years from the beginning of mine development. In the later years, the acid generating mine rock and pyrrhotite tailings will be disposed of in the mined out, West Open Pit. A system of dams will be required downstream of the co-disposal site for water con-trol including settlement of solids, retention time, flooding of acid generating materials, extreme precipitation and other miscellaneous sources of water. The thickener plant will be located above and adjacent to the co-disposal site to take advantage of gravity for distribution of the thickened tailings (thickener underflow) to the co-disposal site and thickener overflow directly to the down-stream water treatment pond.

The unique disposal strategy (filling plan) developed provides containment of the four mate-rial types within a single reduced footprint while preventing contamination of clean waste with acid generating waste and maximizing the available space. The filling plan, which requires a good understanding of the mine development and material scheduling, involves dividing the co-disposal site into two areas to facilitate deposition of the various materials. The low central valley section will be used primarily for acid generating materials (acid generating rock and high sulphur tailings). A dam will be constructed across the western end of the co-disposal area to promote flooding of the valley and to create a pond over the acid generating materials (to prevent acid generation over the long term). The remaining area will be used for non-acid generating materi-als (i.e., clean rock and pyrrhotite reduced tailings). The high sulphur (pyrrhotite) tailings will be transported by pipeline, in slurry form, to the lowest section of the co-disposal area and discharged

Figure 3. Filling sequence—plan view.


89

into a subaqueous environment. The non acid generating tailings will be thickened and disposed of together with the non acid generating mine rock.

As discussed above, the acid generating materials will be placed into a flooded central basin with the acid generating mine rock placed at the west end near the main dam and the slurry tailings placed, starting from the east end. The acid generating rock will be end dumped into the water to create a rockfill platform about 1 m above the water level. Once deposition of the acid generat-ing materials is complete, the water level will be raised to permanently submerge the area and prevent acid generation. This central section will remain open until the later years of the operation as shown on Figures 3 and 4, when it will be completely covered over with co-mingled, non acid generating tailings and waste rock. As shown in Figures 3 and 4, co-mingled, non acid generating tailings and waste rock will initially be placed south of the central, low valley as open pit devel-opment scheduling permits and then gradually cover over the entire co-disposal area. As a result, although the acid generating materials will remain saturated, there will be no pond that could lead to concerns over future exposure of acid generating materials.

For volume estimating purposes (capacity of the co-disposal site), the placed mine rock was assumed to have 30% porosity (void space) and only 50% of this available space would be filled with thickened tailings. In other words, a significant additional capacity exists within the void space of the rock to store tailings without increasing the size of the co-disposal area. Additional storage capacity will become available for potentially acid generating materials in the West Pit once it is mined out. The acid generating material can be placed underwater within this pit and will remain submerged when closed. As can be seen, advanced knowledge of scheduling of the open pit development and production of the various waste materials is critical to the effective opera-tion of this facility. Overall management and scheduling of mine waste disposal activities is quite

Figure 4. Filling sequence—cross-sections.


90

important for this site but the co-disposal scheme developed in consideration of this, will enable this mine to proceed with development.

6 CERRO DE MAIMON PROJECT

The Cerro de Maimon Mine is located about 75 km northwest of Santo Domingo in the Domini-can Republic (Figure 1). The project involves mining and milling oxide ores and sulphide ores bearing gold, copper, silver and zinc minerals with a maximum anticipated production rate of about 2,500 tpd.

The site is situated within a tropical, mountainous region with an average annual precipitation and pan evaporation of approximately 2,012 mm and 1,710 mm, respectively. The annual precipitation for the 100 year dry and wet return periods was estimated to be 1,059 mm and 3,760 mm respectively.

Waste management is challenging due to the quantity and geochemical characteristics of the tailings and waste rock and the limited space available at the site.

Geochemical characterization for waste rock, ores and two streams of oxide and sulphide tail-ings were conducted to identify potential environmental impacts from these materials. In summary, the sulphide ore, footwall waste rock and all the tailings were determined to be acid generating. Some separation of the acid generating waste rock from the non-acid generating was considered possible which would enable development of separate disposal sites. The materials requiring on-site storage include:

− 14 M tonnes of inert overburden waste,− 4.8 M tonnes of potentially acid generating (PAG) waste rock,− 27.2 M tonnes of non-acid generating waste rock, and− 5.2 M tonnes of potentially acid generating tailings.

The main design drivers for waste management for this project are:

− The geochemical characteristics of the tailings and waste rock and preventive measures for ARD management;

− Space availability (minimize footprint of the waste management facility and maximize storage capacity); and

− Overall water management.

To enable mine development in consideration of the site constraints, waste materials and envi-ronmental concerns, the following features were considered in the design of the waste manage-ment system:

− Thickening the tailings prior to disposal to reduce the area required for disposal of poten-tially acid generating waste materials, allow for progressive closure and to simplify site water management.

− Using one site for disposal of both acid generating waste rock and tailings (co-disposal).− Using a co-mingling process to combine placement of both acid generating waste rock and tail-

ings to limit seepage, limit oxygen ingress and reduce the total volume required for disposal.− Constructing several connected storage cells for the co-disposal area to permit operational flex-

ibility and allow for progressive closure.− Using locally excavated clay to provide a low permeability compacted clay liner to minimize

seepage escaping from the co-disposal site.− Constructing a water collection/treatment pond downstream of the co-disposal area to collect all

potentially contaminated water sources including seepage or excess water which might accumu-late in the co-disposal cells.

− Providing a clay cover followed by additional non-acid generating mine waste rock on top of filled cells to limit ingress of water and oxygen and provide space for additional waste materials.


91

Figure 5 shows a plan view of the entire mine site during the first and final years of opera-tion. The co-disposal facility is located to the north side of the open pit and the east side of the existing mountain ridge with the seepage and runoff collection pond at the downstream of the perimeter dam. Thickener overflow will bypass the co-disposal site and discharged into the downstream pond. This will greatly reduce the water management efforts required within the co-disposal site.

The co-disposal facility was designed to accommodate the required storage capacity over the life of the mine for the acid generating tailings and waste rock. The overburden soils and non-PAG waste rock will be hauled to different waste piles within the property boundaries. The co-disposal facility involves three cells for tailings and PAG waste rock which are to be contained by low per-meability perimeter dams and internal dykes, raised progressively, as required. The entire footprint of the tailings cells will be lined with compacted clay using a local colluvium clayey deposit.

The three tailings cells have been designed in such way that each cell can be covered with soil immediately after its storage capacity is fully utilized. The soil cover is to be constructed at a 3% slope outwards to prevent the ingress of water and oxygen by promoting runoff, and hence inhibit oxidation. Subsequently, the non acid generating waste rock will be deposited on top of the tail-ings cell, while tailings are discharged to other cells. The non acid generating waste rock cover will function as an additional oxygen barrier for the underlying acid generating tailings and waste rock. The co-disposal site will accommodate approximately 48% of the total non-PAG waste rock, which would otherwise require additional areas for storage. The concurrent closure of the tailings cells during operation requires careful scheduling based on the production of waste materials dur-ing the operation.

To co-mingle tailings and waste rock, a portion of the acidic waste rock will be directly depos-ited into the thickened tailings in the early stages of deposition for each cell. The majority of the PAG waste rock will be co-mingled with tailings during the final stage of the operation for each cell to provide a competent foundation for the materials placed above (i.e. the non-PAG waste rock). It is assumed that, on average, only 75% of the waste rock voids can be filled with tailings. Figure 6 shows a schematic section of the co-disposal site at the final stage of operation. The perimeter dam will be 55 m in height with a downstream toe berm to provide a suitable factor of safety for both short term and long term stability. The dam will be constructed in stages using a downstream raise method. The side slopes of the non-PAG waste rock pile (i.e. Pile 4) on the tailings cells are designed to provide sufficient factors of safety against failure of the overall slope

Figure 5. Plan view of the co-disposal facility in the first and last year of operation (Cerro de Maimon Project).


92

Figure 6. Cross section of the co-disposal facility in the last year of operation (Cerro de Maimon Project).

Photo 1. Starter dam and tailings cells under construction.

Photo 2. Tailings cell 1 with compacted clay liner.


93

(120 m in height) under both static and seismic conditions. For closure, a topsoil cover and vegeta-tion will be applied to the surface of the entire co-disposal facility.

Photo 1 shows the starter tailings cells 1 and 2 and the perimeter dam during construction. The dam shell is constructed with non-PAG waste rock from the initial open pit development. Photo 2 shows the compacted clay liner in the tailings cells prior to tailings placement.

7 CONCLUSIONS

Innovative disposal methodologies for mine wastes, along with significant advancement in the production of thickened tailings, has provided a means for improved efficiencies with respect to waste and water management at mine sites around the globe. The results can be measured with reduction in costs, environmental impacts and social impacts from mining projects. Sustainable solutions, global financial pressures and risk to both corporations and future generations are key areas where the importance of upfront and thorough evaluations of each mine site, is clearly shown to be necessary for the survival of the mining industry. As a result, mining companies are utilizing innovative alternatives for mine development and, in particular, mine waste management strategies. The two projects presented in this paper have clearly demonstrated a path forward that can benefit existing and future mine developments globally.

The sites described in this paper were areas where mining activities may not have proceeded a few years ago. Companies who design waste management facilities have the responsibility to develop a complete understanding of the site, the waste characteristics, the local regulations, the social concerns of the local population and sustainability options for the area. Although the sites described in this paper used thickened tailings and co-disposal, all disposal options should be con-sidered during the initial mine site study stages to determine the optimum mine waste and water management strategies for the site. Slurry tailings facilities and open mine waste dumps, along with their inherent physical and chemical stability issues, must not be considered as the only pos-sible solution. There are viable alternatives that can and should be explored.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the forward thinking approaches of the management of both URSA Major Minerals and GlobeStar Mining Inc. who allowed and encouraged innova-tion in the development of their respective mine sites. Messrs. Richard Sutcliffe of URSA and J.P. Chauvin of GlobeStar, through their involvement in these projects are enabling the advance-ment of mine waste management technology and the promotion of sustainability in the mining industry.

REFERENCES

Robinsky, E.I. 1999. Thickened Tailings Disposal in the Mining Industry. Published by E.I. Robinsky Associ-ates Limited, Toronto, Canada.

Wickland, B.E., Wilson, G.W., Wijewickreme, D & Klein, B. 2006. Design and evaluation of mixtures of mine waste rock and tailings.



95

Pipeline design for paste and thickened tailings systems

R. CookePaterson & Cooke, Denver, CO, USA

ABSTRACT: This paper discusses design considerations for paste and thickened tailings pipe-line systems. The pipeline flow behavior of conventional tailings, thickened tailings and paste are reviewed. It is noted that there is uncertainty regarding the design of thickened tailings pipelines in laminar flow. Guidelines are presented for the use of centrifugal and positive displacement pumps. The paper concludes with the recommendation that life cycle trade-off studies should be conducted to establish the optimum economic solids concentration for each proposed system.

1 INTRODUCTION

There has been a significant shift towards high concentration backfill and tailings systems over the last two decades:

− Paste backfill technology has been developed for underground mines as an alternate to low concentration hydraulic fill. While research into paste technology was conducted at a number of operations internationally, Dave Landriault was instrumental in the wide scale implementa-tion of paste technology through his work in Canadian Mines (Landriault 2006). Paste backfill technology is now mature although significant operational problems are often experienced with the pipeline distribution systems (Cooke 2007).

− In arid regions (primarily Australia, southern Africa and Chile) there is strong drive to increase the concentration of tailings to reduce metallurgical plant water consumption.

− Thickened and paste tailings impoundments are considered to offer increased stability and reduced environmental impact.

This paper examines design considerations for paste and thickened tailings pipeline transporta-tion systems.

2 PASTE AND THICKENED TAILINGS CLASSIFICATION

The following classification is proposed for tailings and backfill mixtures (Cooke 2006):

− The upper limit for conventional tailings is considered to correspond to the freely settled pack-ing concentration. This typically corresponds to yield stresses of between 5 and 20 Pa.

− Thickened tailings are considered to cover the range from the freely settled concentration to the concentration at which the mixture has a fully sheared yield stress corresponding to 100 Pa. Figure 1 illustrates the slump of a mixture with a 100 Pa yield stress.

− Paste tailings and backfill are considered to be mixtures with yield stresses greater than 100 Pa. The practical upper limit for pipeline transport is about 800 Pa.



96

3 PIPELINE FLOW BEHAVIOR

Figure 2 compares the pressure gradient versus pipeline velocity characteristics for water, conven-tional tailings, thickened tailings and paste.

3.1 Conventional tailings

Conventional tailings systems are operated in turbulent flow with a heterogeneous distribution of solids within the pipeline. The pipeline diameter is selected to ensure that the operating pipeline velocity is always greater than the deposition velocity (i.e. the incipient velocity at which particles deposit on the pipe invert under turbulent flow conditions.

3.2 Thickened tailings

Referring to Figure 3, the pressure gradient characteristic curve for thickened tailings has two distinct zones:

− At low velocities the flow behavior is dominated by the mixture rheology and the flow is said to be laminar. As can be seen from the figure, the pressure gradient in laminar flow is relatively insensitive to changes in pipeline velocity.

− At high velocities, the inertial forces dominate and the flow is turbulent.

The Bingham plastic model is suitable for characterizing most mineral slurries. If the mixture can be assumed to be homogenous, the laminar flow pressure gradients can be estimated using the Buckingham equation (assuming the rheology is known). However due to the wide particle size distribution of most tailings, the flow will not remain homogenous in laminar flow with the con-centration and percentage of coarse particles increasing towards the pipe invert. This results in two significant areas of uncertainty for the design of thickened tailings pipelines in laminar flow:

− There is a risk that particles may settle on the pipeline invert with resulting flow instabilities (Cooke 2002).

− Due to the non-homogeneity of the slurry within the pipeline, the Buckingham equation is not appropriate for scaling up or predicting the pipeline the pressure gradients. This is an area of ongoing research.

This presents a design challenge as high velocities are required to produce turbulent flow (a regime where suitable methodologies are available for predicting the pipeline gradients) resulting

Iron ore tailings

64%m, 100 Pa yield stress

Figure 1. Slump test for tailings with 100 Pa yield stress.


97

Figure 3. Thickened tailings pressure gradient characteristic.

Figure 2. Pipeline pressure gradient versus velocity characteristics.

3Velocity (m/s)

Pip

elin

e pr

essu

re g

radi

ent

Water

Conventional Tailings

Thickened Tailings

Paste

3Velocity (m/s)

Pip

elin

e pr

essu

re g

radi

ent

Water

Laminar flow

Thickened Tailings

Laminar - tubulent transition

Turbulent flow


98

in excessive energy consumption. Gillies et. al. (2007) have proposed a “tentative criterion” for estimating if a stationary deposit will form under laminar flow conditions. But even for flow without a deposit there is no established methodology for predicting pressure gradients for non-homogenous laminar flows. The current approach for designing these systems is to design for laminar flow, but include provision for operating in turbulent flow (at least intermittently) to flush any accumulated solids out of the pipeline. This is done by operating at high velocities, introduc-ing a small percentage of dilution water (to reduce the laminar-turbulent transition velocity) or a combination of these approaches.

3.3 Paste

For paste mixtures, turbulent flow operation is not feasible due to the high velocities required to achieve turbulence. The following considerations should be applied to the design of paste pipelines:

− Experience indicates that deposition is unlikely to occur provided the pipeline pressure gradient is greater that 2 kPa/m.

− The Buckingham equation can be used to scale pipeline pressure gradients provided the test pipe size is large enough that slip does not occur.

− The pipeline pressure gradients are sensitive to pipe diameter changes due to wear. This has important considerations for underground backfill systems where changes in the pipeline fric-tion characteristic can result in extensive slack flow in piping systems.

4 PUMPS

All pipeline systems need an energy source to overcome the pipeline friction losses. Backfill sys-tems primarily use gravity, but are often also equipped with pumps on surface to control the flow rate into the pipeline.

4.1 Centrifugal pumps

Centrifugal pumps are the industry workhorse for tailings and other slurry applications. While centrifugal slurry pumps can handle viscous slurries (as shown in Figure 4) with yield stresses greater than 250 Pa, for reliable operation it suggested that the maximum yield stress is limited to 200 Pa. In addition, the pump should be selected to operate as close to the best efficiency point as possible.

Centrifugal pumps typically have flat head (or pressure) versus flow rates characteristics to the left of the best efficiency point. As thickened and paste tailings also have a flat pipeline character-istic (as shown in Figure 1), it is important that care is taken to ensure a stable operating point.

Due to the significantly increased friction losses in the suction piping, care must be taken to ensure that the pump’s net positive suction head requirements are exceeded to avoid cavitation and unreliable pump operation.

The pressure generated by a centrifugal pump is directly related to the density of the medium in the pump. Accordingly most thickened and paste tailings systems require pressurized flush water into the suction of the first stage pump to flush the pipeline.

4.2 Positive displacement pumps

Two types of positive displacement pumps are typically used for paste applications:

− Hydraulically actuated piston pumps are used for paste backfill applications where the paste yield stresses are generally in the range of 250 to 500 Pa.

− Piston diaphragm pumps are used for higher flow rate paste tailings applications with yield stresses in the range of 100 to 250 Pa.


99

Positive displacement pumps must be provided with a positive suction pressure to overcome the suction valve friction losses and to properly fill the chamber on each stroke. A pressure greater than 40 psi is typically required.

5 CONCLUSION

Engineering design considerations for paste and thickened tailings pipeline systems have been presented. Generally the complexity and costs of these pump and pipeline systems increase

Figure 4. Centrifugal pumps can handle viscous thickened tailings as shown in this photograph.

Figure 5. Paste tailings piston diaphragm pump installation.


100

exponentially with increasing tailings concentration. Before implementing a paste and thickened tailings system, a life cycle trade-off study should be conducted to establish the optimum economic solids concentration.

REFERENCES

Cooke, R. 2002. Laminar flow settling: the potential for unexpected problems. Proc. 15th intern. conf. hydrau-lic transport of solids, Banff, June 2002.

Cooke, R. 2007. Backfill pipeline distribution systems—design methology review. In F. Hassani & J. Archibald (eds), Minefill 2007; Proc. 9th intern.symp. of mining with backfill, Montreal, 29 April–2 May 2007.

Gillies, R.G., Sun, R., Sanders, R.S. & Schaan, J. Lowered expectations: the impact of yield stress on sand transport in laminar, non-Newtonian slurry flows, Proc. 17th intern. conf. hydraulic transport of solids, Cape Town, 7–11 May 2002.

Landriault, D. 2006. They said “It will never work”—25 years of paste backfill 1981–2006. In R. Jewell, S. Lawson & P. Newman (eds), Paste 2006; Proc. 9th intern. seminar on paste and thickened tailings, Limerick, 3–7 April 2006. Australian Centre for Geomechnics.


101

Efficient dewatering solutions on vibrating screens

M. DoerfferUniversity of Mining and Technology, Freiberg, Germany

R. HeinrichW.S. Tyler Canada Ltd., Ontario, Canada

ABSTRACT: Any wet applications in bulk material processes ultimately raise the question of dewatering. Usually the purpose of dewatering is to put the solid material into an either sellable or at least transport-ready state. A variety of machinery options exist for that purpose. Each of these come with there very own range of usage and there pros and cons. This paper will evaluate a range of processing equipment for its dewatering capabilities and operating parameters. Special focus will be set on environmental aspects as energy and water con-sumption and the effects the performance of dewatering equipment has on the entire process. Vibrating screens have been proven to be an economical means for a multitude of dewatering applications ranging from aggregates over coal to food. Operating conditions that affect the dewa-tering performance will be discussed. These parameters are not limited to the mechanical behavior of the screen but include material properties and overall process conditions. In detail the effect of the following parameters will be discussed:

• Feed material properties• Layer of material• Screen media, openings, arrangement• Overall screen dimensions and inclination• Operating parameters like speed and stroke.

Field- and laboratory tests will be discussed and evaluated against there theoretical background. Knowing the influence that above parameters have on the overall dewatering process will allow the operator to tune his vibrating screen to its best performance. In closing dewatering machinery options will be presented based on vibrating screens and on combinations of screens and hydro cyclones.

1 INTRODUCTION

Any wet applications in bulk material processes ultimately raise the question of dewatering. Usu-ally the purpose of dewatering is to put the solid material into an either sellable or at least transport-ready state. A variety of machinery options exist for that purpose. Each of these come with there very own range of usage and there pros and cons.

This paper will evaluate a range of processing equipment for its dewatering capabilities and operating parameters. Special focus will be set on environmental aspects as energy and water con-sumption and the effects the performance of dewatering equipment has on the entire process.

Vibrating screens have been proven to be an economical means for a multitude of dewatering applications ranging from aggregates over coal to food. Operating conditions that affect the dewa-tering performance will be discussed. These parameters are not limited to the mechanical behavior of the screen but include material properties and overall process conditions.



102

2 BACKGROUND

Dewatering is the process to separate a mixture of solids and liquids. During this process none of the components will be altered. If the target of the process is the solids we speak about dewatering. In that case we look for a solids product with a liquid content as low as possible.

There are other applications that look at the liquid as the final product with as little as possible solids in it which won’t be discussed here.

The term “Dewatering” can be further limited to materials smaller 2 ½''. Solid mixtures of that and larger sizes do not have a substantial capacity to carry water.

In order to dewater a bulk material we have to look at the means of how water or liquid is bound to solids (figure).

Table 1. Process classification by remaining water content [1].

Process Remaining water content [%]

Thickening >40Pre-Dewatering <40Dewatering >10Drying >0

Figure 2. Dewatering stages in bulk materials [2].

Figure 1. Water in bulk materials [2].

a) Inner liquidb) Adsorbed liquidc) Adhensive liquidd) Small capilare liquide) Free liquidf) Large capilars liquid


103

Looking at the process of dewatering one can see three basic states of the material:The first stage shows a material fully saturated with water. The entire porosity volume is filled

with water. In this stage the mixture of solids and liquid behaves mostly like a liquid.During dewatering the second stage will occur. In this stage we will find fully water-

filled porosities as well tight solid-solid contacts allowing only adhesive and small capillary water.

The final stage of dewatering is reached when the solids are in the most compacted stage allow-ing only adhesive and small capillary water.

3 MAIN DEWATERING PRINCIPLES

3.1 Sedimentation

Sedimentation uses either gravity or some means of centrifugal force to separate solids from liquid. The focus is usually not only set on the solids material but also on the cleanliness of the liquids. As previously shown the water content of the solids achievable by sedimentation is rather high as naturally the mixture is still fully saturated. Industrial use of sedimentation is found in thickeners.

3.2 Filtration

For this paper filtration is of greater importance. During filtration a means of filter media is used that allows water or other liquids to leave the mixture. The solids will form some sort of a filter cake on top of the filter media. When this cake is established it acts as a filter media in itself and can retain particles that are smaller then the porosity or openings of the filter media. If that is the case we speak of cake-filtration.

The process of cake filtration can be described in 3 phases:

– Creation of the cake. Water drains through filter media.– Compaction of the cake. Water in porosities in the cake is replaced by air.– Final stage. Remaining water content is in balance and cannot be reduced further.

Figure 3. Cake filtration [3].


104

4 MATERIAL PROPERTIES FOR DEWATERING

It is of utmost importance to evaluate all material properties relevant to dewatering before selecting machinery.

4.1 Feed Material Properties

4.1.1 GradationThe single most important material property is its gradation. The gradation defines how much water can be bound in the product. It also defines how small the pores or openings of the filter media have to be.

4.1.2 DensityThe greater the difference between the density of solids and liquid the easier is the separation.

4.1.3 Particle shapeThe shape of a particle not only defines the surface area and therefore its capacity of binding adhesive water but also the resistance against liquid flow in a material.

4.1.4 OtherOther properties like firmness, abrasiveness, toxicity, product value have to be taken into account as well. During the process particles could get destroyed and increase the fines content hence increasing the overall surface area and hindering the dewatering.

Corrosiveness and abrasiveness will play a large role in the selection of the machinery.The value of the product will determine how much sense it makes to dewater to a certain

percentage.

4.2 Finished product—remaining water content

The one product property that comes to mind when speaking about dewatering is the remain-ing water content. This property mainly defines the entire process. It also affects the following processes as the water content will determine whether the material can be transported on a belt conveyer.

4.3 Process and machinery conditions

Before selecting dewatering machinery one should consider the following conditions:

• Allowable gradation• Allowable solids percentage• Throughput• Remaining water content• Efficiency• Economy: investment and operation cost.

5 DEWATERING BY SEDIMENTATION

The following gives a brief overview of machinery and principles using sedimentation for dewatering.


105

5.1 Gravity as separating force

It quickly becomes obvious that none of above principles deliver a “dry” product. Exception to that is stockpile–dewatering which on the other hand takes a rather long time.

When time is of the essence it is worthwhile to enhance gravitational forces by centrifugation.

5.2 Centrifugation as separating force

From the data above it becomes clear that only Hydro Cyclone and Decanter are capable to dewa-ter a relatively wide size range with decent throughputs.

6 DEWATERING BY FILTRATION

The following table compares the major machinery groups using filtration by gravity.The vibrating screen presents the widest application range for in-line processing and comes in

at relatively low remaining water percentages.A variety of other machinery options are available that use a combination of filtration and cen-

trifugal forces. Very common are units applying over- or under-pressure to improve the dewatering efficiency and time. In this area we find all the filter presses.

Table 2. Dewatering by sedimentation.

Feed solids Remaining water DewateringMachinery Size range percentage [Vol. %] content [Wt. %] time

Clarifyer 0.1–0.5 mm <15 20–50 0.5–24 hrs.Thickener 0.1–0.3 mm <10 20–50 0.5–24 hrs.Stockpile > 0.1 <50 5–10 0.5–24 hrs.Sand screw 0.15–63 mm <50 22–25 30–60 sec.

Table 3. Dewatering by centrifugal forces.

Feed solids Remaining percentage Throughput water content Dewatering timeMachinery Size range [Vol.%] [m3/h] [Wt.%] [sec.]

Hydro cyclone .005–0.5 mm 2–30 0.2–650 20–30 <30Centrifuge 0.0001–0.1 mm 0.005–3 <4 3–6 <30Separator 0.0001–0.1 mm 0.002–6 <200 3–6 <30Decanter 0.001–5 mm 4–40 <200 3–6 <30

Table 4. Filtration by gravity.

Feed solids Remaining percentage Throughput water content Dewatering timeMachinery Size range [Vol. %] [m3/h] [Wt. %] [sec.]

Silo 0.063–2 mm 50–90 – 4–8 8–10 hrsVibrating screen 0.063–63 mm 50–80 35–80 10–20 10–30 secFine sand recovery wheel 0.063–63 mm 20–50 20–50 20–25 10–30 sec


106

Figure 4. Dewatering systems by size range.

Figure 5. Throughput by dewatering system.

0 10 20 30 40 50 60

Fine Sand Recovery Wheel

Vibrating Screen

Stationary Screen

Silo

Decanter

Seperator

Centrifuge

Hydro Cyclone

Stock Pile

Clarifyer

Dew

ate

rin

g S

yste

m

Size Range [mm]

0

100

200

300

400

500

600

700


Vibrating Screen

Stationary Screen

Silo

Decanter

Seperator

Centrifuge

Hydro Cyclone

Stock Pile

Clarifyer

De

wa

teri

ng

Sy

ste

m

Throughput [m3/h]


107

Figure 6. Feed solids content by dewatering system.

0 10 20 30 40 50 60 70 80 90


Vibrating Screen

Stationary Screen

Silo

Decanter

Seperator

Centrifuge

Hydro Cyclone

Stock Pile

Clarifyer

De

wa

teri

ng

Sy

ste

m

Feed Solid Content [%]

0 10 20 30 40 50


Vibrating Screen

Stationary Screen

Silo

Decanter

Seperator

Centrifuge

Hydro Cyclone

Stock Pile

Clarifyer

De

wa

teri

ng

Sy

ste

m

Remaining Moisture [%]

Figure 7. Remaining moisture content by dewatering system.


108

7 SYSTEM COMPARISON

7.1 Feed size range

The size range that can be fed to the individual dewatering systems is shown below. It can be seen that only screens (stationary or vibrating) and of course the stockpile cover wide range of sizes. Other units mostly focus on the fine end of the scale.

7.2 Throughput

Looking at the achievable throughput hydro–cyclones and fine-sand-recovery-wheels come out first for small size while vibrating screens clearly dominate in the mid-size range.

7.3 Feed solids content

An important aspect is the condition of the feed or it’s solid to water ratio. A number of dewatering systems like clarifier, separator and centrifuge clearly focus on the liquid side of the scale while again screens cover a wide range.

7.4 Remaining water content

The remaining water content is usually the scale to be used to evaluate dewatering options. Some of the systems deliver exceptional values that all come with a downside like long dewatering times or very high machinery costs. Others are more focused on the liquid end of things like the clarifier. Screens (vibrating or stationary) are in mid range.

7.5 Conclusion

It can be said that from all options presented the vibrating screens are the most versatile. Espe-cially in the application range of dewatering of bulk materials in mining, sand and gravel or recy-cling they are second to none.

Figure 8. Dewatering on inclined screens.


109

8 DEWATERING ON VIBRATING SCREENS

8.1 Inclined screens

The material is usually introduced as slurry. On an inclined dewatering screen the accelerations along with a portion of the gravitational force will cause the material to travel towards the dis-charge end while the water is being screened out by means of proper screening media (figure).

The resulting force for material travel is indicated as the yellow arrow. The force causing the water to separate from the solids is gravity enhanced by the vertical vector of the g-forces pro-duced by the screen.

The problem in this design is that the gravitational forces are only partially used towards dewa-tering but also towards material transport. The later will ultimately drag water into the product.

In order to improve dewatering one would have to decrease the inclination of the screen which in return will also decrease the material travel rate drastically. In other words the “dry” product will end up quite wet.

This can raise a couple of problems. A customer could be quite hesitant to pay for a high amount of water in the product. A product with high water content will flow back on a conveyer belt and will cause severe damage to the rollers of the belt by washing fine solid product into the bearings. Finally the amount of water in the product is a loss if the customer doesn’t pay for it just as well as it is a loss if the plant runs a closed water circuit.

8.2 Horizontal screens

The horizontal screens used for dewatering are actually not exactly set at 0º. It has proven to be very beneficial to set them up at a negative incline of about 3º.

As for the applications discussed above the material is usually introduced as slurry. Other then on inclined machines the only force resulting in material travel is the g-force produced by the screen. This g-force is aligned a 45º and transports the material uphill. The gravitational forces enhanced by the vertical portion of the acceleration of the machine are fully utilized towards dewatering.

In operation those forces will build a material layer on the screen that is pushed out of the wet zone towards the discharge end. A back dewatering field is used to reduce the amount of water right after feeding the screen. The thick layer of material acts as a filter cake and not only presses water out but traps fine particles that would be lost in a thin layer screening process.

Furthermore a dam at the discharge end is normally used to further enhance product quality.The main advantages of that technology can be summarized as follows:Large angle between material transport and gravity –> Good separationFilter-cake bridges openings and “traps” fine particles –> Minimal material lossWater won’t run uphill + Filter-cake presses water out –> Excellent dewatering.

Figure 9. Dewatering on horizontal screens.


110

Figure 10. TYCAN XL-Class 8 × 24.

Figure 11. Dewatered material.


111

8.3 Field data

The following describes field data of an 8 × 24 Tycan XL-Class.The machine is being fed at a rate of 300 tph solids and 750 USGPM of water. The transport

speed of material over the screen is 0.19 m/s.The resulting product water content ranges from 10 to 15%, which delivers a belt conveyable

product (figure). It can be seen how the material breaks of at the end of the discharge as a very well dewatered product whereas the last visible water is about 2 ft away from that point.

REFERENCES

Böhringer, Paul; Höffl, Karl: Baustoffe wiederaufbereiten und verwerten, AVS-Institut GmbH-Verlag 82008 Unterhaching, S. 299–301.

Schubert, Heinrich: Aufbereitung fester mineralischer Rohstoff, Band III: Flüssigkeitsabtren nung, 2. völ-lig neu bearbeitete und erweiterte Auflage, Leipzig, Deutscher Verlag für Grundstoffindustrie, 1984, S. 64–66.

Schubert, Heinrich: Aufbereitung fester mineralischer Rohstoff, Band III: Flüssigkeitsabtren nung, 2. völ-lig neu bearbeitete und erweiterte Auflage Leipzig, Deutscher Verlag für Grundstoffindustrie, 1984, S. 64–66.



113

High pressure washing technology Hydro-Clean

S. PalomboW.S. Tyler Canada, Ontario, Canada

J. VarelaHaver & Boecker Latinoamericana Ltda., Monte Mor, Brazil

ABSTRACT: Washing can be defined as the process used to eliminate or reduce the dirt (pres-ence of fines or slimes) of one material. It is also the basic and oldest process applied to ore prepa-ration for further concentration. This process, in combination with commnution and screening, form the so called washing plant and in many cases produce end products. This work describes the factors that influence the performance of the washing process and main washing technologies used in the mineral processing industry. The Hydro-Clean represents, since 1998, a state of the art technology in this area and is com-prised of one of the most innovative solutions developed at HAVER & BOECKER. HAVER & BOECKER design and manufacture technology for mining and related industries. Today the Hydro-Clean is manufactured and sold throughout the Haver Screening Group (HSG) which is comprised of HAVER & BOECKER Maschinenfabrik Münster, HAVER & BOECKER Latinoamericana—Brazil and W.S. Tyler Canada. The Hydro-Clean is a washing technology based on utilization of high pressure water jets. To follow we will discuss how the Hydro-Clean’s different modes of operation result in several advantages related with mine waste reduction, recovery of saleable product, lower capital costs, superior washing results, less stress for the material, a wider range of application and less mainte-nance cost. Aside from these positive factors, its lower operation costs related to the consumption of energy and water are presented. It is important to note that the quantities, tonnages, data and management concepts presented in the paper are accurate for the time of writing and are subject to change and modification as operations progress.

1 INTRODUCTION

A United Nations Educational, Scientific and Cultural Organization (UNESCO) estimate warns that by 2050, between 2 billion people in 48 countries and 7 billion people in 60 countries will face water shortage problems (UNESCO, 2008). Prevention is the best way of reducing water and electrical power shortages. In Brazil, the law 9.433/97 defines the utilization and payment of water from rivers. Since 2007 the mining companies have agreed to pay for the use of river water from the Paraíba do Sul Basin, one of the most important in São Paulo State. In the case of the rivers from Piracicaba, Capivari and Jundiaí Basins, the taxation for the mining sector began in 2006 (IBRAM, 2007). In the U.S., states are also now taking water consumption and preservation more seriously as is evident in Colorado’s new law to protect water from contamination associated with in situ uranium mining. Recently, “Gov. Bill Ritter signed House Bill 1161—a measure Fort Collins lawmaker John Kefalas and mining executive Richard Clement hailed as one of the tough-est in the country. The law, which goes into effect July 1, requires mining companies to restore water to previous quality or to state standards.” Pamela Dickman (2008).



114

The principal uses of water are human consumption, transport, electrical energy generation, recreation, irrigation (agricultural purpose) and in several industrial applications. The biggest Brazilian mining company, Companhia Vale do Rio Doce (CVRD), uses water in all of its opera-tions. The main water consumption is in the mineral processing plant which reaches 137 million m3/annum.

The aim of mineral processing plants is to prepare products for further industrial applica-tions. All ore types are composed by particles of different size distribution and composition. Both characteristics are the main control parameters in mineral processing plants. Size dis-tribution and the liberation factor define the preparation stages and consequently the mineral concentration processes. For this reason, a mineral characterization program is necessary before any processing route evaluation, as described in Fernandes et al. (2007), Luz et al. (2004) and Hoeffl (1993).

Normally, one washing stage is present in all mineral processing plants, except when the mate-rial needs to be dry processed. However, wet processes are associated with higher costs for water management, sludge treatment and reject disposal. Other items to consider when implementing a wet process are freshwater availability and wastewater discharge according to local environmental protection legislation.

In fact, when very fine fractions can not be efficiently removed by dry process, a wet process is the only feasible method for satisfying the stringent product quality specifications. Therefore to achieve greater cost effectiveness, water and energy costs need to be reduced. For these reasons the Hydro-Clean technology comes to provide one of the best solutions to water management in the mining industry.

2 IMPORTANCE OF SIZE FRACTION FOR THE PROCESS

Particle size distribution has an influence in many process areas, particularly in the washing stage. Table 1 illustrates how fine material can be classified for further efficient processing. The table below provides an idea of the agglomeration degree of sand, silt and clay fractions.

One example of fines contamination, in terms of deleterious matter, is observed in building aggregate products. For example silt and clay size fractions can represent agglomerated particles present among or adhering to the surface of coarse particles, According to technical specifications such as DIN 4226 (1983), requirements in terms of size fraction are used to guarantee concrete quality. Krellmann and Hoppe (2000) report, however, that in many cases customers demand further reduction in the amounts of fine particles allowed as contamination in their aggregate product.

Theoretically, all material within a class can be processed with similar operation conditions however past practice has demonstrated the necessity to perform laboratory tests such as the wash ability test developed by HAVER. Alternatively, on-site test work (see figure 1) can be done to determine the applicability of the equipment.

Depending on a project’s objective, key factors determine the type and capacity of wash-ing equipment. Such factors relate to material characteristics or equipment aspects and process parameters.

Table 1. Classification of fines material.

Fraction description Fraction size (μm) Remarks

Sand 60–2000 material feels rough and coarse when rubbed by hand; it is difficult to form a sphere and it very easily falls apartSilt 2–60 material feels very soft, smooth and silky but it is very fragile and makes one’s hand very dirtyClay <2 material feels extremely fine, with plastic behavior


115

3 WASHING TECHNOLOGIES

The main application of the so called washers and their purpose in mineral processing plants is to remove adherent material from the surface of mineral particles and or break up agglomerated material. The mechanical energy present at traditional washing processes comprises of friction,

Figure 1. Hydro-Clean pilot plant.

Figure 2. Screw and log washer.


116

impact and shear forces. The main washing technologies available on the market are screw wash-ers (widely applied to classification), log washers and scrubbers, their modes of operation are briefly described below.

The screw and the log washer are basically tubs with a shaft inside them, (see figure 2). The screw washer has a fly like a screw mounted on the shaft, whereas the log washer has paddles mounted to it’s shaft. Both are installed at an angle, so that the material, which is filled in at the lower end, is transported to the upper end. During this process, water constantly flows in the tub, solves the dirt and carries it away. The solving of the dirt is mainly provided by the scrubbing action supplied by the paddles and screw.

The scrubber is a horizontally installed drum, see (figure 3), with shelves on its liners. These shelves lift the material which after reaching a certain height causes a cascading on itself, scrubbing and breaking dirt away. Water is constantly filled into the scrubber to solve the dirt and carry it out of the drum. A scrubber deploys more shearing action to the material than a log or screw washer.

4 HYDROCLEAN TECHNOLOGY

The Hydro-Clean is a completely new washing technology for the mineral processing industry. The first application of the Hydro-Clean was for washing aggregate materials. Aside from this conventional application, today there are units in operation within the recycling industry (build-ing gravel) and minerals industry (diamonds, gold, limestone and gypsum). Other industries have demonstrated interest in this technology, such as iron ore, bauxite, nickel, kaolin, phosphate, coal, emeralds, glass and plastic recycling.

The newly developed Hydro-Clean is a high-pressure washing system. It can be used for economi-cal and environmentally friendly cleaning of sticky clay, soil and other impurities from raw material with a size fraction of 0 to 150 mm. The water pressure to be adjusted at the equipment can reach up to 200 bar with a water and energy consumption, respectively, between 6–42 m3/h and 10–265 kWh. The intense water pressure and hydraulic force is determined beforehand and, in most cases, lies in the range of 60–180 bar. It’s continuous mode of operation can operate with a feed rate up to 400 t/h.

4.1 Technical principle

The Hydro-Clean consists of a vertical washing drum, which has a feed hopper mounted on one side and a discharge conveyor belt on the other side (see figure 4). The washing chamber, the

Figure 3. Scrubber.


117

Figure 4. Main components of the Hydro-Clean.

central element of the Hydro-Clean, consists of an upright cylinder which is lined with poly-urethane panels. The washing chamber contains a rotor, which is mounted on its top side and consists of several water nozzles. Some other special features of the Hydro-Clean include the variable height adjustment at the rotary wash head, the discharge belt and controlled feeding system that can be optimally adjusted to the bulk material and can yield outstanding cleaning results.

The Hydro-Clean mode of operation begins with the material being continuously fed by a conveyor belt into the feed hopper (see figure 5). Small water jets are mounted on the side walls of the hopper, which create a low pressure downstream current, which helps the material, particu-larly sticky material, to flow into the washing chamber. The height of the material in the hopper is constantly monitored by a laser level indicator. From there the material passes a slide gate into the washing container and forms a column of material.

The material is cleaned by being exposed to high pressure streams of water that come from the washing rotor and spray nozzle combination, located on the top third of the cylinder. The nozzles are adjusted into the movement direction of the rotor, so that the water, distributed to the reactor, is plowing through the material creating a shoveling effect. The cleaning process is assisted by the friction and shear forces resulting from the material movement around the chamber in a vortex (see figure 6).

The liberated fines material and process water are discharged from the washing chamber through the openings in the polyurethane panels at the side and collected by a waste water pipe (located behind the chamber) and is sent to water treatment or further screening. The washed


118

Figure 6. Washing rotor and spray nozzles.

Figure 5. Contaminated gravel being transported to the Hydro-Clean feed hopper.

material passes through the run-out hopper onto the variable speed discharge conveyor and is sent to a washing screen where the dis-agglomerated contamination is rinsed off (see figure 7).

These steps make the Hydro-Clean different from all washing technologies available in the market. Furthermore, the Hydro-Clean’s ability to incorporate automation through advanced plc control makes it the most technological advancement in the washing market today.

Rotor

Water Jets

“Shoveling Effect”Angled water injectionStd PU Panel Lining


119

4.2 Technical data and sepcifications

Currently the Hydro-Clean is available in four different models, respectively, HC 350, HC 700, HC 1000, HC 2000. The type numbers derive from the diameter of the drum in mm (See figures 8 & 9). Mobile and semi-mobile plants are also available both for production and testing applications.

Figure 7. Dis-Agglomerated material exiting the washing drum and entering the final rinse stage at the horizontal vibrating screen.

Figure 8. Hydro-Clean model data sheet.

Type HC 350 HC 700 HC 1000 HC 2000 T (Twin)

Capacity up to 20 t/h up to 100 t/h up to 200 t/h up to 400 t/hPressure up to 2900 psi up to 2900 psi up to 2900 psi up to 2900 psiElectrical PWR 54 HP 34–79 H P 69–158 HP 136–299 HPWater Req. GPM 27 35–46 70–92 140–184Length 5 ft 8.5 ft 10.5 ft 10.5 ftWidth 3 ft 8 ft 9 ft 12 ftHeight 7 ft 10 ft 11 ft 11.5 ftWeight (Empty) 6944 lbs 12460 lbs 17167 lbs 23283 lbs


120

5 COMPARISON WITH OTHER WASHING PROCESSES

In several plants the washing stage is performed on wet classifying screens but it is well known that the use of spray systems on vibrating screens at 3–5 bars, has a limited effect for sticky mate-rial. This practice can only be effective if the deleterious particles are easily removed (fractions between 60–2000 μm). Figures 10 and 11 compare two processing plants, the first one without Hydro-Clean and the second one with Hydro-Clean. It can be observed that for the production of 80 t/h producing the same 5 size fractions the water consumption can be reduced from 140 m3/h to 60 m3/h.

Table 2 shows a specifications comparison between some washing technologies. This table indicates a specific water consumption of 0.45–1.7 m3/h per ton of material for the screw. For the Log Washer and Scrubber this consumption can reach 0.95 and 1.45 m3/h per ton, respectively, and 0.2 m3/h per ton for the Hydro-Clean. The specific energy consumption of the washers indi-cate that the Hydro-Clean has the smallest value (0.3–0.5 kWh/t).

Table 3, compares the Log Washer (36′′ × 35′) with the Hydro-Clean (1000/140), not only is the energy and water consumption compared but also the maximum particle size that can be processed and the washing retention times. The table reveals the high pressure technology of the Hydro-Clean is the superior of the two in all categories. The better performance with regards to capacity is also evident and further outlines the water consumption benefit. A water consumption up to 115 m³/h for the Log washer 36′′ × 35′ and up to 18 m³/h for the Hydro-Clean (1000/140) is shown. These values indicate a clear advantage, in terms of water savings, when considering the Hydro-Clean for a mineral processing plant.

The better washing results of Hydro-Clean can be explained as a function of efficient power use in the process as related to power density (E) applied in the process. The power density index is a relation between the power used and the volume of the washer (kW/m³) as shown in equation (1).

E =

Installed Power (kW)

Machine Volume (m )3

(1)

Figure 12 illustrates the energy density index of the some washing technologies available in the market. The highest energy density index of Hydro-Clean is a characteristic for this high pressure washer.

Figure 9. Hydro-Clean Model HC 1000.


121

Figure 10. Flow sheet of a washing plant without Hydro-Clean.

Figure 11. Flow sheet of a washing plant with Hydro-Clean.

Table 2. Specifications of alternative washing technologies.

Max. Power Specific water Max. Empty capacity consumption consumption grain size weightEquipment (t/h) (kWh/t) (m3/h per ton) (mm) (t)

Screw Washer 1100 0.4–2.25 0.45–1.7 100 2.73–18Log Washer 190 0.15–2.2 0.1–0.95 140 15–32.5Scrubber 1700 0.01–1.5 0.2–1.45 500 16–68Hydro-Clean 400 0.3–0.5 0.1–0.2 150 1.5–15.7

Table 3. Comparison between Log Washer 36′′ × 35′ and Hydro-Clean 1000/140.

Log Washer 36′′ × 35′ Hydro-Clean 1000/140

Capacity (t/h) 50–125 50–200Preferred particle size (mm) 4.7 × 50 0 × 100Water consumption (m3/h) 11–115 18Installed power (kW) 112 90Retention time (min) 3 <0.5


122

6 “GREEN” BENEFITS

With the ever increasing difficulty for mining companies to acquire mine permitting for water use continued development by mining equipment manufacturers is necessary to help ease the proc-ess in permitting acquisition and help preserve our valuable environment. In addition to deliver-ing unprecedented degrees of cleanliness and some major technical advantages the Hydro-Clean offers the following green benefits:

• Water Consumption—Water is one of the world’s most precious resources. Only 0.52 gpm to 1.32 gpm of water is needed to clean 1 ton of dirt or clay contaminated material. Thanks to the Hydro-Clean’s robust filtering technology the water can be re-circulated through a client settling system and requires the addition of only 10% fresh water. This translates into water consumption savings of up to 75% compared to traditional log washer technology. This not only saves our resources and dramatically cuts the operation expenses, but makes permitting for new plants or plant expansions a lot easier.

• Turning Waste Piles to Sellable Product—Due to its unparalleled cleaning abilities, the Hydro- Clean transforms the customer’s waste pile, which has an impact on the required land resources into a sellable material. Not only does this aid the environment but directly translates from the cost side of the income statement to the sales side, translating into direct bottom line profits. See figure 13.

• Spare Parts—Since the only “tool” the Hydro-Clean utilizes is water, the wear and tear to the unit is minimal. Application dependant the drum lining needs to be replaced once a year and from time to time the nozzles require exchange. Minimizing the use of spare parts leads to less scrap, less logistical costs and ultimately lower maintenance expenses.

• Ambient Integration—To reduce the impact on the environment it is necessary to streamline systems and make them smaller and more efficient. The standard Hydro-Clean weighs only a total of 8 tons compared to comparable cleaning equipment that can weigh more than 3 times as much. This allows the customer to opt for a much smaller and leaner building structure, which reduces the area of influence on the environment. In addition, it significantly reduces construc-tion costs.

Figure 12. Energy density index for main washing technologies.

Washing

E


123

7 SUMMARY

This state of the art technology results in many advantages when comparing with traditional wash-ing equipment:

• unparalleled cleaning degrees as the Hydro-Clean’s spray output from the high pressure nozzles is able to enter porous surfaces.

• lower capital costs, as a result of the Hydro-Clean’s reduced weight which results in a smaller steel structure and smaller conveyor requirements.

• lower operation costs due to the lower usage of energy and water.• green advantages.• less stress on the material, when compared to a scrubber where the material can be damaged.• wider range of application, as a scrubber should not be filled with material <6.35 mm (forma-

tion of clay balls) and a log washer with material containing more than 15% of plastic and soluble clay (<63 μm).

• superior flexibility based on the various ways the Hydro-Clean can be adjusted. Adjustments include being able to change the distance between the nozzles and the feed material, size of the nozzles, number of the nozzles, retention time in the chamber, gate opg. between the hopper and the drum, change of the motor speed for the pump (thus a change in water pressure, adjust-ment of the rotor position (angular or straight), whereas the main competing machines can only change their angle and the speed of their shaft respectively in the drum.

• less maintenance costs when compared with the screw and log washer which need to renew their shaft nearly on an annual basis due to sealing issues with the bearings and experience continual renewal of paddles and flies.

• easy to empty thus easy to maintain and to repair.

8 CONCLUSION

The importance of the washing process as one of the first stages in a mineral processing plant and washing technologies currently available in the market have been outlined. As demonstrated, the

Figure 13. Practical washing results.

0%

2%

4%

6%

8%

10%

12%

Perc

en

tag

e <

63 μ

m

14%

16%

limestone

0 - 45 mm

gravel

0 - 45 mm

gravel

2 - 8 mm

sand

0 - 2 mm

material

before

after


124

high pressure Hydro-Clean is a leader in state of art of washing technology and comprises one of the most innovative solutions and cleaner technologies for sustainable development in mining and related industries. It has been applied with very successful results mainly because of its water and energy consumption benefits. Its continuous mode of operation can operate with feed rates up to 400 t/h with a water and energy consumption of 0.1–0.2 m3/h per ton and 0.3–0.5 kWh/t, respectively.

REFERENCES

UNESCO, 2008. Web access: http://www.unesco.org/water/ (22/01/2008).The Reporter-Herald, Pamela Dickman (Publish Date: 5/21/2008).IBRAM, 2007. Cobrança gera utilização consciente da água. Ano II, N° 11, pp. 16.Krellmann, J., & Hoppe, J. (2000). “Cost effective separation of silt and clay size material”. Aufbereitungs

Technik 41 (11), pp. 529–534.Fernandes, F. R. C., Matos, G. M. M., Castilhos, Z. C., & Luz, A. B. (2007). “Tendências tecnológicas Brasil

2015—Geociências e Tecnologia Mineral”. CETEM/MCT, Rio de Janeiro, 380 p.Hoeffl, K. (1993). Zerkleinerungs- und Klassiermaschinen. 2. ueberarbeitete Auflage, Schluetersche Verlag-

sanstalt und Druckerei GmbH & Co., Hannover, 431 p.Luz, A. B., Sampaio, J. A., Almeida, & S. L. M. (2004). “Tratamento de minérios”. 4ª. Edição, CETEM/MCT,

Rio de Janeiro, 867 p.DIN 4226 (1983). Zuschlag fuer Beton Teil 1 und 3. Beuth Verlag, Berlin.HAVER (2006). Internal Report, 17 p.


125


Remote monitoring of a high hazard coal waste impoundment in mountainous terrain case study

J.D. QuarantaDepartment of Civil and Environmental Engineering, West Virginia University, Morgantown, WV, USA

L.E. BantaDepartment of Mechanical and Aerospace Engineering, West Virginia University, Morgantown, WV, USA

J.A. AltobelloATK Tactical Propulsion and Controls, Rocket Center, WV, USA

ABSTRACT: Following the breakthrough and release of coal slurry from the Martin County Coal Corporation impoundment near Inez, Kentucky on October 11, 2000 the United States Con-gress requested the National Research Council (NRC) to examine ways to reduce these types of accidents. The NRC completed their study titled “Coal Waste Impoundments” which identi-fied numerous areas of concern and the committee presented recommendations for improving the design, operation, and safety of coal slurry impoundments. This project addresses the National Research Council’s findings for introducing state-of-practice electronic instrumentation for moni-toring parameters within the embankment and slurry pool area of an impoundment. A research project was initiated to design, install, and operate an automatic, remote data collection system for use at a coal waste impoundment. The research objectives included identification of monitor-ing methods, approaches, and limitations in mountainous terrain and fluctuating environmental conditions.

The project included development of a prototype wireless data collection system for moni-toring impoundment performance and environmental indicators at a candidate high hazard fine and coarse coal refuse impoundment constructed using the upstream design technique in Boone County, West Virginia. The instrumentation system included the engineering design, instrument system fabrication, assembly, and field construction of a prototype automatic wireless data col-lection system for monitoring impoundment performance (weather data, piezometric water levels, pH, and Specific Conductance. Data results are presented with discussions on system long-term performance, advantages, and challenges.

1 INTRODUCTION & BACKGROUND

West Virginia, with its mountainous terrain and historical coal mining production, has a legacy of pre-law (Surface Mining Control and Reclamation Act of 1977) and post-law coal slurry impound-ment sites. These sites were historically used as impounding structures for coal slurry, process black water, and coarse refuse disposal. These structures are regulated by the US Department of Labor—Mine Safety and Health Administration (MSHA) and by the West Virginia Department of Environmental Protection (DEP).

Following the breakthrough and release of coal slurry from the Martin County Coal Corpora-tion impoundment near Inez, Kentucky on October 11, 2000 the United States Congress requested the National Research Council (NRC) to examine ways to reduce these types of accidents. The NRC completed their study titled “Coal Waste Impoundments” which identified numerous areas


126

of concern and the committee presented recommendations for improving the design, operation, and safety of coal slurry impoundments (NRC, 2002). In 2003 the Coal Impoundment Project began as a program of the National Technology Transfer Center (NTTC) at Wheeling Jesuit Uni-versity (WJU). Between 2003 and 2005 research performed by West Virginia University (WVU) included the field evaluation of nineteen coal impoundments within West Virginia (Quaranta et al., 2004). Findings indicated no occurrences of electronic instrumentation in use for the real-time collection and emergency warning annunciation of important safety related parameters. The field findings also identified that coal company operators and their design engineers were not familiar with the new instrumentation equipment/systems.

The NTTC initiated this project to address the National Research Council’s findings for intro-ducing state-of-practice electronic instrumentation for monitoring parameters within the embank-ment and slurry pool area of an impoundment and investigate the applicability of a remote and autonomous impoundment instrumentation system for West Virginia. This project was initiated to design, install, and operate an automatic, remote data collection system for use at a coal waste impoundment with objectives of determining monitoring methods, approaches, and limitations in mountainous terrain and fluctuating environmental conditions unique to the state.

Inspection and monitoring of coal waste impoundments are required by MSHA under the Code of Federal Regulations, Section 77.216 (a); the regulations include a 7 day impoundment inspec-tion cycle with record keeping of the phreatic water level at the piezometer stations. In Decem-ber 2007, MSHA released the updated, draft design manual for coal refuse impoundments and has dedicated Chapter 13—Instrumentation to this topic (MSHA, 2007). This project application received a waiver from MSHA for the system installation and operation based on a research and development request.

2 OBJECTIVES AND TECHNICAL APPROACH

The main objectives of this project were to design and demonstrate a cost effective data acqui-sition system as a retrofit at an existing coal waste impoundment to collect data relevant to impoundment operation. Key variables monitored included: phreatic surface level, the pH and specific conductance of the groundwater, and the seepage flowrate. Various weather parameters monitored, included temperature, barometric pressure, rainfall, and wind speed. This paper is lim-ited to reporting on the instrumentation selected and installed at a coal waste impoundment in a mountainous region of West Virginia for accuracy and an overall evaluation of instrumentation suitability, reliability, and cost.

3 INSTRUMENTATION SYSTEM DESIGN, SETUP, AND CALIBRATION

The candidate coal refuse impoundment and industry partner was selected and are located in Boone County, West Virginia. Boone County is a rural and steep mountainous area located in the heart of the Appalachian coal mining region. The mine and impoundment facilities were under full 24/7 operation with coal production averaging 22 tons of slurry per hour and approximately 32,500 tons per hour of raw coal.

Figure 1 is a relief map of the impoundment that was chosen for this project. The map illustrates the mountainous terrain and elevation changes that led to the decision to use a wireless commu-nications system that is cost effective and provides a communications range that is sufficient for the data collection at a coal impoundment. Due to the remote location of the impoundment and constant site construction operations, a wireless communication system was proposed. The fol-lowing equipment constraints were identified:

i. Mountainous terrain and continual construction with large machinery requires that a wireless communications system be installed.


127

ii. Limited standpipe diameter requires that only a few sensors are installed at each location, thus a data logger capable of handling only a few sensors is required.

iii. All sensors and communications devices must be compatible with data logger.

Sensor components of the monitoring system included:

Piezometers and Water Level sensors: used to monitor the water level within the pool (slurry), toe, and the phreatic level in the coarse refuse structure.

pH sensors: located in open pipe piezometer, pool, and toe to detect acidity levels.Specific Conductance sensors: monitored metal ions and salts in the seepage water.Toe Pump Flowrate: monitored water discharge to determine total water balance thru the

impoundment.Temperature (water): monitored as indicator of chemical reactions, infiltration, and calibra-

tion of other sensors;Weather Station: fully automated to gather data on: ambient air temperature, barometric pres-

sure, rainfall collection, and wind speed on a real-time basis at the impoundment crestImpoundment Pool Elevation: monitored free water surface and slurry levels pumped into

the impoundment pool.

After preliminary site visits and review of site operation plans and mapping, the instrument stations were laid out as shown in elevation view in Figure 2 and the installed sensor details are listed in Table A. The features of the monitoring system included individual instrumentation arrays coupled with wireless technology for transmission of data to a central logging computer.

Figure 1 below is a relief view of the impoundment. The illustration shows the different sta-tions being used for this project and the individual equipment located at each station is listed

Figure 1. Impoundment elevation relief with remote data collection stations.


128

in Table A. The horizontal distance through the impoundment is approximately 800 meters. Table B lists the elevations of each of the piezometers and their relative elevation to the radio base station.

Figure 2, below illustrates the communication pathways used for station communication with the host computer. Every station either communicates directly to the host computer or communi-cates via a radio repeater that relays information from one station to another.

The selection of sensor equipment was to identify a stand-alone data acquisition system using commercially available equipment. Numerous vendors and products were investigated for an off- the-shelf package suitable for this application. However, no wireless communications systems

Table A. Instrumentation details.

Piezometer/Station Instrumentation

Toe Wedgewood CSIM11 pH, Wedgewood CSIM11 ORP, Campbell Scientific Water Conductance

P14 Pressure Systems Hydrostatic Water Level Transducer, Campbell Scientific Water Conductance

P13 Pressure Systems Hydrostatic Water Level Transducer and VW Piezometer.P7 Pressure Systems hydrostatic water Level Transducer, Wedgewood CSIM11 pHWeather Station HOBO Rain Gauge, HOBO Barometric Pressure, HOBO Wind Speed, HOBO

TemperaturePool Pressure Systems Hydrostatic Water Level Transducer, Campbell Scientific

Water Conductance, Wedgewood CSIM11 pH, Wedgewood CSIM11 ORP

Table B. Piezometer elevations.

Piezometer Top elevation Tip elevation Top elevation in relation to base station

7 487.6 m (1599.7 ft) 438.9 m (1440 ft) 127.9 m (419.7 ft)13 411.8 m (1351.2 ft) 377.6 m (1239 ft) 52.1 m (171.2 ft)14 453.9 m (1489.3 ft) 411.8 m (1351 ft) 94.2 m (309.3 ft)

Figure 2. Impoundment radio modem communication pathway.


129

offered by geotechnical companies were identified that were competitively priced or for which the distributors had full knowledge of their products’ range capability and sensor compatibility. Other drawbacks to these systems include proprietary software and the lack of flexibility in configuring the system.

The project pursued selecting sensors that could be integrated with general purpose data log-ging and communications equipment with a minimal amount of effort and cost. This solution required the commercial data acquisition software be modified to suit the reporting requirements for the impoundment. The wireless communication pathway setup incorporated a repeater and booster station for transmitting the data due to the lack of direct line-of-sight transmission and blind areas caused by the site’s steep slopes. The individual data loggers and radio modems were battery powered and used solar collector arrays for recharging.

To consolidate the raw data, it was necessary to establish a link to a host computer to provide sufficient memory to store all the data from the various sensors. Because the coal impoundment is such a large structure and the terrain is so mountainous, it was decided that for this project it would be impractical to hardwire all the equipment together, so in order to store all the data on a computer the information was transmitted using radio frequency modems.

Data acquisition used Onset’s Microstation data logger which is capable of handling four input signals from many types of sensors that produce a 0 to 5V output which includes most the sen-sors used in this project. A couple of sensors, including the pH and Specific Conductance (SpC) sensors, produce outputs that are bipolar (output ranges from a negative voltage to a positive voltage). The input adapters for the data logger are not capable of detecting negative voltages, so compensation for the negative voltages (ranging to − 700 mV) was accomplished by inserting a 1.5 volt alkaline battery in-line with the positive output signal of the sensor to boost the resulting readings.

MaxStream’s 9X-Tend radio modems were used to transmit the data logger signals. These modems provide up to 64.4 km (40 miles) line-of-sight data transmission range with optional high-gain antennas. These antennas were needed as line-of-sight was not possible due to the ter-rain in southern West Virginia. The 1 watt power capabilities of these modems provided enough power to circumvent the communication problems that the terrain presents. These radio modems work on the 900 MHz frequency range which carries no permit requirements for use. Also, these modems interfaced well with Onset’s Microstation data loggers. Figure 3 shows the Piezometer 14 station setup.

The data loggers use a computer communications port with a serial cable to connect directly to a computer. The modems were connected to the serial port for transmitting to the base com-puter using both stand-alone, repeater, and end-node configuration. The versatility to function in

Figure 3. Piezometer 14 equipment.

Solar Panel

Data Logger

Radio Modem

Battery


130

different modes allowed several different configurations to communicate the data. These configu-rations include direct links (line-of-sight), and indirect links (non-line-of-sight using a modem as a data repeater). The modems come with an easy to use software application which allowed programming for different channels (up to 9), and permitted data transfer rates and power con-sumption to be accomplished in a short period of time.

The final advantage that is provided by the 9X-Tend modems is that all programming is stored in non-volatile memory. This was beneficial since even in accidental or purposeful power out-ages the modems’ code would not be altered or lost. This feature ensured that unless a modem’s program is changed directly by a computer, the code will remain the same for accurate data transmission.

4 RESULTS AND INTERPRETATION

The data loggers used for the project were set at collecting the field data at four hour intervals at seven days per week. Figure 4 A through C illustrate data recovered from Piezometers 14 and 13. Piezometer 14 is up-gradient of Piezometer 13 horizontally 167 meters and vertically 45 meters. Figure 4 A is a graph of the daily rainfall data between November thru April 2006, and Figures 4B and 4C are the hydrostatic water level transducer levels recorded. In Figure 4B the water level in Piezometer 14 steadily increases over the course of three months consistent with increases in rain-fall levels. This increase is on the magnitude of approximately four feet. The end of the increase seems to correspond to the increased rainfall in the area, but the rise early does not seem to cor-respond to any monitored data. Gaps in the Piezometer 14 data were attributed to low voltage in the batteries indicating that the solar recharge was not occurring. The corresponding rainfall data response for Piezometer 13 is graphed in Figure 4C. Piezometer 13 appears to remain within a narrow response range and to be affected by the rainfall. Piezometer 13 also does not appear to show a similar response trend at Piezometer 14 to the increased rainfall values. The difference in piezometer response may be attributed to the 3-Dimensional affect of the seepage flow possibly discharging to surface bench drains or reduced hydraulic conductivity of the refuse around the piezometer.

Figures 5 A, B, and C are graphs of the weather rainfall data and the specific conductance changes at Piezometer 14 and at the Toe Drain of the impoundment. The specific conductance of the seepage water can be affected by rainwater on the measurement of dissolved solids in water, which is an indication of the specific conductance. For the specific conductance at P14, the drop in the readings coincides (with a 5 day lag) with the significant amount of rainfall experienced between December 20, 2006 and January 2, 2007. This indicates that the rainfall has affected the readings of the sensor. These reading remain at a reasonably constant value as the seepage flow continues. The specific conductance response at the impoundment toe drain remains reasonably steady.

Figure 6 shows the graph comparing the data collected by the hydrostatic water level trans-ducer at Piezometer 13 as compared with the field data collected by the impoundment inspector at the mine using a manual weighted tape m-scope. Table C shows the field data collected over a six month time interval. Table C contains the elevations of the m-scope data tend to follow the same trend as the data collected by the hydrostatic water level transducer. This table contains the reduced general statistics for the data, with all elevation readings displayed in feet for simplifica-tion purposes. The missing data in the sensor reading column is a result of corrupt data due to a low input voltage to the sensor.

The difference of the mean measurement with the HSWLT sensor is approximately 0.03 m (0.10 feet). The average difference between the Field vs. HSWLT sensor is -0.021 m (-0.07 feet). The final comparisons were of the maximum and minimum absolute difference from one data set to the other. The maximum was 0.07 m (0.23 feet), and the minimum was 0.00 m (0.00 feet). These statistics show that the hydrostatic sensor used in this project at Piezometer 13 con-tains electrical inconsistencies in its measurements; although, it can also be assumed that there


131

Figure 4. A, B, and C: Rainfall amounts versus Piezometer 14 and 13 response.A) Rainfall, B) Piezometer 14, C) Piezometer 13.

0

0.2

0.4

0.6

0.8

1

1.2

11/1

/200

6

11/1

5/20

06

11/2

9/20

06

12/1

3/20

06

12/2

7/20

06

1/10

/200

7

1/24

/200

7

2/7/

2007

2/21

/200

7

3/7/

2007

3/21

/200

7

4/4/

2007

4/18

/200

7

Date

Rain

fall (

in)

1356.00

1358.00

1360.00

1362.00

1364.00

1366.00

1368.00

9/2/

2006

9/16

/200

6

9/30

/200

6

10/1

4/20

06

10/2

8/20

06

11/1

1/20

06

11/2

5/20

06

12/9

/200

6

12/2

3/20

06

1/6/

2007

1/20

/200

7

2/3/

2007

2/17

/200

7

3/3/

2007

3/17

/200

7

3/31

/200

7

4/14

/200

7

4/28

/200

7

Date

Wate

r E

levati

on

(ft

)

Missing data result of dead battery

P13 Water Levels

1240.00

1240.50

1241.00

1241.50

1242.00

1242.50

1243.00

9/17

/200

6

10/1

/200

6

10/1

5/20

06

10/2

9/20

06

11/1

2/20

06

11/2

6/20

06

12/1

0/20

06

12/2

4/20

06

1/7/

2007

1/21

/200

7

2/4/

2007

2/18

/200

7

3/4/

2007

3/18

/200

7

4/1/

2007

4/15

/200

7

Date

Wa

ter

Ele

va

tio

n (

ft)

A

B

C


132

Figure 5. A, B, C: Rainfall versus specific conductance data at P14.A) Rainfall Data, B) Piezometer 14, C) Toe drain.

1.000

1.500

2.000

2.500

3.000

3.500

4.000

11/1

3/20

06

11/2

7/20

06

12/1

1/20

06

12/2

5/20

06

1/8/

2007

1/22

/200

7

2/5/

2007

2/19

/200

7

3/5/

2007

3/19

/200

7

4/2/

2007

4/16

/200

7

Date/Time

Sp

. C

on

d.

(mS

cm

-1)

1.00

1.50

2.00

2.50

3.00

3.50

4.00

11/1

3/20

06

11/2

7/20

06

12/1

1/20

06

12/2

5/20

06

1/8/

2007

1/22

/200

7

2/5/

2007

2/19

/200

7

3/5/

2007

3/19

/200

7

4/2/

2007

4/16

/200

7

Date/Time

Sp

. C

on

d. (m

S c

m-1

)

0

0.2

0.4

0.6

0.8

1

1.2

11/1

/200

6

11/1

5/20

06

11/2

9/20

06

12/1

3/20

06

12/2

7/20

06

1/10

/200

7

1/24

/200

7

2/7/

2007

2/21

/200

7

3/7/

2007

3/21

/200

7

4/4/

2007

4/18

/200

7

Date

Rain

fall (

in)

A

B

C

Figure 6. Piezometer 13 field data levels vs. HSWLT levels.

1240.00

1240.50

1241.00

1241.50

1242.00

1242.50

1243.00

1243.50

1244.00

9/13/2006

9/20/2006

9/27/2006

10/4/2006

10/11/2006

10/18/2006

10/25/2006

11/1/2006

11/8/2006

11/15/2006

11/22/2006

11/29/2006

12/6/2006

12/13/2006

12/20/2006

12/27/2006

1/3/2007

1/10/2007

1/17/2007

1/24/2007

1/31/2007

2/7/2007

2/14/2007

Date

Ele

vati

on

(F

t)

Sensor Data EAC Data


133

Table C. General comparison statistic for field and sensor data at P13.

P-13 field Difference field vs.EAC data reading (ft) Sensor reading (ft) sensor (ft) %Error

9/20/06 12:00 PM 1241.509/27/06 12:00 PM 1241.55 1241.55 0.00 0.0010/4/06 12:00 PM 1241.65 1241.69 −0.04 0.0010/11/06 12:00 PM 1241.40 10/18/06 12:00 PM 1241.50 1241.64 −0.14 −0.0110/25/06 12:00 PM 1241.55 1241.78 −0.23 −0.0211/1/06 12:00 PM 1241.55 11/8/06 12:00 PM 1241.55 1241.75 −0.20 −0.0211/15/06 12:00 PM 1241.80 11/22/06 12:00 PM 1241.58 1241.58 0.00 0.0011/29/06 12:00 PM 1241.55 1241.58 −0.03 0.0012/6/06 12:00 PM 1241.50 1241.67 −0.17 −0.0112/14/06 12:00 PM 1241.65 1241.78 −0.13 −0.0112/20/06 12:00 PM 1241.94 1241.87 0.07 0.0112/27/06 12:00 PM 1241.66 1241.73 −0.07 −0.011/3/07 12:00 PM 1242.00 1242.01 −0.01 0.001/10/07 12:00 PM 1241.85 1241.89 −0.04 0.001/17/07 12:00 PM 1242.00 1242.06 −0.06 0.001/24/07 12:00 PM 1241.80 1241.84 −0.04 0.001/31/07 12:00 PM 1241.65 1241.73 −0.08 −0.01MEAN = 1241.66 1241.76

Sensor vs. Field(ft)

MEAN DIFF. = 0.10

Avg. Diff. = -0.07

Max Diff. = 0.23

Min Diff. = 0.00.

*MEAN DIFF. = The difference between the means of the measurements.

** Avg. Diff. = The average of the difference between the measurements.

are inconsistencies in the manually read data as well. The inconsistencies, or “noise,” in the manually read data can be attributed to moisture or mud that has collected on the inside of the standpipe causing a false reading. If sensor measurements can be considered precise to within an average 0.04 m (0.14 feet) of an actual water level then this is completely capable of producing reliable data.

4 PROJECT SETBACKS AND SOLUTIONS

While most of the field problems had definite solutions, some problems were just temporarily solved or the recurring nature of those problems was lessened without eliminating the problem. For several of these types of problems, more permanent solutions or recommendations are presented.

4.1 Battery power loss

The first problem encountered in the operation of this project was the rapid loss of power from the 12 volt batteries used to power the radio modems and some of the sensors. Initially, a 7.5 amp-hour (AH) battery was installed to supply the power while a single 5 watt solar panel was used to recharge the battery. It was found that the quiescent current draw and the transmitting current draw


134

of the modem caused a power drain that was more rapid than the 5 watt solar panel could handle. The batteries were found to be below operational standards within approximately two weeks. The solar panel was originally designed with the assumption that there were on average approximately 6 hours of sunlight per day. This amount would be enough to keep the batteries charged; however, it was found that over the course of a few cloudy days the power drawn from the battery was unre-coverable. The primary problem was that the face of the dam on which the sensors and solar panels were located faces northwest, and is flanked by high ridges to the east and west. Thus most of the panels received direct sunlight only a few hours per day, from mid morning to early afternoon. The remainder of the day the panels were shaded by either the ridges or the dam itself. In addition, the panels could not be mounted at a steep enough angle to shed snow and sleet in winter, so their effectiveness was further impaired by precipitation in freezing weather.

The first solution proposed for this problem was to add a second 5 watt solar panel, in parallel, to aid in the charging of the batteries. It was calculated that the modems would cause a loss of 2.64 Amp-hours (AH) over the course of 24 hours (assuming a current draw of 0.110 A). Assum-ing 6 hours of sunlight per day, one 5 watt solar panel (0.33 A) would provide 1.98 AH per day. This is a net of −0.66 AH per day. This would cause full discharge of the 7.5 AH batteries in 11 days. By adding an additional solar panel, the net would increase to +1.32 AH per day. This would theoretically ensure the batteries would never fully discharge. However, this solution proved to only increase the power life of the batteries by approximately two to four days, so a second solu-tion was proposed.

The second solution involved upgrading the batteries from 7.5 AH batteries to 18 AH batteries. It was theorized that the 7.5 AH battery could not fully store all of the available energy from the two solar panels, and any series of cloudy days would exhaust the battery before the sun had a chance to recharge it. With two solar panels the increased battery capacity resulted in a more reli-able system, by allowing a wider range in days for the solar panels to perform their intended jobs. Now the batteries at most of the stations have proven to last approximately 4 to 5 weeks, with the battery at Piezometer 14 lasting as long as 9 weeks even in winter. This has been due, in most part, to the station at P14 receiving the most direct sunlight while the weather station receives the least due to the location below a hillside that blocks most of the winter sunlight. As spring and summer approach the ability of the solar panels to recharge the batteries should become more sufficient and should allow the 18 AH batteries to last indefinitely. A proposed future resolution for this problem for any project similar in nature to this one would be to replace the radio modems with ones that would require less power. Alternately, timing circuitry could be installed to switch the radio modems off for the 99% of the time they are not transmitting data. These changes would also alleviate the problem of insufficient power to the hydraulic water level sensors as was discussed earlier.

4.2 Sensor communication errors

Another problem encountered in this project was the occasional loss of data due to unknown rea-sons for sensor communication errors. These errors typically would cause the loss of several data points in succession but would resolve themselves without intervention. This occurred both at the toe of the impoundment with the ORP and specific conductance sensors as well as with the smart sensors located at the weather station. As for the communication errors at the toe, for unexplained reasons the loss of approximately ten data points was encountered, but with the “rebooting” of the data logger no additional errors have been recorded.

The weather station was a different story. An unexplained communication error was encoun-tered, and it resulted in the loss of approximately two days worth of data (48 points of data per sensor). After the data logger was “rebooted” the problem was not resolved. The solution was to replace the data logger with a new, unused unit. This has proven to be an effective solution, and no more communication errors have been experienced. Due to the location of the weather station near the top of the dam, it is conjectured that a nearby lightning strike may have disabled the data logger but there is no way to confirm that theory.


135

4.3 Equipment siting

Most of the data acquisition and telemetry equipment must necessarily be sited where the piezom-eters in the dam are located. As the dam grows, new piezometer wells are added and existing ones are extended by adding lengths of casing to the top of the well. As the dam grows, the crest moves upstream, and the wells eventually become part of the downstream face of the dam. Once a well is off the crest of the dam, it becomes stable and no more additions are necessary, but for those piezometers on the top of the dam, there is constant danger of damage from heavy machinery and the reasonably frequent need to add more casing as the dam grows. The sensors in the piezometer well must be pulled up, the new segment of casing added, and the sensors lowered back down into the well.

Occasionally the ground will shift within the dam, causing the piezometer casing to be either distorted or crushed, and preventing the removal or reinsertion of the sensors. For these reasons, it is probably best to use automated instrumentation on wells only after they are off the crest of the dam and on the downstream face, and to continue to use manual methods to measure phreatic levels and other parameters of interest for the wells on the crest and upstream sides of the dam.

5 SYSTEM COST

The costs for the fabrication and installation of the instrumentation system built for this project are outlined in Table D. This system was constructed as a prototype, wireless, stand-alone system with capabilities for expansion as needed.

The costs in Table D may be applied for each instrumented station at an impoundment. A compu-ter base station would be needed but the cost is not included here as any site computer could serve this function. Based on a site having five water level monitoring stations the estimated equipment cost would be $8675. The cost of a weather station is approximately three thousand dollars which brings the total system cost to approximately $12,000.

6 CONCLUSIONS AND RECOMMENDATIONS

From this work it is clear that development of a wireless system to collect and analyze data is possible, but not trivial. The main issues are the siting of the data logging and radio equipment and the provision of reliable power to the electronics, with the two issues being related. South-facing impoundments with clear paths to a location for the data collection computer will work well. Sites such as the one in which this work was done are more problematic, due primarily to the difficulty of providing power to the equipment. Even this site would have been more feasible had there been a better line of sight from the piezometer locations to the maintenance building

Table D. Instrumentation equipment system cost per station.

Component Cost ($)

Hydrostatic Water Level Transducer 635Ancillary sensor (pH, ORP) 135HOBO brand Micostation Data Logger 200XTend radio modem 300High gain antenna 12518 AH battery 40Solar panels 200NEMA 4—Hinged panel box/post 100

Total Cost/Station $1735


136

where the data collection computer was housed. If the piezometer locations could be chosen from the outset with these requirements in mind, some of the problems experienced in this work would have been alleviated. Alternately, a better site for the repeater would have helped, but none was available without using machinery to clear trees, and that action was not feasible in this project. In retrospect, the best solution might have been to use lower-powered transmitters and install them in a denser network with multiple repeaters instead of using a single repeater and high-powered modems. Future work in this area should concentrate on reducing the power demands of the wireless system and on improving its reliability. Alternatives to the equipment used in this work have recently become available that trade reduced radio range (∼400 meters line-of-sight) for low power consumption (6–12 months on 2 D-cells with no recharging, depending on data collection rates).

We believe that further development of this concept is warranted from both a safety and from an economic standpoint. Each station, depending on the sensors included, can cost around fifteen-hundred dollars. However, the ability to collect and analyze data continuously is of great interest from a safety perspective, and may justify the expense. The ability to do this without sending a person to physically collect the data is lucrative from an economic standpoint. Together, the fac-tors argue heavily for further work to refine and commercialize this technology.

ACKNOWLEDGEMENTS

This project was made possible through funding from the National Technology Transfer Center (NTTC) at Wheeling Jesuit University through a grant from the US Mine Safety and Health Administration. The authors wish to thank Mr. J. Davitt McAteer, Program Director, and Mr. Paul Myles, Project Manager for their support in completing this work; the authors further wish to thank the MSHA Pittsburgh Safety and Health Technology Center, Mine Waste and Geotechnical Engineering Division for their involvement and support of this work. The authors wish to thank the engineers and field personnel from the Eastern Associated Coal—Patriot Coal Division for all of their kind support in helping to make this research possible.

REFERENCES

MSHA. 2007. Engineering And Design Manual: Coal Refuse Disposal Facilities Advance Draft For Industry Review And Comment. U.S. Department of Labor, 937 pp.

NRC [National Research Council]. 2002 Coal Waste Impoundments: Risks, Responses, and Alternatives. Washington, DC: National Academy Press, 230 pp.

J.D. Quaranta, B. Gutta, B. Stout, D. McAteer, and P. Ziemkiewicz, “Improving the Safety of Coal Slurry Impoundments in West Virginia,” Tailings 2004, Vail CO.


Geotechnical considerations



139

Experimental characterization of the influence of curing under stress on the hydromechanical and geochemical properties of cemented paste backfill

E. Yilmaz, T. Belem, M. Benzaazoua & B. BussièreUniversité du Québec en Abitibi-Témiscamingue, Rouyn-Noranda, Quebec, Canada

ABSTRACT: This paper describes the results of laboratory investigations intended to inves-tigate the influence of curing under stress on the hydromechanical and geochemical properties of Cemented Paste Backfill (CPB). A new laboratory equipment named CUAPS (curing under applied pressure system) that mimics the in situ placement and curing conditions of CPB mate-rial was used, along with conventional mould samples. A big difference was observed between the Uniaxial Compressive Strength (UCS) obtained from CUAPS-consolidated CPB and conven-tional mould-unconsolidated CPB samples. An evolution of physico-chemical and geochemical properties of the CPB material is compared for binder contents of 3, 4.5 and 7 wt%, and curing times of 7, 14 and 28 days. The results have shown that the water drainage due to consolidation has an advantageous effect on CPB hardening. Consequently, the performance of the CUAPS-consolidated CPB is always more realistic than the performance of conventional plastic mould-unconsolidated CPB, whose strength is known to be underestimated.

1 INTRODUCTION

The mining industry uses mine waste materials (mainly tailings and waste rocks) for the prepa-ration of backfill, which fills underground voids created by ore extraction. The choice of mine backfill systems such as rock fill, slurry fill or paste backfill plays a major role in the productiv-ity, safety, and economic benefits of operating underground mines (Hassani & Archibald 1998; Bussiere 2007). However, compared with other backfill systems, cemented paste backfill (CPB) has begun to be increasingly used by most modern mines worldwide, due to its significant cost advantages and the potential for placing the full plant sulphidic tailings in underground stopes (Landriault 2001).

CPB is usually prepared by mixing the tailings, binder and mixing water to form a composite construction material. After its hardening, the CPB material acts as secondary ground support. Physico-chemical and mineralogical properties of CPB ingredients greatly affect both strength and stability performance (Amaratunga & Yaschyshyn 1997; Archibald et al. 1999; Benzaazoua et al. 1999, 2002, 2004; Belem et al. 2000; Mohamed et al. 2001; Yilmaz 2003, Yilmaz et al. 2003; Kesimal et al. 2004, 2005; Godbout 2005; Klein & Simon 2006; Ouellet 2007). Uniaxial Compressive Strength (UCS) is the key index parameter mostly utilised for CPB stability design (Stone 1993, Belem & Benzaazoua 2008a, b). The UCS requirements are closely related to the assortment of roles undertaken by CPB. Overall, the typical target UCS values for CPB vary from 0.25 to 4.35 MPa for a wide range of applications varying from mine tailings disposal to roof sup-port in underground mines (Hassani & Bois 1992, Belem & Benzaazoua 2008a).

CPB design is frequently based on the evaluated properties of laboratory-prepared backfill material (conventional plastic mould samples). However, experience indicates that it is not easy to ensure that all the physical and mechanical conditions prevailing underground are observed when the CPB sample is prepared, cured and tested in the laboratory (Servant 2001). Besides,



140

the mechanical and curing properties of in situ CPB material vary appreciably, depending on the methods of preparation, placement and the conditions of the mine environment (Belem et al. 2002; Le Roux et al. 2005). In order to obtain a more realistic mechanical response of CPB material, it is very important to take into account the intrinsic and extrinsic factors affecting in situ backfill performance. Previous studies indicate that the UCS magnitudes of in situ CPB material are often 2 to 4 times higher than those obtained from laboratory-prepared samples of identical batch mix and curing time (Servant 2001; Belem et al. 2002; Cayouette 2003; Revell 2004; Le Roux et al. 2005). The discrepancy may be attributed to the fact that the hardening of CPB material cured under an effective field stress increases the rate of strength development and the ultimate strength, as reported recently by a number of authors (Belem et al. 2002, 2006; Helinski et al. 2006; Fourie et al. 2006; Grabinsky and Simms 2006; Yilmaz et al. 2006, 2008a). As well, Revell (2004) has indicated that scale effects have a large impact on CPB strength performance. To the knowledge of the authors, there is no appropriate laboratory apparatus and/or standardized test procedure for CPB that mimics the field mixing, placement and curing conditions on a laboratory scale. Moreover, only few studies showed that consolidation both under self-weight and time-dependent surcharge loading could have a notable effect on the overall performance and quality of CPB materials (Belem et al. 2002, 2006, 2007; Benzaazoua et al. 2004; Le Roux et al. 2005; Yilmaz et al. 2006, 2008a, b). In the literature, there is a lack of knowledge about the characteristics of CPB cured under constant and/or variable applied pressure. Consequently, an understanding of the influence of curing under stress on CPB performance is required, for a more reliable and better quality of backfill design.

The main objective of this study is to evaluate the effects of binder content and curing time on CPB hydromechanical performance, as well as to observe the resulting geochemical properties. To achieve this objective, a new laboratory consolidometer named CUAPS (curing under applied pressure system) that allows for the simulation of the in situ placement and curing conditions of CPB materials is used (Benzaazoua et al. 2006). A comparative analysis of the above-mentioned properties of the CPB cured in both conventional plastic moulds and under variable applied pres-sures was performed using three different binder contents (3, 4.5 and 7 wt%) and curing times (7, 14 and 28 days).

2 MATERIAL AND METHOD

2.1 Characteristics of paste backfill ingredients

Tailings: the tailings sample used in this study was taken from a Canadian gold mine, located in the western part of the province of Quebec. The sample’s particle size distribution (PSD) was

Figure 1. Particle size distribution curves of the tailings sample used: a) cumulative, b) incremental.

Clay (less than 2 μm) Silt (2 – 50 μm)

50

60

70

80

90

100

Typical range of paste backfill in Canadian mines

(a)

Clay (less than 2 μm) Silt (2 – 50 μm) Sand (50 −2000μm)

0

10

20

30

40

0.01 0.1 1 10 100 1000

Particle diameter (log μm)

Cu

mu

lati

ve

vo

lum

e (

%)


Clay (less than 2 μm) Silt (2 – 50 μm)

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Inc

rem

en

tal

vo

lum

e (

%)

Sand (50 − 2000μm)

0.01 0.1 1 10 100 1000


(b)


0.5

1.0

1.5

2.0

10 100 1000




141

determined using a Malvern Mastersizer laser diffraction-type particle size analyser (Fig. 1), fit-ting with a typical range of PSD curves of 11 mine tailings sampled from the underground hard rock mines located in the provinces of Quebec and Ontario, Canada (Ouellet 2007).

Table 1 also shows the main physico-chemical characteristics of the tailings sample. It can be observed that the fines content (minus 20 μm) is 44 wt% which corresponds to medium size tail-ings, according to Landriault (2001). Tailings are well-graded with a coefficient of uniformity C

u

(= D60

/D10

) of 8 and with a coefficient of curvature Cc (= D

302/D

60*D

10) of 1.1. Based on the Uni-

fied Soil Classification System (USCS), the studied mine tailings are classified as low plasticity silt (ML). The specific gravity G

s of the tailings sample that was measured with a Micromeritics®

Accupyc 1330 helium gas pycnometer is 3.7, reflecting a relatively high content of metal-barren sulphides.

The chemical composition analyses of metals and total sulphur (Stotal

) were conducted on the tailings sample using a Perkin-Elmer® Model Optima 3100 RL ICP-AES (inductively coupled plasma-atomic emission spectroscope) after complete acid digestion by HNO

3, Br

2, HF and HCl.

Dilute HCl was used to extract sulfates and the solution obtained was analyzed by ICP-AES. From the chemical composition analysis carried out on the tailings sample, it can be observed that the contents of iron Fe, sulphur S, aluminium Al, and calcium Ca were 27.4 wt%, 20.6 wt%, 2.8 wt% and 0.57 wt%, respectively.

Binding agent: the binder used for the preparation of CPB samples is a blend of Ordinary Portland cement or Type I (PCI) and ground granulated blast furnace slag (Slag). The blending ratio of PCI to Slag is 20/80. Three different binder proportions of 3, 4.5 and 7 wt% were chosen to evaluate the effect of curing conditions on the UCS performance of CPB samples. Proportions were calculated on the basis of weight, relative to the total dry tailings.

Moreover, the chemical composition analyses were conducted on the studied binder samples using ICP-AES method, as explained in the previous section. Table 2 shows the main chemical composition and physical properties of the binding agents used in the experiments.

Mixing water: two types of water: namely, as-received tailings pore water and tap mixing water were chemically analyzed using ICP-AES method. Table 3 summarizes the results of the chemical

Table 1. Main physico-chemical characteristics of the tailings sample used.

Ss D

10 Fines C

u C

c Al Ca Fe S

total

Parameter (m2/kg) Gs (μm) (<20 μm) (wt%) (wt%) (wt%) (wt%) (wt%) (wt%)

Tailings 2170 3.7 4.26 44 wt% 8 1.1 2.8 0.57 27.4 20.6

Ss: specific surface; D

10: effective particle diameter.

Table 2. Chemical composition and some physical properties of the binders used.

Ss

Al2O

3 CaO Fe

2O

3 K

2O MgO Na

2O SO

3 SiO

2

Parameter (m2/kg) Gs (wt%) (wt%) (wt%) (wt%) (wt%) (wt%) (wt%) (wt%)

PCI 1580 3.1 4.86 65.76 2.44 0.83 2.21 2.11 3.67 19.51Slag 3540 2.8 10.24 31.41 0.55 0.51 11.3 2.01 3.27 36.22PCI-Slag 2840 2.9 4.26 42.82 0.64 0.55 6.19 2.03 3.35 30.91

Table 3. Chemical and geochemical analysis results of pore and tap waters used as mixing water.

Parameter/ EC Eh SO42 Si Mg Al Ca Cu Fe

element (mS/cm) pH (v) (ppm) (ppm) (ppm) (wt%) (ppm) (ppm) (wt%)

Pore water 7.42 9.4 0.147 4883 0.891 1.83 0.212 559 0.286 0.011Tap water 0.274 7.8 0.431 138 0.901 2.27 0.01 40.9 0.835 0.066


142

and geochemical analyses. It can be observed that the pore water is highly aggressive in terms of sulphate content (SO

42– content around 4880 ppm) but also contains calcium (Ca content about

560 ppm). While Ca comes from the lime added during ore treatment, sulphates can come from the cyanide elimination process (tailings are treated before backfilling), and those produced from sulphide reactivity (Benzaazoua et al. 2004).

Additionally, the values in terms of pH, Eh (redox potential) and EC (electrical conductivity) of pore and tap waters were 9.41, 0.147 volt and 7.42 mS/cm, and 7.82, 0.431 volt and 0.274 mS/cm, respectively. The relatively high value of EC within the pore water of as-received tailings indicates the presence of many dissolved conducting ions.

2.2 Experimental program

Consolidometer: a new experimental apparatus named CUAPS (Curing Under Applied Pressure System) was developed for simulating field placement and curing conditions of laboratory-prepared CPB materials. CUAPS apparatus has been described previously by Benzaazoua et al. (2006). The CUAPS allows the operator to estimate the more realistic UCS values of the CPB materials cured under constant or variable pressure up to 400 kPa. The operating principle of CUAPS is that CPB are one-dimensionally consolidated by axial pressure applied over time. A Polycarbonate Perspex tube in the CUAPS apparatus forms a highly resistant portion against volume change after a series of pressure is applied. Following an increase in total pressure, excess pore water within CPB material is expulsed through a drainage port located at the bottom of CUAPS. During curing, drainage water is collected in order to calculate drainage rate and to ana-lyze the chemical composition for the different recipes.

Backfill sample preparation: CPB ingredients (i.e. tailings, cement and water) were thoroughly mixed using a double spiral mixer for about 7 minutes. After mixing, the measured slump of all mixtures was set to approximately 7 inches by adding more or less mixing water. The slump test was conducted according to ASTM C143 standard. The initial gravimetric water content w

g for all

CPB samples was kept constant at 28.2 wt%.Table 4 lists the initial bulk properties of each CPB batch mixture. These values represent volu-

metric binder content Bv, solids mass concentration C

w, volumetric solids concentration C

v, water-

to-cement ratio w/c, specific gravity Gs, void ratio e, wet density ρ, dry density ρ

d, and volumetric

water content θ. It can be observed from Table 4 that the w/c ratios reduce from 9.68 to 4.31 when the amount of binder added to the CPB batch is increased from 3 to 7 wt%. Also, ρ

d value was

1810, 1800, and 1790 kg/m2 for 3, 4.5 and 7wt% binder contents, respectively.Immediately after mixing, the prepared paste materials were poured into both conventional

plastic moulds and transparent polycarbonate round tubes (CUAPS cell samples holders). Tubes and moulds have a shape factor of 2 that corresponds to height-to-diameter ratio. CPB samples were then sealed and stored in a humidity chamber maintained at ∼80% relative humidity and 24 °C ± 2°C in order to mimic curing conditions similar to those observed in the underground stopes of the mine under study. Figure 2 shows CUAPS-consolidated CPB samples and plastic mould-unconsolidated CPB samples being cured in a controlled-humidity chamber.

One-dimensional consolidation tests: as part of the present investigation, one-dimensional consolidation tests, inspired by ASTM D2435 and D4186 standards were carried out. CPB sam-ples were consolidated using the CUAPS apparatus which allows a pressure range of 0−500 kPa.

Table 4. Initial bulk properties of different paste backfill material prepared for a given batch mix.

ρ Bv C

w C

v ρ

d S

r

Parameter (kg/m2) (wt%) (wt%) (wt%) Gs w/c e (kg/m2) (wt%) θ

3.0 wt% 2320 3.8 78 49.1 3.68 9.68 1.04 1810 100 0.54.5 wt% 2310 5.7 78 49.2 3.66 6.55 1.03 1800 100 0.57.0 wt% 2300 8.8 78 49.3 3.64 4.31 1.03 1790 100 0.5


143

Before running the tests, a seating pressure of 15 kPa was applied to provide contact between sam-ple and loading platen. Then, an example pressure sequence of 0.5, 25, 50, 100, 200 and 400 kPa, based on a load increment ratio LIR of 1, was applied to CPB samples and the axial deformations during each pressure increment were measured with LVDTs (linear displacement variable trans-ducers) through a HOBO U12-4 data logger at a time interval of 0, 2, 4, 6, 8, and 10 hours. All pressures were incrementally applied during the first day of curing and unloaded down to 0.5 kPa after the desired curing time.

Unconfined compression tests: A total of 36 CPB samples, 9 CUAPS-consolidated and 27 con-ventional mould-unconsolidated samples were cured for uniaxial compressive strength (UCS) tests after periods of 7, 14, and 28 days. After curing, all CPB samples were subjected to the UCS tests according to ASTM C39 using a computer−controlled MTS 10/GL press, which has a nominal load-ing capacity of 50 kN and a displacement speed of 1 mm/min. The UCS magnitude of each sample is recorded (peak or ultimate stress) along with the elasticity modulus of deformation. For a given binder proportion and curing time, only one-test is done for CUAPS-consolidated backfill samples whereas the average of three tests is taken for conventional plastic mould-unconsolidated backfill samples.

Bulk properties determination: after UCS testing, the calculated and/or measured CPB geotechni-cal index (bulk properties) parameters are determined as follows: gravimetric water content w (%), solids specific gravity G

s, degree of saturation S

r (%), void ratio e (or porosity n), volumetric water

content θ, dry density ρ (g/cm3), solid concentration Cw (%) and specific surface area S

s (m2/kg). The

Ss parameter, based on the BET (Brunauer, Emmett, and Teller) method, is evaluated by the nitrogen

N2 isotherm adsorption using a Micromeritics® surface analyzer model Gemini III.Chemical and geochemical analysis: during the one-dimensional consolidation test, a quantity

of water is collected from each CUAPS-consolidated CPB sample at a specific time interval. After the acidifying process, analysis of the chemical composition of the collected water is carried out using a Perkin-Elmer ICP−AES (Optima 3100 RL). Also, the pH, redox potential (Eh) and electri-cal conductivity (EC) parameters are also measured using a Benchtop pH/ISE meter Orion Model 920 A coupled with a Thermo Orion Triode combination electrode (Pt-Ag-AgCl).

3 RESULTS

3.1 Drainage water of consolidated CPB samples

Figure 3 shows the evolution of cumulative drainage water Wd versus elapsed time for CUAPS-

consolidated CPB samples prepared with a binder content of 3, 4.5 and 7 wt% as a function of elapsed time (7, 4.5 and 7 days of curing).

Figure 2. Photos of (a) CUAPS-consolidated CPB samples and (b) conventional mould-unconsolidated CPB samples curing in a humidity chamber.


144

It can be observed from Figure 3 that the Wd parameter is greatly affected by the hardening

process, depending on the amount of binder added to the CPB. The drainage rate decreases with increases of both binder content and curing time. For a 28-day curing time, the volume of drainage water for CPB with 3, 4.5 and 7 wt% binder was 19.6 wt% (253.5 mL), 18.8 wt% (242.7 mL) and 17.1 wt% (217.8 mL), respectively. As well, the pressure varying from 0.5 to 400 kPa plays a key role in the volume of water collected from samples. Overall, a pressure of 100, 200 and 400 kPa that corresponds to an elapsed time varying between 120 and 600 minutes induced a significant gap in the magnitude of W

d. However, a lower pressure of 0.5, 25 and 50 kPa did not have a big

influence on the water collection, even at early curing times (up to 7 days).

3.2 Vertical strain of consolidated CPB samples

Figure 4 shows the evolution of the vertical strain εv (= ΔH*100/H

0) of CUAPS-consolidated CPB

samples cured under variable pressures as a function of binder content (3, 4.5 and 7 wt%).For a given applied pressure, Figure 4 shows that the total ε

v observed from CPB samples con-

taining 3 wt% binder is slightly greater than those observed from CPB samples containing the 4.5 and 7 wt% binder content. For pressure greater than 200 kPa, the settlement behaviour is relatively similar. For a binder content of 3, 4.5 and 7 wt%, the observed ε

v value varies from 12.5 to 13.4%,

from 12.2 to 13.3%, and from 11.2 to 12.1 for samples 1, 2 and 3, respectively. The shape of the curves is different particularly for pressure less than 200 kPa. As the amount of binder added to CPB material increases, the ε

v value (for pressure less than 200 kPa) decreases mainly because of

binder hydration during the first day

3.3 Characterization of CPB mechanical strength

Figure 5 shows the difference in terms of strength development between CUAPS-consolidated CPB samples and plastic mould-unconsolidated CPB samples as a function of binder content and curing time. Strength proportionally increases with binder content as expected.

0

4

8

12

16

20

24

28

0.1 1 10 100 1000 10000 100000

Elapsed time (log min)

Cu

mu

lati

ve d

rain

ag

e w

ate

r (w

t%)

1000000

4

8

12

16

20

24

28

0.1 1 10 100 1000 10000


Cu

mu

lati

ve d

rain

ag

e w

ate

r (w

t%) 0

4

8

12

16

20

24

28

0.1 1 10 100 1000 10000 100000


Cu

mu

lati

ve d

rain

ag

e w

ate

r (w

t%)

rednib%tw7rednib%tw5.4rednib%tw3

)c()b()a(

Sample 3

Sample 1Sample 2Sample 3

Sample 1Sample 2


Figure 3. Evolution of CPB drainage water : a) 3 wt%, b) 4.5 wt%, and c) 7 wt% binder content.

0

2

4

6

8

10

12

14

16

0.1 1 10 100 1000Applied pressure (log kPa)

Vert

ical str

ain

(%

)

0

2

4

6

8

10

12

14

16

0.1 1 10 100 1000

Applied pressure (log kPa)

Vert

ical str

ain

(%

)

0

2

4

6

8

10

12

14

16

0.1 1 10 100 1000

Applied pressure (log kPa)

Vert

ical str

ain

(%

)

rednib%tw7rednib%tw5.4rednib%tw3

(a) (b) (c)




Figure 4. Variation of CPB vertical strain with applied pressure: a) 3%, b) 4.5%, and c) 7 wt% binder.


145

Generally, the CUAPS−consolidated CPB samples systematically produce higher mechanical strength (higher UCS value) than conventional mould-unconsolidated samples, regardless of the curing time. This difference is of 57.9%, 64.3% and 58.2% for 3 wt% binder content, 52.3%, 54.3% and 51.8% for 4.5 wt% binder content, and 50.5%, 27.1% and 19.8% for 7 wt% binder content (Fig. 5 a, b, c). The highest UCS gap was obtained for CPB with 3 wt% binder content. One of the probable reasons for such a high UCS variation, could be the high drainage of excess water from these mixtures during the consolidation process, which reduces the corresponding porosity and permeability. During the drainage, the stiffness of consolidated samples increases as a function of increasing density.

Another reason is that the hardening of CPB materials being cured under different effective stresses greatly contributes to the strength development of backfill. If a series of effective stress is applied just before the commencement of curing which contributes to CPB hardening, as is being practiced in the present investigation, the rate of strength increase (short-term property), ultimate strength (mid-term property) and durability (long-term property) will be greater, compared to the application of pressure after the completion of hydration. It has been experimentally shown from Belem et al. (2002) and Yilmaz et al. (2006, 2008a) that the primary bonds in cement in CPB may be broken due to hydrates damage caused by the applied pressure (overloading) in early stages of strength development, which thus reduces the strength and stiffness.

3.4 Characterization of CPB geotechnical index parameters

Figure 6 shows the variation of the geotechnical index (bulk properties) parameters of CUAPS-consolidated and conventional plastic mould-unconsolidated CPB samples containing a binder content of 3, 4.5 and 7 wt% as a function of curing time. The measured geotechnical index

Figure 5. Comparison of CPB strength: a) 3 wt%, b) 4.5 wt%, c) 7 wt%, and d) mixed results.

0

500

1000

Binder content (wt%)

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

Mould @ 7 daysMould @ 14 daysMould @ 28 days

0

500

1000

1500

2000

2500

3000

3500

4000

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

(a) CUAPS @ 3.0 wt%

Mould @ 3.0 wt%

0

500

1000

1500

2000

2500

3000

3500

4000

Curing time (days)

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

(c)

CUAPS @ 7.0 wt%

Mould @ 7.0 wt%

3500

4000

(b) CUAPS @ 4.5 wt%

Mould @ 4.5 wt%

51.8%

02 3 4 5 6 7 80 7 14 21 28 35

0 35

500

1000


Co

mp

res

siv

e s

tre

ng

th (

kP

a)



0

500

1000

1500

2000

2500

3000

3500

4000

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

(a) CUAPS @ 3.0 wt%

Mould @ 3.0 wt%

0

500

1000

1500

2000

2500

3000

3500

4000

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

0

500

1000

1500

2000

2500

3000

3500

4000

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

(a) CUAPS @ 3.0 wt%

Mould @ 3.0 wt%

CUAPS @ 3.0 wt%

0

500

1000

1500

2000

2500

3000

3500

4000

Curing time (days)

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

(c)

CUAPS @ 7.0 wt%

Mould @ 7.0 wt%

0

500

1000

1500

2000

2500

3000

3500

4000

Curing time (days)

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

0

500

1000

1500

2000

2500

3000

3500

4000

Curing time (days)

Co

mp

res

siv

e s

tre

ng

th (

kP

a)

(c)

CUAPS @ 7.0 wt%

Mould @ 7.0 wt%

CUAPS @ 7.0 wt%

Mould @ 7.0 wt%

3500

4000

(b) CUAPS @ 4.5 wt%

Mould @ 4.5 wt%

51.8%

3500

4000

(b) CUAPS @ 4.5 wt%

Mould @ 4.5 wt%3500

4000

3500

4000

(b) CUAPS @ 4.5 wt%

Mould @ 4.5 wt%

CUAPS @ 4.5 wt%

Mould @ 4.5 wt%

52.3%

54.3% 51.8%

Difference:


146

parameters are as follows: the gravimetric water content wg (%), dry density ρ

d (g/cm3), specific

surface Ss (m2/kg), degree of saturation S

r (%) and void ratio e.

There is a noteworthy variation in water content, dry density and void ratio. The CUAPS-consolidated CPB samples have a w

g value varying between 18.8−11.2 wt%, a ρ

d value vary-

ing between 2.13−2.08 g/cm3, a void ratio e value varying between 0.7−0.57 while conventional mould-unconsolidated CPB samples have a w

g value varying between 25.2−17%, a ρ

d value vary-

ing between 1.85−1.87 g/cm3, a void ratio e value varying between 0.95−0.71. In addition, the final water content w

g and degree of saturation S

r are highly influenced by the drainage ability of

the CPB material. The highest binder content (7 wt%) exhibits the lowest wg and S

r. This could

be basically explained by the different initial w/c ratios (see Table 4) and the amounts of water required for binder hydration.

It was also shown by a number of authors (Belem et al. 2002, 2006; Benzaazoua et al. 2004; Yilmaz et al. 2006; 2008a,b) that drainage of the excess water existing within the CPB sam-ple contributes to a better hardening process (water trapped during the cement hydration and/or hydrated phase precipitation as gypsum) for CPB material and consequently, the reduction of total void ratio of backfilling. Moreover, the rate of water drainage is significantly affected by the settling (i.e. improved consolidation and/or density increase) of CPB sample being cured under a series of pressure during the first step of curing, and so higher strengths.

Moreover, specific surface area Ss is closely related to the hardening phase formation dur-

ing curing, and the ultimate mechanical strength. The higher the binder content used, the greater the overall S

s value of the CPB due to hydrate growth, becomes. The UCS values increase with

the increasing Ss values for a given binder content, Overall, one can say from these tests that the

CUAPS−consolidated CPB samples give lower wg and S

r, and higher UCS than those obtained

0.50

0.55

0.60

0.65

0.70

0.75

0.80

0.85

0.90

0.95

1.00

2 3 4 5 6 7 8


Vo

id r

ati

o e

(d)


CUAPS @ 7 daysCUAPS @ 14 daysCUAPS @ 28 days

10

12

14

16

18

20

22

24

26

2 3 4 5 6 7 8


Gra

vim

etr

ic w

ate

r co

nte

nt

(wt%

) (a)



65

70

75

80

85

90

95

100

2 3 4 5 6 7 8


De

gre

e o

f s

atu

rati

on

(%

)

(c)



4

5

6

7

8

9

10

11

12

2 3 4 5 6 7 8


Sp

ec

ific

su

rfac

e a

rea

(m

2/g

)

(b)Mould @ 7 daysMould @ 14 daysMould @ 28 days


2 3 4 5 6 7 82 3 4 5 6 7 8

(d)





(a)



(a)





(c)











Figure 6. Evolution of CPB geotechnical index parameters as a function of binder content and curing time.


147

using conventional mould-unconsolidated CPB samples due to the variation in the initial void ratio e

0 and the influence of time−dependent consolidation settlement.

4 DISCUSSION

4.1 Correlation between the mechanical strengths

From the compression test results shown in Figure 7, it seems that a strong correlation exists between the UCS obtained from CUAPS-consolidated CPB samples (UCS

CUAPS) and the UCS

obtained from conventional mould-unconsolidated CPB samples (UCSmould

). After normalizing the UCS values for binder content and curing time, it was possible to observe the correlation between the two sets of UCS data. This normalization consisted of multiplying each UCS variable by the binder content/curing time ratio (UCS*B

w%/t). Figure 7 shows the relationship between

UCSCUAPS

and UCSmould

. It should be noted that at first approximation, it appears that this relation-ship could be well-described by an exponential function such as y = a*exp(b*x). Indeed, a coeffi-cient of correlation R = 0.98 was obtained, thus indicating the strong relationship existing between the two types of UCS variables.

The exponential relationship describing the correlation between these two types of variables (UCS

CUAPS, UCS

mould) is as follows:

UCS UCSCUAPS mould× ⎛

⎝⎜⎞⎠⎟

= × × ×⎡⎣⎢

⎤⎦

B

t

B

tw w% %. exp .360 952 0 001399 ⎥⎥

⎛⎝⎜

⎞⎠⎟

(1)

where Bw%

= binder content (wt%); t = curing time (day); UCS values are in kPa.Finally, the predictive CUAPS-consolidated backfill samples equation from known mould-

unconsolidated backfill samples is given as follows:

UCS UCSCUAPS mould× ×⎛

⎝⎜⎞⎠⎟

× × ×⎛⎝⎜

⎞⎠

360 952 0 001399.exp

.

%

%t

B

B

tw

w ⎟⎟

(2)

0

500

1000

1500

2000

2500

3000

3500

0 200 400 600 800 1000 1200 1400 1600

Normalized UCS_mould (= UCSmould*Bw/t)

Nor

mal

ized

UC

S_C

UA

PS

(=

UC

SC

UA

PS*B

w/t)

y = UCSCUAPS*Bw/tx = UCSmould*Bw/tCoeff. of correlation: R = 0.98


y = a*exp(bx)



y = a*exp(bx)

Figure 7. Correlation between UCSCUAPS

and UCSmould

based on binder content and curing time.


148

The above equations are specific to the CPB tested and to the loading sequence applied in this study (see the section on one-dimensional consolidation tests). However, ongoing research on the influence of different pressure loading rates on CPB consolidation characteristics, and conse-quently on its strength development, is underway. After these tests, it would be possible to develop a more general equation that could be used to improve prediction of CPB strength.

4.2 Chemical analysis of drainage water

Table 5 shows the variation of chemical analysis of drainage water obtained from CUAPS-consolidated CPB samples during 1-D consolidation tests. Overall, the calcium Ca, silicon Si and sulphates SO

42- concentrations of CPB samples containing 3, 4.5 and 7 wt% binder content

decrease with increasing curing time. The variation in terms of chemical composition can be explained by the different chemical compositions of both binder and tailings pore water and tap water (see Table 3) used for CPB mixture, relating to the hydration and precipitation process at early ages. However, the variation is also due to the fact that the denser the cementitious matrix is, the more this later act as a filter of colloidal species from the pore waters.

It can be interpreted from Table 5 that, for a given binder content, calcium Ca concentration is higher at early ages (7 days) and begins to reduce appreciably when the curing time is increased up to 28 days. The reason for this phenomenon can be explained by the dissolution of Ca-bearing mineral within the cement during early hydration stages. At this time, Ca remains trapped in porous media. The dissolved Ca is expulsed from the sample along with drainage water as a func-tion of pressure. In a similar manner, Si concentration is reduced with increased curing time, but at lower rates than Ca, due to the fact that Si occurs as gel phases in solid form. One can also say that sulphate concentration is relatively higher (6692, 6390, and 6279 ppm for 3, 4.5 and 7 wt% binder, respectively at 7 days of curing) than the one observed from the tailings pore water (4883 ppm, also Table 3). This means that sulphate concentration increases during mixing at early curing ages due to aeration contributing to sulphide oxidation by oxygen intake. Also, the sulphates can come from the sulphated mineral dissolution from the binder.

4.3 Geochemical analysis of drainage water

Figure 8 shows the variation in the geochemical analyses of water collected from the CUAPS-con-solidated CPB samples containing a binder content of 3, 4.5 and 7 wt%. The measured parameters are pH, redox potential (Eh) and electrical conductivity (EC).

For all CPB mixtures, as the amount of binder used in the tailings material sample increases, the pH and EC parameter decreases as a function of curing time. However, the Eh value increases with increasing binder content and curing time. In terms of magnitude, the highest pH change (from 11.5 at 7-day curing to 7.9 at 28-day curing) is observed for the sample containing 7 wt%

Table 5. Chemical composition of water collected from CUAPS-consolidated CPB samples.

Binder Curing Al Ca Fe Mg Si SO42–

content (day) (ppm) (ppm) (ppm) (ppm) (ppm) (ppm)

3.0 wt% 7 0.03 436 0.07 0.08 13.3 6692 14 0.02 429 0.13 0.04 12.5 6411 28 0.03 284 0.15 0.05 11.7 4945

4.5 wt% 7 0.01 420 0.06 0.02 15.0 6390 14 0.02 302 0.12 0.05 14.0 6100 28 0.03 247 0.17 0.01 13.5 3056

7.0 wt% 7 0.01 412 0.08 0.02 28.0 6279 14 0.03 251 0.13 0.78 27.5 5572 28 0.04 194 0.22 0.18 26.7 2885


149

binder content (Fig. 8a). This shows that pH decreases with increasing binder content and curing time. Furthermore, the Eh values, as shown in Figure 8b, indicate that the CUAPS-consolidated CPB sample that contains 7 wt% binder acts differently while a binder content of both 3 wt% and 4.5 wt% acts in the same manner. After 7 days of curing, all Eh values are almost the same (0.141, 0.148, 0.154 volts for 3, 4.5 and 7 wt% binder contents, respectively).

In addition, pH reduces while Eh increases with increased curing time, due to the enhanced reactivity of sulphide during the sample preparation and mixing. Oxidation of pyrite accelerates acidity (i.e. the decrease of pH) and gives rise to a loss of strength over time. Degree of saturation S

r is closely related to the Eh parameter. Eh increases with the decrease of S

r which contributes to

the increase of oxidation by molecular oxygen diffusing through CPB samples.At the highest binder content (7 wt%), the S

r value is much lower compared to other binder

contents (3 and 4.5 wt%) and the Eh value becomes higher, as can be seen in Figure 8b. However, as curing time increases, the gap observed between 7 and 4.5 wt% binder contents increases from 0.17 to 0.25 after 14 days, from 0.2 to 0.26 after 28 days. The electrical conductivity EC value decreases from 6.5 to 6.2 mS/cm, from 5.9 to 5.5 mS/cm, and from 5.3 to 4.4 mS/cm when curing time is increased from 7 to 28 days (Fig. 8c). Also, Eh versus pH curve, as indicated in Figure 8d, shows the drainage water exhibits an oxidizing and alkaline medium.

5 CONCLUSIONS

This paper presents comparative results of laboratory testing conducted in order to better under-stand the hydromechanical and geochemical properties of consolidated (using a new laboratory apparatus called CUAPS) and unconsolidated (by conventional plastic moulds) CPB samples.

Figure 8. Geochemical analysis of drainage water: a) pH, b) Eh, c) EC, and d) pH vs. Eh.

7

8

9

10

11

12

13

14

0 7 14 21 28 35Curing time (days)

pH

of

dra

ina

ge

wa

ter

3.0 wt%

4.5 wt%

7.0 wt%

0.1

0.15

0.2

0.25

0.3


Eh

of

dra

ina

ge

wa

ter

(V)

3.0 wt%

4.5 wt%

7.0 wt%

0.1

0.15

0.2

0.25

0.3

0 2 4 6 8 10 12 14pH

Eh

(V

)

Oxidizing and alkaline zone

4

4.5

5

5.5

6

6.5

7


EC

of

dra

ina

ge

wa

ter

(mS

/cm

)

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

)b()a(

)d()c(

7

8

9

10

11

12

13

14


pH

of

dra

ina

ge

wa

ter

7

8

9

10

11

12

13

14


pH

of

dra

ina

ge

wa

ter

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

0.1

0.15

0.2

0.25

0.3


Eh

of

dra

ina

ge

wa

ter

(V)

0.1

0.15

0.2

0.25

0.3


Eh

of

dra

ina

ge

wa

ter

(V)

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

0.1

0.15

0.2

0.25

0.3

0 2 4 6 8 10 12 1pH

Eh

(V

)

Oxidizing and alkaline zone

4

4.5

5

5.5

6

6.5

7


EC

of

dra

ina

ge

wa

ter

(mS

/cm

)

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

3.0 wt%

4.5 wt%

7.0 wt%

)b()a(

)d()c(


150

From the tests performed, one can conclude that the one-dimensional consolidation of CPB samples being cured under a series of pressure increments is essential to a better understanding of events linked to placement and curing conditions occurring in underground stopes during the backfilling process. The CUAPS-consolidated backfill samples always give higher mechanical strengths than those obtained from conventional plastic moulds (undrained) for a given binder content and curing time. Geotechnical index parameters show that void ratio is dramatically reduced when CPB samples are prepared using CUAPS cells that allow drainage during a curing time of 7, 14 and 28 days. The reduction of void ratio gives rise to a stiffer, and consequently more resistant, CPB material. From the chemical composition analysis, it can be pointed out that the concentrations of sulphate SO

42- continuously decrease over time, depending on binder content

used in the CPB mixture. Geochemical analyses have also shown that drainage water exhibits an alkaline to neutral oxidizing medium, based on the relationship between pH and Eh.

These results showed that CUAPS apparatus gives higher CPB mechanical strength than UCS values from laboratory−prepared CPB samples (often considered an underestimation of the true value). Consequently, more realistic mechanical strength results can be obtained using this new apparatus. More work is underway to validate the values obtained by the CUAPS apparatus and in situ results.

ACKNOWLEDGEMENTS

This research was partly financed by the Canada Research Chair on “Integrated Management of Sulphide Mine Waste using Fill Technology” and the Industrial NSERC-Polytechnique-UQAT Chair on “Environment and Mine Wastes Management”. The authors would like to express their appreciations to the Canadian Foundation for Innovation for the financial support provided for designing and manufacturing the CUAPS setups used in the present study. Special thanks go to URSTM chemists and technicians, principally Mélanie Bélanger and Mélinda Gervais for their technical support on a range of chemical and geochemical analyses.

REFERENCES

Amaratunga, L.M. & Yaschyshyn, D.N. 1997. Development of a high modulus paste fill using fine gold mill tailings. Geotechnical and Geological Engineering 15(3): 205–219.

Archibald, J.F., Chew, J.L. & Lausch, P. 1999. Use of ground waste glass and normal Portland cement mix-tures for improving slurry and paste backfill support performance. CIM Bulletin 92(1030): 74–80.

ASTM Designation D2435, 1999. Standard test method for one-dimensional consolidation properties of soils. Annual Book of ASTM Standards, 04.08, American Society of Testing Material, MD, 207–216.

ASTM Designation D4186, 1999. Standard test method for one-dimensional consolidation properties of soils using controlled-strain loading. Annual Book of ASTM Standards, 04.08, American Society of Testing Material, Easton, MD, 477–481.

ASTM Designation D806, 1999. Standard test method for cement content of soil-cement mixtures. Annual Book of ASTM Standards, 04.08, American Society of Testing Material, Easton, MD, 85–87.

ASTM Designation C143~

, 1999. Standard test method for slump of hydraulic cement concrete. Annual Book of ASTM Standards, 04.01, American Society of Testing Material, Easton, MD, 68–76.

ASTM Designation C39, 1999. Standard test method for compressive strength of cylindrical concrete speci-mens. Annual Book of ASTM Standards, 04.02, American Society of Testing Material, 15–23.

Belem, T., El Aatar, O., Bussière, B., Benzaazoua, M., Fall, M. & Yilmaz, E. 2006. Characterization of Self weight consolidated paste backfill. In R. Jewell, S. Lawson & P. Newman (eds), Proceedings of the 9th International Seminar on Paste and Thickened Tailings, April 3–7, Ireland, ACG, 333–345.

Belem, T. & Benzaazoua, M. 2008a. Design and application of underground mine paste backfill technology. Geotechnical and Geological Engineering 26(2): 147–174.

Belem, T. & Benzaazoua, M. 2008b. Predictive models for prefeasibility cemented paste backfill mix design. In International Symposium on Post-Mining 2008, February 6–8, Nancy, France, Research Group for the Impact and Safety of Underground Works, 1–13.


151

Belem, T., Benzaazoua, M. & Bussière, B. 2000. Mechanical behaviour of cemented paste backfill. In D. Leb-oeuf (ed.), Proceedings of the 53rd Canadian Geotechnical Conference and the 1st Joint IAH-CNC/CGS Conference, September 18–20, Montréal, Canadian Geotechnical Society, 373–380.

Belem, T., Benzaazoua, M., Bussière, B. & Dagenais, A.M. 2002. Effects of settlement and drainage on strength development within mine paste backfill. In L. Hinshaw (ed.), Proceedings of the 9th International Conference on Tailings and Mine Waste, January 27–30, Vail, Balkema: Rotterdam, 139–148.

Belem, T., El Aatar, O., Benzaazoua, M., Bussiere, B. & Yilmaz, E. 2007. Hydro-geotechnical and geochemi-cal characterization of column consolidated cemented paste backfill. In: F. Hassani & J. Archibald (eds), Proceedings of the 9th International Symposium in Mining with Backfill, April 29–May 2, Montreal, Quebec, Canadian Institute of Mining, Metallurgy and Petroleum, 1–10 (No 2523).

Benzaazoua, M., Ouellet, J., Servant, S., Newman, P. & Verburg, R. 1999. Cementitious backfill with high sulfur content: physical, chemical and mineralogical characterization. Cement and Concrete Research 29(5): 719–725.

Benzaazoua, M., Belem, T. & Bussiere, B. 2002. Chemical factors that influence the performance of mine sulphidic paste backfill. Cement and Concrete Research 32(7): 1133–1144.

Benzaazoua, M., Belem, T. & Yilmaz, E. 2006. Novel lab tool for paste fill. Can. Min. J. 127(3): 31–31.Benzaazoua, M., Fall, M. & Belem, T. 2004. A contribution to understanding the hardening process of

cemented pastefill. Minerals Engineering 17(2): 141–152.Bussière, B. 2007. Colloquium 2004: Hydrogeotechnical properties of hard rock tailings from metal mines

and emerging geoenvironmental disposal approaches. Canadian Geotech. J. 44(9): 1019–1052.Cayouette, J. 2003. Optimization of the paste backfill plant at Louvicourt mine. CIM (Canadian Institute of

Mining, Metallurgy and Petroleum) Bulletin 96(1075): 51–57.Fourie, A., Helinski, M. & Fahey, M. 2006. Filling the gap—a geomechanics perspective. Australian Centre

for Geomechanics Newsletter 26(5): 1–4.Godbout, J. 2005. Évolution des propriétés hydriques des remblais miniers cimentés en pate durant le curage.

M.Sc. Thesis, Université de Montréal École Polytechnique, Québec, 1–190.Grabinsky, M. & Simms, P. 2006. Self-desiccation of cemented paste backfill and implications for mine

design. In R. Jewell, S. Lawson & P. Newman (eds), Proceedings of the 9th International Seminar on Paste and Thickened Tailings, April 3–7, Limerick, Ireland, ACG, 323–332.

Hassani, F. & Archibald, J. 1998. Mine backfill Hand Book, Montreal, Quebec, Canadian Institute of Mining, Metallurgy and Petroleum, 1–263.

Hassani, F.P. & Bois, D. 1992. Economic and technical feasibility for backfill design in Quebec underground mines. Final report 1/2, Canada—Quebec Mineral Development Agreement, Research and Development in Quebec Mines. Contract No. EADM 1989–1992, File No. 71226002.

Helinski, M., Fahey, F. & Fourie, A.B. 2007. Numerical modelling of cemented paste backfill deposition. Journal of Geotechnical and Geoenvironmental Engineering ASCE 13(10): 1308–1319.

Kesimal, A., Yilmaz, E. & Ercikdi, B. 2004. Evaluation of paste backfill test results obtained from different size slumps with varying cement contents for sulphure rich tailings. Cement and Concrete Research 34(5): 1817–1822.

Kesimal, A., Yilmaz, E., Ercikdi, B., Alp, I. & Deveci, H. 2005. Effect of properties of tailings and binder on short-and long-term strength and stability of cemented paste backfill. International Journal of Materials Letters 59(28): 3703–3709.

Klein, K. & Simon, D. 2006. Effect of specimen composition on the strength development in cemented paste backfill. Canadian Geotechnical Journal 43(3): 310–324.

Landriault, D. 2001. Backfill in underground mining: Underground mining methods engineering fundamen-tals and international case studies. In W.A. Hustrulid & R.L. Bullock (eds), Littleton, Colorado, Society for Mining, Metallurgy and Exploration, Chapter 69, 601–614.

Le Roux, K.A., Bawden, W.F. & Grabinsky, M.W.F. 2005. Field properties of cemented paste backfill at the Golden Giant mine. Mining Technology: IMM Transactions section A, 114(2): 65–80.

Mohamed, A., Hossein, M. & Hassani, F. 2001. Hydromechanical evaluation of stabilized mine tailings. International Journal of Environmental Geology 41(7): 749–759.

Ouellet, S. 2007. Mineralogical characterization, microstructural evolution and environmental behaviour of mine cemented paste backfills. PhD Dissertation, University of Quebec at Abitibi-Témiscamingue (UQAT), Rouyn–Noranda, Quebec, 1–310.

Revell, M.B. 2004. Paste—How strong is it? In Z. Aimin (ed.), Proceedings of the 8th International Sympo-sium on Mining with Backfill, September 19–21, Beijing, China, Metals Society of China, 286–294.

Servant, S. 2001. Détermination des paramètres mécaniques des remblais miniers faits de résidus ciments. Master’s Thesis, McGill University, Montreal, Quebec, 1–53.


152

Stone, D.M.R. 1993. The optimization of mix designs for cemented rockfill. In H.W. Glen (ed.), Proceedings of the 5th International Symposium on Mining with Backfill, October 15–20, Johannesburg, SA, South African Institute of Mining and Metallurgy, 249–253.

Yilmaz, E. 2003. Investigation of the strength characteristics of the cemented paste backfill samples prepared from sulphide-bearing mine tailings. Master’s Thesis, Black Sea Technical University (KTU), Trabzon, Turkey, 1–117.

Yilmaz, E. Kesimal, A. & Ercikdi, B. 2003. Strength properties in varying cement dosages for paste backfill samples. In L. Hinshaw (ed.), Proceedings of the 10th International Conference on Tailings and Mine Waste, October 12–15, Vail, Fort Collins, Colorado, Balkema: Rotterdam, 109–114.

Yilmaz, E., Benzaazoua, M., Belem, T. & Bussiere, B. 2008a. Influence of applied pressure on the hydro-mechanical properties of cemented paste backfill. Geotechnical and Geological Engineering (submitted on 14 January 2008).

Yilmaz, E., Belem, T., Bussiere, B. & Benzaazoua, M. 2008b. Consolidation characteristics of early age cemented paste backfill. In Proceedings of the 61st Canadian Geotechnical Conference and 9th Joint IAH-CNC/CGS Conference, September 21–24, Edmonton, Alberta, CGS, (accepted for presentation).

Yilmaz, E., El Aatar, O., Belem, T., Benzaazoua, M. & Bussiere, B. 2006. Effect of consolidation on the performance of cemented paste backfill. In Proceedings of the 21st Annual Underground Mine Support Conference, April 11–12, Val d’Or, Quebec, 1–14.


153

Guidelines for stabilizing historic mine workings

J.F. LupoAMEC Earth and Environmental, Englewood, CO, USA

ABSTRACT: Historic mine workings are present throughout several regions of the United States. The presence of these workings can pose a significant challenge to engineering projects requiring competent foundation conditions, such as roads, buildings, dams, etc. Mine workings such as shafts, open stopes, drainage tunnels, and drifts may leave unstable voids in the subsurface that may become public safety hazards, compromise foundation stability, and adversely impact engineered structures. In order to stabilize historic mine workings, geotechnical investigations and engineering analyses are required to assess site conditions and the associated risk to nearby or proposed structures. The design of a suitable stabilization approach to historic mine workings needs to consider many facets of rock and soil mechanics, including:

• Existing and induced stress state within the region of the mined workings;• Rock mass quality;• Crown pillar stability;• Cavability of the surrounding rock mass; and• Settlement of backfilled materials.

This paper presents practical guidelines for the stabilization of historic mine workings. The paper also discusses important implementation and constructability issues which can affect the success of the project.

1 INTRODUCTION

Throughout several regions of the United States there are numerous historic mine workings that have been left abandoned without proper stabilization or in-place controls to prevent access. The presence of these workings can pose a significant challenge to engineering projects requiring competent foundation conditions, such as roads, buildings, dams, pipelines, etc. Mine workings such as pits, shafts, open stopes, drainage tunnels, and drifts may become unstable over time and present public safety hazards, compromise foundation stability, and have adverse impacts to engi-neered structures, as illustrated in Figures 1 through 4.

While there are numerous methods that may be used to stabilize historic mine workings, the selected stabilization method must consider a number of factors, including:

• Type of mine workings• Size, shape, and depth of workings• Type of existing mine support (if present)• Rock mass characteristics• Characteristics of the overburden materials (soil, waste rock, etc)• Depth of groundwater (perched, permanent, confined and unconfined conditions)• Surface water drainages• Environmental issues (acid drainage, gas production, etc).



154

Figure 1. Surface collapse over mine shaft.

Figure 2. Small surface collapse over mine working.


155

Figure 3. Large-scale surface collapse.

Figure 4. Large-scale surface subsidence.


156

This paper presents practical guidelines for the stabilization of historic mine workings. The paper also discusses important implementation and constructability issues which can affect the success of the project.

2 FIELD INVESTIGATIONS FOR MINE WORKINGS

An important aspect to stabilizing historic mine workings, is understanding the different types of workings and the ground conditions around the workings. In general terms, mine workings can be categorized into two groups, surface workings and underground workings. Even though these workings may be related to each other, the method of stabilization will differ for each group.

Surface mine workings may consist of:

• Open pits• Waste rock piles• Tailings ponds or impoundments• Surface water/sedimentation ponds• Structures (head frames, process buildings, etc).

Underground mine workings may include:

• Shafts (manway, ore, and ventilation)• Stopes• Rooms• Drifts• Drainage tunnels• Winzes.

Surface mine workings are generally easier to understand, since they are entirely exposed at the ground surface. Therefore, stabilization methods can be planned and executed with only limited amount of uncertainty and risk. On the other hand, underground mine workings are more difficult to understand since only portions of the workings may be exposed, or have no surface expression at all. Given these conditions, it is difficult to assess the overall stability of the workings and the extent of workings beneath the surface.

Field investigations for surface mine workings generally focus on site reconnaissance and surveying to locate the type and extent of surface working; test pit excavation and sampling for material characterization, and possibly geotechnical drilling to assess slope stability for open pits.

Field investigations for underground mine workings are typically more extensive, using a com-bination of geotechnical and geophysical methods. Geotechnical field investigation methods often include:

• Geotechnical drilling through overburden• Corehole drilling through rock• Exploratory drilling on grid• Test pit excavation.

While applicable geophysical methods may include:

• Resistivity surveys• Video surveys• Gas monitoring• Ground penetrating radar• multi-channel analysis of surface waves (MASW)• Surface high-resolution seismic


157

• Crosshole seismic tomography• Reverse vertical seismic profiling• Borehole sonar mapping• Radio imaging method.

It is critical that the field investigation program be carefully designed to gather information that will have direct influence on the method of stabilization. Table 1 presents common field investiga-tion methods for some underground mine workings.

It is important to note that not all geophysical methods work for all mine workings. The size, shape, and characteristics of the overburden and rock mass may influence the effectiveness of the method. It is often worth trying more than one geophysical method to assess the presence on mine workings.

For the most part, the mine working shape (depth, width, and lateral extent) and the charac-teristics of materials around the working generally play a primary role in stabilization method selection and design. The mine working shape may be defined using a combination of drilling (geotechnical, corehole, and exploratory) and geophysical methods. Often geophysical surveys are conducted to provide a first pass on locating the mine workings. This is followed by explora-tory drilling with an air-track rig to confirm the location and extent of the working. Geotechnical and corehole drilling may then be used to collect samples for material characterization testing.

Table 1. Common field investigation methods for underground mine workings.

Mine Working Investigation Type

Shafts • Geotechnical drilling around shaft to assess stability• Video survey to evaluate the size of opening, condition of existing shaft

support, location of drifts that may intercept shaft at depth, and presence of groundwater

• Geophysical surveys combined with exploratory drilling to assess depth and extent of drifts that may intercept the shaft at depth

• Gas monitoring for the presence of noxious gases

Stopes • Drilling (geotechnical, coring, exploratory) to assess thickness and condition of rock mass in crown pillar

• Drilling combined with geophysical surveys to evaluate width and lateral extent of stope

• Video survey to evaluate the size and shape of stope, and conditions within the stope (open, collapsed, etc)

Drainage Tunnels • Drilling (geotechnical, coring, exploratory) to assess thickness and condition of rock mass in crown pillar

• Drilling combined with geophysical surveys to evaluate width and lateral extent of tunnel

• Video survey to evaluate condition of tunnel, existing tunnel support, and drainage water discharge rate


Rooms • Drilling (geotechnical, coring, exploratory) to assess thickness and condition of rock mass in crown pillar

• Drilling combined with geophysical surveys to evaluate width and lateral extent of room

• Video surveys to evaluate the room size, condition of the room, and existing ground support. In addition, the video survey can be used to assess the condi-tion of support pillars.



158

3 MATERIAL CHARACTERIZATION

Material characterization is required to not only classify materials, but to also develop estimates of engineering properties that can be integrated into the stabilization method and design. It is often convenient to classify materials around mine workings as either overburden (soil and waste rock) or as rock mass, as the method of characterization is different for each material as described in the following sections.

3.1 Overburden Characterization

Characterization of overburden materials rely on obtaining representative samples of materials from test pits and geotechnical drilling programs. These samples are then subjected to a suite of laboratory tests to characterize the materials and obtain estimates of engineering properties. A typical laboratory testing program would consist of the following:

• Particle size• Atterberg limits• In-place density and moisture content• In-place shear strength from triaxial compression or direct shear testing on relatively undis-

turbed samples• In-place permeability from laboratory tests on relatively undisturbed samples• Remolded moisture-density testing if the overburden is to be used in the stabilization program• Shear strength from triaxial compression or direct shear testing on remolded samples• Permeability testing on remolded samples.

For coarse-rockfill (e.g. waste rock), laboratory tests may not be practical. For these materials, empirical methods are often used to estimate engineering properties. Some practical empirical meth-ods are presented in Leps (1970), Donaghe & Cohen (1978), De Mello (1977), Barton & Kjaernsli (1981), Marachi & Chan (1972), Matheson (1986), Matheson & Parent (1989), and Hunter & Fell (2003). These empirical methods are based on experience with large rockfill dams and waste rock piles, but are very useful for estimating the engineering properties for rockfill overburden.

3.2 Rock Mass Characterization

The characterization and development of engineering properties for rock masses is based on empirical methods used in the mining industry for the design of open pits and underground mines. Rock mass characterization methods are typically based on either the Rock Mass Rating (RMR) system developed by Bieniawski (1976) or the Q Index developed by Barton et al (1974).

The RMR system is a linear empirical method with assigned numerical values based on six rock mass parameters:

• uniaxial compressive strength of rock material,• rock quality designation (RQD),• spacing, condition, and orientation of discontinuities, and• groundwater conditions.

The sum of the six numerical values, with the addition of an adjustment based on the strike and dip orientation of the discontinuities, is then categorized into five rock mass classes ranging from very good to very poor rock. Each class has a range of assigned strength parameters (cohe-sion and friction angle) and average stand-up times for assumed life-of-design use. The RMR was originally drawn from civil engineering case histories, but it is still in general use for both mining and civil projects (Bieniawski, 1989). Several modifications have been suggested to make it more applicable to the mining environment, including Bieniawski’s (1989) guidelines for support of tunnels in rock, Laubscher’s (1977, 1984) MRMR or RMRL

system and Cummings’ (1982) MBR or modified basic RMR system.


159

The Q system is a logarithmic empirical method defined as:

QRQD

J

J

J

J

SRFn

r

a

w= * *

(1)

where:

RQD is the rock quality designation, J

n is the joint set number,

Jr is the joint roughness number,

Ja is the joint alteration number,

Jw is the joint water reduction factor, and

SRF is the stress reduction factor.

The three quotients represent the block size (RQD/Jn), interblock shear strength (J

r/J

a), and

active stress for the rock mass (Jw/SRF), respectively (Hoek, 2000). The numerical values for the

J variables are assigned from tables and are based on field observations.The Q rating is assigned in structural regions, similar to the RMR, as defined by the RQD

intervals. Ranging from 0.001 to 1,000, Q is then categorized into one of nine rock mass catego-ries ranging from exceptionally poor to exceptionally good rock. To translate this Q into support guidelines for underground excavations, Barton et al (1974) defined an equivalent dimension, D

e,

to incorporate the final design use into the support guidelines, using the Excavation Support Ratio (ESR), which provides a correlation to type of mine working.

While both the RMR and Q system are very useful for characterizing rock mass and providing estimates of engineering properties, the Hoek-Brown Method (Hoek & Brown, 1980a, b, 1988, Hoek, 1990, Hoek, 1994, and Hoek et al, 2002) may also be used. This method requires the fol-lowing parameters to be estimated:

• Intact rock strength• Hoek-Brown constant m

i based on rock type

• Geological Strength Index (GSI) based the structures observed in the rock mass.

Using a series of empirically-based ratings, the method provides an estimate of the rock mass strength and deformation modulus, which can be used as input into the stabilization design.

The estimates of rock mass engineering properties from the RMR, Q, and Hoek-Brown methods can be used to evaluate the stability of mine workings. However, it must be real-ized that the rock mass properties will likely change over time as exposure to moisture and freeze-thaw conditions will degrade the rock mass, reducing its strength. Therefore, stability assessments and designs should take into account changing conditions over time due to rock mass weathering.

4 STABILIZATION GUIDELINES

As presented in Lupo and Eddy (2002) and Hutchinson et al (2002), the selection and design of a stabilization methods for mine workings should be risk based. A risk based approach requires an understanding of overall risk and uncertainties for the project, with consideration to safety, environmental, and financial risk.

A summary of risk-based stabilization methods for underground mine workings is presented in Figure 5. This summary was developed from experience on stabilization projects conducted by the author at sites involving over 400 mine workings. In general terms, the stabilization methods include the following:

• Access restriction with fencing: While technically this is not a stabilization method, it does address public safety concerns and is normally used for low risk applications only.


160

Figure 5. Stabilization guidelines chart.


161

Figure 7. Geogrid cap installation.

Figure 6. Crown pillar collapse and backfill.


162

• Backfill: This stabilization method consists of placing either non-structural or structural backfill into a collapsed mine void or excavated open pit. Non-structural backfill is defined as material placed using a very broad specification and may or may not be compacted. Non-structural backfill is normally used where uncontrolled settlement within the fill can be tol-erated. Structural backfill is defined as material that is placed under controlled conditions following a detailed specification with compaction, and is used in areas where uncontrolled settlement cannot be tolerated (e.g. under foundations, roads, etc) and is used in moderate to high risk applications.

• Collapse crown pillar with backfilling: This method of remediation (illustrated in Figure 6) is used when the crown pillar over a mine working is determined to be unstable or unacceptable for the design loads. This method involves collapsing the crown pillar by drilling and blasting, then backfilling the surface depression. The backfilled depression is then capped with either a geogrid, cemented rockfill, or concrete cap, depending risk and predicted settlement.

• Geogrid cap: This stabilization method consists of placing a structural cap over the mine work-ing. The structural cap consists of geogrid layers within compacted structural backfill (see Figure 7) and is normally used for applications where the applied surface loads will be low to moderate, although some high load applications have been used successfully. The geogrid cap is typically designed using the methods presented by Giroud et al (1990). Since the cap is not a rigid structural element, settlement may occur under load, so these caps should only be used where limited settlement is acceptable.

• Concrete/cemented rockfill caps and bulkheads: Concrete (reinforced or non-reinforced) and rockfill caps are typically used in areas with moderate to high risk, where the structural support needs to be of high quality. This stabilization method consists of constructing either a slab or tapered plug over the working. If the mine working is open at the surface and cannot be backfilled, the slab or plug must be constructed over an artificial support (I-beams, timber planking, etc). Concrete caps (Figure 8) provide a high integrity, structural element with pre-dictable properties. Cemented rockfill consists of a blend of rockfill, sand, and cement binder (usually between 3 to 15%), and provides an alternate to concrete caps. While a cemented rockfill cap or plug is less “controlled” than a concrete cap, it is typically less expensive, so a

Figure 8. Reinforced concrete cap placement.


163

thicker cap or plug can be used. Figure 9 presents a photograph of the placement of a cemented rockfill plug.

• Flowable fill: Flowable fill consists of a high slump (8 to 11 inch), lean concrete which is often used to fill drainage tunnels and drifts. A flowable fill is often used in combination with a bulkhead so that filling can be isolated to sections of the working. This method of stabilization provides a low-compressibility backfilling method that can be used for moderate to high risk applications.

In selecting a stabilization design, there are a number of important construction issues that should also be consider. These include:

− Site accessibility for the proposed stabilization method.− Maximum depth for stabilization. This often requires detailed engineering analyses to estimate

the extent of surface subsidence and potential for a “glory hole” to form by caving.− Development of a site-specific safety plan to used during construction.− Weather conditions that can affect the construction method.− Availability of materials.

Depending on the site conditions, the method of stabilization may need to be modified to account for construction issues.

5 CLOSURE

The purpose of this paper was to present practical guidelines for the stabilization of historic mine workings. The stabilization guidelines presented herein were developed based on work conducted on over 400 historic mine workings in the US. Experience from these projects has shown that successful stabilization of mine workings requires detailed field investigations and engineering

Figure 9. Cemented rockfill plug placement.


164

analyses. The ultimate design of the stabilization method must be selected based on the overall project risk, site conditions, and with consideration to construction issues.

REFERENCES

Barton, N. & Kjaernsli, B. 1981, Shear Strength of Rockfill Norwegian Geotechnical Institute Publication No. 136.

Barton, N.R., Lien, R. & Lunde, J. 1974. Engineering Classification of Rock Masses for the Design of Tunnel Support. Rock Mech. 6(4), 189–239.

Bieniawski, Z.T. 1976. Rock Mass Classification in Rock Engineering, Exploration for Rock Engineering, Proc. of the Symp., (ed. Z.T. Bieniawski) 1, 97–106. Cape Town: Balkema.

Bieniawski, Z.T. 1989. Engineering Rock Mass Classifications. New York: Wiley.Bieniawski, Z.T. 1993. Classification of Rock Masses for Engineering: The RMR System and Future Trends.

Comprehensive Rock Eng. Vol. 3.Carter, T.G., Alcott, J. & Castro, L.M., 2002. Extending applicability of the crown pillar scaled span method

to shallow dipping workings, NARMS-TAC, R. Hammah, W. Bawden, J. Curran, and M. Telesnicki (eds), University of Tortonto.

Cummings, R.A., Kendorski, F.S. & Bieniawski, Z.T. 1982. Caving Rock Mass Classification and Support Estimation, U.S. Bureau of Mines Contract Report #J0100103. Chicago: Engineers International Inc.

Donaghe, R.T. & Cohen, M.W., 1978. Strength and Deformation Properties of Rockfill U.S. Army Corp. of Engineers, Waterways Experiment Station Technical Report No. S-78–1.

De Mello, V.F.B. 1977, Reflections on Design Decisions of Practical Significance to Embankment Dams 17th Rankine Lecture, Geotechnique,27, No. 3, 281–354.

Giroud, J.P., Bonaparte, R., Beech, J.F. & Gross, B.A. 1990. Design of Soil Layer-Geosynthetic Systems Overlying Voids, Geotextiles and Geomembranes No. 9, pp. 11–50.

Hoek, E., Carranza-Torres, C. & Corkum, B. 2002. Hoek-Brown failure criterion—2002 edition, NARMS-TAC 2002, Hammah et al (eds), University of Toronto, ISBN 0 7727 6708 4.

Hoek E. & Brown, E.T. 1980a. Empirical strength criterion for rock masses, J. Geotech. Engng Div., ASCE, Vol. 106 (GT9), 1013–1035.

Hoek, E. & Brown, E.T. 1980b. Underground Excavations in Rock, London, Instn Min. Metall.Hoek, E. & Brown, E.T. 1988. The Hoek-Brown failure criterion—a 1988 update. Proc. 15th Canadian Rock

Mech. Symp. (ed. J.C. Curran), 31–38, Toronto, Dept. Civil Engineering, Univ. of Toronto.Hoek, E. 1990. Estimating Mohr-Coulomb friction and cohesion values from the Hoek-Brown failure crite-

rion, Intnl. J. Rock Mech. & Mining Sci. & Geomechanics Abstracts, Vol. 12, No. 3, 227–229.Hoek, E. 1994. Strength of rock and rock masses, ISRM News Journal, Vol. 2, No. 2, 4–16.Hutchinson, D.J., Phillips, C. & Cascante, G, 2002. Risk Considerations for Crown Pillar Stability Assess-

ment for Mine Closure Planning, Geotech. Geolog. Engng., 20, 41–63.Laubscher, D.H., 1977. Geomechanics Classification of Jointed Rock Masses—Mining Applications, Trans.

Instn. Min. Metall. 86, A1–8.Laubscher, D.H., 1984. Design Aspects and Effectiveness of Support Systems in Different Mining Condi-

tions, Trans. Instn. Min. Metall. 93, A70–A82.Leps, T.M. 1970, Review of shearing strength of rockfill, Journal of the Soil Mechanics and Foundations

Division, Vol. 96, pp. 1159–1170.Lupo, J.F. & Eddy, G. 2002. Geomechanics and construction issues related to the remediation of historic

underground workings, NARMS-TAC, R. Hammah, W. Bawden, J. Curran, and M. Telesnicki (eds), Uni-versity of Tortonto.

Marachi, N.D. & Chan, C.K., 1972. Evaluation of properties of rockfill materials, J. Soil Mechanics and Foundations Division, ASCE Vol. 98, SM1.

Hunter, G. & Fell, R. 2003. Rock modulus and settlement of concrete face rockfill dams, J. Geotechnical and Geoenvironmental Engineering, Vol. 129, No. 10.

Matheson, G.M., 1986. Relationship between compacted rockfill density and gradation, J. Geotechnical and Geoenvironmental Engineering, Vol. 112, No. 12.

Matheson, G.M. & W.F. Parent, 1989. Construction and performance of two large rockfill embankments, J. Geotechnical and Geoenvironmental Engineering, Vol. 115, No. 12.


165

Landscape design for soft tailings deposits

G. McKenna & V. CullenBGC Engineering Inc., Vancouver, BC, Canada

ABSTRACT: Stabilization and reclamation of tailings ponds is often one of the most expensive components of mine reclamation, sometimes costing more than the entire tailings costs during mine operation. There are numerous ways to reduce these costs through better tailings technology selection, design, planning, and monitoring while providing greater landscape performance and better long-term environmental performance. Careful site selection, dyke design, and technology selection are critical to creating tailings landforms that are easy to reclaim. Methods and moni-toring to minimize the volume of off-spec material are vital, along with plans to deal with the inevitable volume of soft material. As closure approaches, full geotechnical and hydrogeological characterization of the tailings are required for stabilization, cover and reclamation designs. Typi-cally all areas of the tailings landform (slopes, beaches and ponds) must be made safe and useful. Landform design is a framework that is being successfully used to manage these soft tailings deposits.

1 INTRODUCTION AND BACKGROUND

Tailings management is an important aspect of many mining projects. Today, it is recognized that the initial capital and the operating cost of tailings is a significant component of overall project economics. The potential environmental impact of tailings and process-affected water is almost always an important part of the project’s environmental assessment and operating permit. Advances in the past several decades have put tailings management front and centre for most mining operations.

However, many projects overlook or underestimate the closure costs and the time involved to stabilize and reclaim tailings ponds, in particular those areas of the ponds underlain by soft tailings.

For new mining operations, tailings technology selection has become an important aspect of the initial mine planning process. A variety of new technologies are available that can have operational benefits (in particular conservation of water, reagents, or heat) and most offer stronger, denser tail-ings for closure and reclamation. Even with this investment, reclamation of these soft tailings is often still challenging, even more so if large quantities of off-spec tailings have been produced.

Both mature and new mining operations need to better understand and plan for tailings recla-mation to meet shareholder and stakeholder expectations. This paper presents recent experience and guidance for tailings technology selection, planning, closure and reclamation and introduces “landform design” as a useful methodology.

2 STATE OF PRACTICE (CONVENTIONAL TAILINGS TECHNOLOGY)

Tailings is the mine waste produced by the extraction of minerals from host rocks or sediments by the mining industry. Ore is finely ground to liberate minerals of interest, producing sand or silt sized milled tailings. In some cases, the ore is hosted in natural sediments—processing pro-duces a tailings slurry with the same grainsize distribution to that of the ore. In most cases, the



166

milling/extraction uses large quantities of water and produces a water-mineral tailings slurry which is pipelined from the extraction plant/mill to a nearby impoundment to form a permanent tailings landform within the mine’s closure landscape. At some mines, the tailings slurry is further processed prior to deposition.

Conventional tailings technology can be defined as the pumping, pipelining, and deposition of segregating tailings slurries. Tailings slurry is deposited behind earth or rockfill dams, or can be contained by dams made from the compacted coarse fraction of the tailings themselves. The tailings slurry segregates as it is deposited, with the coarsest materials (usually sand-sized mate-rial) dropping out first and forming a beach, with the finer materials (typically silt and clay-sized materials) being carried farther down the beach, usually forming a process-water capped low-density “pond”. The tailings in this pond zone at the distal end of the beach is often referred to as fine tailings, slimes or sludge. At most mines, the fines content of the deposited tailings increases with distance down the beach.

Reclamation of the upper sandy tailings beaches is often straightforward, especially if the water table (phreatic surface) is at least one to two metres below the beach surface. Farther down the beach, where the fines contents are higher, the tailings often remains saturated, and will exhibit undrained behaviour (or even liquefaction-like behaviour during trafficking), making the beaches weak, hindering utility of mine reclamation equipment, and bringing process-affected water up into the cover being placed. The fine tailings in the pond is often fluid-like, with both very low density and very low permeability, and takes decades or longer to consolidate. Even fully consoli-dated, saturated fines-rich tailings can remain extremely weak. Consolidation settlement of soft tailings is often in the range of metres over years or decades, to tens of metres over a century for clay-rich fluid tailings.

By way of definition, soft tailings are those, that due to their low shear strength, present a chal-lenge to stabilize, cover, cap and reclaim (Jakubick & McKenna, 2001). Their very low bearing capacity requires use of special methods of stabilization before soil placement by traditional earth-moving equipment. Fluid-like soft tailings have shear strengths (or viscosities) typically measured in pascals, whereas soil-like soft tailings will have strength of a few to tens of kilopascals. The liquid limit (corresponding to a shear strength of roughly two kilopascals) is a useful dividing line between fluid-like and soil-like soft tailings. The cost of specialized methods for soft tailings stabilization are high—often 20 to 50 times the cost (per unit area) of typical mine reclamation projects. Costs to reclaim soft tailings are usually more than CDN $100,000 per hectare and can reach more than CDN $1,000,000 per hectare for fluid tailings. Thus the costs required to stabilize and reclaim soft tailings is usually tens to hundreds of millions of dollars for most mines, often comparable to the costs of tailings management during ore production.

3 CONVENTIONAL SOFT TAILINGS STABILIZATION AND RECLAMATION

Most tailings ponds have a soft tailings zone that can be remediated either by using a permanent water cap or by stabilizing and capping the tailings to form a dry landscape. The degree of analyti-cal work that is applied to the design of a stabilization program varies widely—some remediation programs receive little analysis and rely on a brute-force approach, while others receive high levels of investigation, modeling, and engineering (often including development of numerical models for consolidation and deformation). While there is a considerable body of successful stabilization and reclamation work worldwide, little is reported in the literature.

Usually a soil-like state (in terms of both strength and density) must be achieved before any capping technologies can be attempted. Stabilization of fluid tailings (or enhancing the strength of soil-like soft tailings) may include combinations of chemical amendment, crust management, hydraulic sand capping, use of vertical drains, and surcharging with capping material (with or without geofabrics and/or geogrids).

Much of soft tailings stabilization work is done by trial and error. Generally, capping is carried out through mechanical placement of a rock or sand layer over soft tailings by starting around the


167

perimeter where the tailings is strongest, and slowly progressing towards the downstream end of the deposit where the tailings are softest. Often a geotextile and/or geogrid layer is placed just prior to covering. Typically the reclamation front advances down the beach slowly over many years, allowing the softest material to strengthen due to consolidation or crusting. Costs of treating the process-affected water removed from the tailings impoundment can also be high.

4 ALTERNATIVE TAILINGS TECHNOLOGIES

Adapting a framework created by Dawson (1994), it can be useful to classify various mine tail-ings streams and products as either wet or dry and either coarse or fine. With advances in tailings technology, it may be useful to indicate some intermediate states as well (see Table 1). Typically, the fine, wet streams cause the most difficulty in terms of containment and reclamation due to their low strengths at deposition and slow rate of strength gain over time.

Because every ore body and every mining operation is unique, there is no tailings technology that is inherently better than another. As part of a technology selection process, all applicable tech-nologies should be evaluated at a screening level, and a formal decision analysis used to choose the best technology or more commonly, the best suite of technologies, based on a life-of-mine plan and economics. Selection criteria typically include operational, environmental, and economic (capital and operating costs), goals, and constraints. As mines typically operate over periods of decades, the tailings technologies are typically enhanced or even replaced over time as experience is gained.

Even with good operational practices, most tailings technologies result in some percentage of the tailings deposited being off-spec. Newer technologies such as paste, thickened tailings, com-posite tailing, and other types of tailings that involve co-disposal or other methods of chemical or mechanics beneficiation may produce deposits that are generally easier to reclaim, but off-spec tailings can unexpectedly disrupt the operation and require further treatment. Most miners under-stand that regardless of methods or technologies employed, there is almost always some volumes/areas of soft tailings that require formal design consideration and management.

5 CHALLENGES FOR SOFT TAILINGS

This section describes of the key challenges related to managing soft tailings.

5.1 Low bearing capacity for reclamation equipment

Dozers typically have bearing pressures of 50 to 100 kPa necessitating tailings shear strengths in of 20 to 30 kPa to depths of several metres for adequate bearing capacity/trafficability. Haul truck bearing pressures are approximately equal to their tire pressures (600 to 800 kPa) necessitating

Table 1. A classification of some tailings technologies.

Dry (Materials stand at angle of repose) Intermediate Wet (Slurries that form beaches and ponds)

Coarse Filtered tailings Unsaturated sand

Cyclostacked tailings Tailings sand slurry

Intermediate Fines-rich filtered tailings

Co-disposal fines/ waste rock

Co-disposal sand/clay slurrySand-rich thickened tailings

Fine Freeze-thaw tailings Evaporation-dried tailings

Centrifuged tailings Paste tailingsFines-rich thickened tailings Fine tailings (sludge)


168

tailings shear strengths in excess of 100 to 200 kPa for trafficability. Bearing capacity calculations are complicated by layering (especially due to crusting); empirical trials are required to develop safe work practices

Typically, strategies to overcome the low bearing capacity involve some combination of strength-ening the tailings, use of geogrids to spread the load, use of very small equipment, and running the equipment on thin lifts of capping material rather than directly on the equipment. Often the edge stability of these embankment layers governs the cover placement design.

Hydraulic placement of capping layers (especially of coarse tailings slurry is available) is one method to overcome some of these limitations when practical. Water capping is also often a viable alternative, provided the long-term storage of water in the reclaimed landscape is acceptable. Both of these strategies require considerable design and planning.

5.2 Consolidation and settlement

When soft tailings are saturated there is a strong upward flow of consolidation water. Every metre of settlement of the tailings involves the expression of a metre thickness of process-affected water. The quantity and quality of this water is very important to post-reclamation landscape performance—typically this water must be kept out of the root zone of the revegetation plants, and it often needs to be transmitted laterally for collection and water treatment. If the tailings is water-capped, the contribution of the seepage water to overall capping-water quality is critical. Consoli-dation rates are typically fairly fast at the start, reducing to low levels over years or decades. The consolidation rate can be modified by the use of drains (vertical band drains or underdrainage), or by controlling the permeability of the material deposited (by adjusting grainsize or use of coagu-lants or flocculents during deposition).

Settlement of the deposit can have other important impacts on landscape performance. If the tailings is contained behind a dyke, the settlement can cause large quantities of water to pond against the dyke or elsewhere, or cause deepening of any water cap. Where regulations require no ponded water, long-term settlements present unique challenges. More subtly, differential settlement can cause waterlogging of the root zone or disrupt drainage layers. Wetland performance is particu-larly susceptible to changes in the depths of ponded water due to consolidation settlements.

5.3 Gas generation

Organic or inorganic generation of gas from tailings deposits can have major impacts on the geo-chemistry of the pore waters and the effects of water or soil covers and their vegetation.

5.4 Seepage

Seepage of process-affected waters from soft tailings into dykes and into the substrates below the tailings deposits can impact surrounding ecosystems and may necessitate long-term water treatment.

5.5 Tailings slopes

While not usually a soft tailings issue, the dykes are integral to the tailings landform as a whole. Where dykes are partially or wholly constructed of tailings sands, reclamation of the downstream slopes is often extremely difficult. The tailings are usually fine-grained cohesionless sands and silts and susceptible to gully erosion, dusting, may provide a pathway for uptake of metals into vegetation, and seepage of process-affected water from the slopes or internal drains may necessi-tate long-term water treatment. The longevity of internal drains remains a concern for many dykes, especially where the drains are required to ensure geotechnical stability. While not dealt with further in this paper, these tailings slopes require the same level of care and attention for closure and reclamation than the soft tailings they contain as they are part of the same landform.


169

It should be noted that there is an industry-wide interest in moving away from tailings slurry deposition, that “dry tailings” may be able to provide an alternative, which may reduce or elimi-nate the need for tailings dams and fluid containment. The result of changes to the extraction process or processing of the tailings would result in an unsaturated tailings stream that would be conveyed or trucked to its final location. While this approach has great potential, it is not without its own challenges including cost of transportation, resaturation of tailings (leading to seepage and first-time wetting collapse), compaction costs, erosion, and other potential issues. The ability to deal with some off-spec wet tailings may also necessitate large additional capital costs. More work in this area is needed, and the use of landscape design to create acceptable reclaimed tailings landscapes and landforms remains important.

6 LANDSCAPE DESIGN FOR SOFT TAILINGS

There is no magic bullet for soft tailings reclamation. Alternative tailings technologies seldom ensure the absence of soft tailings and not all operations or climates are amenable to many of the most promising techniques. The use of enhanced consolidation such as underdrains or wick drains may not be practical or economical in many cases, especially for fine tailings rich in clay miner-als. Reprocessing of deposited tailings as a stabilization method is only rarely practical. However, the process of landscape design is starting to be used successfully to overcome these soft tailings reclamation challenges, particularly when applied to all aspects of the tailings technology selec-tion, operation, and closure.

McKenna, (2002) describes landscape design (actually, landscape engineering) as follows:

Landscape engineering is the interdisciplinary application of engineering and other applied sciences to the design and creation of anthropogenic landscapes. It differs from, but embraces traditional reclamation. It includes scientific disciplines: botany, ecology, for-estry, geology, geochemistry, hydrogeology, and wildlife biology. It also draws upon applied sciences: engineering geomorphology, landscape architecture, and mining, geotechnical, and civil engineering. Landscape engineering builds on the engineering strengths of declar-ing goals, determining initial conditions, iteratively designing, predicting performance based on knowledge of the design, monitoring performance, and adjusting designs to meet the declared goals. It builds on the strengths and history of reclamation practice. Its distinguish-ing feature is the marriage of landforms, substrates, and vegetation throughout all phases of design and construction, which previously have been kept as separate disciplines.

As most people involved in this work are not engineers, the term “landscape design” is often used instead of “landscape engineering” to be more inclusive, helping to engage the scientists and other miners necessary to the work.

The remainder of this paper describes some of the key ways that landform design can be used to successfully stabilize and reclaim soft tailings deposits, and the paper finishes with some specific steps that have proven useful in recent projects.

6.1 Up-front planning

One of the mantras of landscape design is to set reasonable landscape performance goals, then to ensure these goals are achieved through good management, design, and operation. In particular, lofty (often unobtainable, ill-defined, or high-risk) goals are to be avoided (McKenna, 2002) in favour of more modest and practical ones. As an example, promising a maintenance-free land-scape immediately after reclamation is not recommended. Instead, one might agree to design the landscape to:

− be safe for people and wildlife under the proposed land use.− avoid catastrophic breach with outflows of water or soft tailings.


170

− minimize the amount of process-affected water that impacts the root zone of vegetation.− collect and treat process-affected water treated until safe to discharge to the environment.

Often there are lengthy and heated debates about what constitutes a reasonable set of goals—this debate is an important part of the design process. If the goals are clearly set out at the beginning of the project, they can be used to guide the entire design, operation, and closure of the tailings facility. CEMA-RWG Landscape Design Subgroup (2005) provides a landscape design checklist that is used in the oil sands region of Alberta that forms a useful starting basis.

Site selection is an especially important (and generally irreversible) aspect of soft tailings man-agement. The location of the tailings deposit greatly affects the long-term geotechnical stability and the seepage performance. Closure and reclamation must be included as major factors in the site selection and layout.

Similarly, tailings technology selection is critically important. Special attention should be paid to minimizing the volume and aerial extent of soft tailings, stabilization methods, the post-reclamation consolidation and settlement rates, and methods to deal with off-spec materials. Because of the inherent uncertainty in tailings behaviour, especially prior to the commencement of mine operations, conservativism and especially flexibility should be important factors in any decision analysis or design.

Dyke design is an important aspect as well. The dyke can be designed to promote (or minimize) underdrainage (depending upon the performance goals), and should be designed to allow for long-term changes in the hydrogeology, especially degradation of internal filter drains. Minimizing the potential for catastrophic breaching in the long-term is critical as is designing the structures to require acceptable amounts of monitoring and maintenance.

As an overall comment, the tailings dykes and impoundments should be designed to be easy to reclaim, with operational plans to minimize the amount of soft tailings produced. Good up-front work, and careful monitoring and management throughout the operation and closure are indicated.

6.2 Minimizing the volume of off-spec material

Mines cannot generally afford to have tailings operations that exclusively produce tailings within a narrow band of parameters. There will be process upset conditions, startup and shutdown condi-tions, line flushing, changing ore-qualities, etc. Tailings systems need to be designed to handle certain levels of variability of product quality as well as off-spec materials. Tailings plans that involve shutting down the mill or the mine when the tailings is off-spec seldom last beyond the initial optimistic planning stages. That said, some operators use online tailings analyzers to control the extraction operations, but this is an exception to the rule.

Conventional tailings, using large water-capped tailings ponds, handle off-spec tailings well—low-density or high-fines tailings simply run off the beach and is captured in the pond. Higher quality feed is only needed for dyke construction or beaching. Because the specs for the fluid tailings are low to non-existent, there is little off-spec material. However, the resulting soft tailings deposit is usually difficult and expensive to reclaim.

Where higher-spec tailings are required, more controls are used. Thickeners are an obvious example—chemical dosages and residence times are controlled to produce tailings that meet cer-tain specifications (typically density driven). Where drying beds are employed, the drying areas are sized to be large enough to allow certain volumes of low-density slurries (which take longer to dry and form shallower slope angles), and the operation is designed to allow for upsets such as extended periods of rainy weather.

Even short periods of production of off-spec materials can have significant impact on many operations, unless there are specific provisions for dealing with off-spec materials. Some mines have dedicated areas for placement of off-spec tailings, others allow for greater working of the material (such as additional landfarming of low-density tailings during evaporation drying operations), or have long-term plans for reprocessing the tailings or geologic containment of off-spec tailings in mined out pits capped with water. Too often, there is not enough attention paid to


171

off-spec tailings production, resulting is costly remedial measures being required for stabilization at closure. Close monitoring of the process, the deposits, and having sound plans to deal with the inevitable volume of soft material produced are all needed.

As mentioned above, hydraulic sand capping of the soft deposits can greatly reduce capping costs, enhance consolidation, and lead to trafficable plateaus. Cycloning the tailings stream to produce a coarse sand slurry and using this to cap the soft tailings deposits is used in the oil sands extensively with considerable success, though it is not without its challenges.

6.3 Use of full site investigation for soft tailings deposits

As closure approaches, full geotechnical and hydrogeological characterization of the tailings is required, involving sampling, piezometers, and often vane and cone penetrometer testing (for stratigraphy and soil density) as part of the landscape design process. These programs often required specialized equipment to access the beaches and ponds (and especially the areas of beach that are too soft for tracked equipment but not fluid enough to float a barge). These programs are often quite involved, take place over several seasons, and require specialized expertise. A paral-lel laboratory program to develop consolidation parameters for the tailings and models for shear strength gain are usually required.

The results of hydrogeology of the soft tailings deposits are often surprising. Where the geo-logic materials that underlie the tailings are more permeable than the soft tailings, they can act as underdrains, causing a certain amount of downward drainage. Seepage into pervious dykes can also enhance tailings consolidation and strength adjacent to these dykes. The hydrogeology program needs to investigate all possibilities, provide the boundary conditions for consolidation analysis of the tailings, and allow prediction of upward, downward, and lateral fluxes of process-affected water. The program must also support a good understanding of the water balance for the soil or water cap on the tailings deposits.

6.4 Employing a multidisciplinary team

Most tailings deposits are reclaimed to a terrestrial state with vegetation that supports agreed upon lands uses for the reclaimed landscape. In many cases, wildlife habitat is selected as an important land use. Where water covers are used, wetland function or lake ecosystem establishment may be important.

In addition to the geotechnical and hydrogeology specialists, the team needs to include other specialists—often a surface water hydrologist, a soil scientist, a revegetation expert, and a wildlife expert. A geochemist, aquatic biologist, agrologist, or wetlands specialist may also be required depending upon the situation. Management support and adequate budget and schedule are critical. Often research is required to provide design parameters for the team’s efforts.

7 A MODERN APPROACH

Work on stabilizing and reclaiming several tailings deposits over the past decade have yielded a generalized approach. Much of this work is based on dealing with soft tailings in the oil sands, but also at several metal and mines. The methodology has benefitted from discussions with practition-ers from Europe, Australia, and North America and visits to dozens of tailings operations to learn firsthand from practitioners.

7.1 Design for closure

Up-front planning can greatly reduce the risks and costs of reclaiming soft tailings. As mentioned above, site selection, dyke design, and tailings technology selection are all important activities.


172

7.2 Monitor deposition performance

Dykes are typically highly monitored for geotechnical and seepage performance. This effort and diligence needs to be extended to monitoring the slurry properties (relative to various targets) and the performance of the soft tailings deposits. Much of the monitoring is needed to ensure that the soft tailings management is effective and on-track to produce the desired end condition, and that there are effective measures for dealing with the inevitable off-spec tailings. This monitoring is especially important during times of change—movement to new orebodies, new processes, and especially with ramping up of production which often results in an increase in the rate of produc-tion of off-spec tailings and a reduced capacity for management. The same diligence that goes into monitoring the paystream (the ore production and processing) needs to be applied to waste streams such as tailings.

7.3 Plan for closure

The closure plan for the soft tailings deposits (and the related dykes) is typically updated every five years based on monitoring results, stakeholder and regulatory concerns, and direction from mine management. Advances in tailings technology should be assessed and tailings plans updated if needed. One of the main outputs of a closure plan is a site-wide surface water and groundwater plan—the tailings operation must be aligned with this plan to create a topography that supports the final drainage of the reclaimed tailings slopes and plateaus and a plan for collecting and managing seepage waters. One of the main benefits of a good closure plan is that it allows selection of the best location of the outlet for surface water drainage for the reclaimed tailings plateau—typically one wants to ensure that the ground conditions at this outlet are adequate to support a long-term outlet, and that the tailings placement plan supports drainage to this location—this can mean beaching tailings from the periphery towards this future outlet, and locating the reclaimed water system (and any fine tailings removal system) near the future outlet location.

The conceptual landform design plan for tailings beaches and pond that is developed during the closure planning phase often yields a similar result. The upper beaches are landform graded (see Schor & Gray, 1995) to create some topographic diversity and promote runoff. One or more vege-tated waterways (which are often fen-like) gather this water and transmit it to a shallow marsh-live wetland adjacent to the outlet. The outlet must structure is designed for the probable maximum flood (PMF) and transmits water to the base of the dyke; sometimes construction of a waste rock dump at this location helps to provide greater stability and lower gradients for this outlet. If no water can be ponded over soft tailings, provisions for periodically lowering the outlet elevation or adding more material to accommodate settlement of the soft tailings is needed and complicates design and construction.

7.4 Conduct site investigation

When the tailings deposition is nearing completion, a geotechnical and hydrogeological full site investigation of the beaches and soft deposits is required as described above.

7.5 Preparation of settlement map

A critical element of the landscape performance is the amount and rate of future consolidation set-tlement. Results from the site investigation are used to predict post-reclamation settlements. The materials, topography, and in particular the configuration of the outlet facilities for surface water drainage need to be designed or modified to accommodate the settlement. Ideally the settlement should be largely complete within human timeframes.

7.6 Design a robust cover

Soft tailings covers can have numerous duties including transport of consolidation water, protec-tion of root zones from poor quality waters, providing a growth medium for a viable ecosystem,


173

controlling ingress of water or oxygen, transmitting surface water, all the while accommodating large and variable amounts of settlement. Whereas typically the focus is on finding economical and safe methods of capping soft tailings, successful work involves a multidisciplinary team to guide the design and construction to meet these broader objectives. Often test fills are employed to better understand the geomechanics of the cover placement and its performance once in place.

7.7 Monitor stabilization and cover construction

Designs are typically updated based on the field conditions and experience gained. Designs often change considerably over the years of construction.

7.8 Ensure adequate monitoring and maintenance

Both monitoring and maintenance are integral to the design and construction of the tailings sta-bilization programs. Depending upon the length of time for settlement, both may continue over decades or longer.

8 CLOSING COMMENTS

Stabilization and reclamation of soft tailings deposits is an important element of tailings manage-ment. While every deposit and operation is unique, a broader landscape design-based multidisci-plinary approach to soft tailings management is indicated to reduce risks and costs and to better meet landscape performance goals. The same level of effort into advancement and recognition of tailings management and the development of new tailings technologies need to be applied to a more holistic approach to managing soft tailings. Landscape design is offered as one such methodology.

ACKNOWLEDGEMENTS

The information presented in this paper come from learnings from numerous mines and working with many fine engineers, scientists, and operators. In particular, we’ve relied upon experience at Syncrude and Suncor oil sands mines near Fort McMurray, Alberta, the Alcan alumina refining operations near Perth, Australia, the Aughinish Alumina refinery near Limerick, Ireland, the Wis-mut uranium tailings stabilization project near Chemnitz, Germany, and numerous metal mines in Canada and the Western US. Many of the learnings have come from work with exceptional individuals including Alex Jakubick, Andy Robertson, Bill Shaw, Clara Qualizza, Dave Sego, Don Scott, Ed McRoberts, Gord Livingstone, Iain Bruce, Lee Barbour, Mike O’Kane, Nordie Morgenstern, Robert Mahood, Richard Dawson, and Sean Wells.

REFERENCES

CEMA-RWG Landscape Design Subgroup. 2005. Landscape Design Checklist, Cumulative Effects Manage-ment Association, Fort McMurray, Alberta, Canada.

Dawson, R.F. 1994. Mine waste geotechnics. Doctoral Thesis, University of Alberta, Edmonton, 239 p.Jakubick, A.T. & McKenna, G.T. 2001. Stabilization of soft tailings: practice and experience, Eighth Inter-

national Conference on Radioactive Waste Management and Environmental Management (ICEM’01). American Society of Mechanical Engineers, Brugge, Belgium.

McKenna, G.T. 2002. Sustainable mine reclamation and landscape engineering. PhD Thesis, University of Alberta, Edmonton, 661 p.

Schor, H.J. & Gray, D.H. 1995. Landform grading and slope evolution. Journal of Geotechnical Engineering, 121(10): 729–734.



Liners, covers, and barriers



177


Liner system design for tailings impoundments and heap leach pads

J.F. LupoAMEC Earth and Environmental, Englewood, Colorado, USA

ABSTRACT: The use of geosynthetics in the mining industry has expanded significantly over the past 10 to 15 years. One of the areas where there has been considerable development in the utilization of geosynthetics is in the design of liner systems for tailings impoundments and heap leach pads. Modern liner system designs for these facilities often include a combination of geomembranes, geotextiles, GCLs, geonets, and geopipes.

An important aspect to liner system design, which is often overlooked, is the compatibility between geosynthetic materials with the other components of the liner system (e.g. foundation, drainage materials, protection layers, etc), and the loading conditions for the facility. This is par-ticularly critical where liner systems are exposed to high static and earthquake loading conditions, steep slopes, soft foundations, and lack of suitable construction materials.

This paper discusses design approaches for liner systems for heap leach pads and tailings impoundments, which stress the importance of compatibility between materials and site condi-tions. The author has successfully used the concept of compatibility in liner system design for a number of facilities around the world, some with normal loads exceeding 4 MegaPascal (MPa). The discussions presented in the paper highlight lessons learned while applying compatibility concepts to actual projects.

1 INTRODUCTION

The application of geosynthetic materials in the mining industry over the past 10 to 15 years has seen significant changes, with geosynthetics now being used in a wide range of projects. This is particularly true for geosynthetics in the design of liner systems for heap leach pads and tailings impoundments. Modern liner system designs for these facilities now often include a combination of geomembranes, geotextiles, GCLs, geonets, and geopipes, depending on the application.

In addition to the increased use of geosynthetic materials, our understanding of the perform-ance of geosynthetics in the harsh mining environment has evolved. As many of the worlds mines are located in areas with harsh environments (e.g. extreme temperature and precipitation, sig-nificant seismic events, steep terrain, soft foundations, etc), geosynthetic materials have had to perform beyond “typical” conditions. Under these harsh conditions, liner system design concepts must be modified to focus on the compatibility between geosynthetic materials with the other components of the liner system (e.g. foundation, drainage materials, protection layers, etc). The issue of material compatibility becomes critical when the liner system is used in steep terrain and subjected to high loads.

This paper discusses design approaches for liner systems for heap leach pads and tailings impoundments, which stress the importance of compatibility between materials and site condi-tions. These methods have been used for a number of facilities around the world, some with nor-mal loads exceeding 4 MegaPascal (MPa).


178

2 LINER SYSTEM DESIGNS

Liner systems are often considered to consist solely of the material between the ground surface and the ore or tailings (e.g. a geomembrane liner, GCL, etc) and used to provide environmental containment of process solutions. However, modern heap leach pads and tailings impoundments have liner systems that also include elements to assist with solution collection, drainage, and ore leaching in ore heaps. In the design of liner systems, experience has shown that it is important incorporate all of these elements into the design to ensure compatibility under the anticipated operational loading conditions. At first, this may not seem intuitive; however, as illustrated in the next sections, there is close interaction between all elements of the liner system. It is the interac-tion between these elements that will dictate the overall performance of the liner system.

2.1 Heap leach pads

Heap leach pads (HLPs) are used as part of the mining process for the recovery of gold, copper, silver, uranium, nickel, and other metals and non-metals from ore. An HLP is a lined facility onto which ore is placed, either using truck haulage or conveyor placement. Once the ore is placed on the HLP, leach solutions are applied at a controlled rate to percolate through the ore and the leach-ate is collected at central locations. The leaching solution varies depending on the type of ore to be processed, and may consist of a strong acid (e.g. sulphuric acid) or, in the case of gold and silver heap leaching, a dilute cyanide solution. The selection of the leach solution and application rate is determined through leach column tests.

There are several types of HLPs that may be designed depending on the ore leaching character-istics, site topography, ore production rate, and processing requirements. Types of HLPs include:

• Traditional pads with dedicated lined leach pads and external solution collection ponds,• Valley leach facilities with internal solution storage,• On/off or reusable leach pads with several active cells and external solution collection ponds,

and• Hybrid leach pads, which may consist of a combination traditional, valley leach, and reusable

leach pads.

An HLP is commonly lined with natural or geosynthetic materials to promote solution recov-ery and to provide environmental containment of the leach solutions and ore. Liner systems for HLPs often include a single composite, or double composite liner with a leakage collection layer, as shown in Figure 1. Single composite liner systems generally consist of a geomembrane liner placed over a compacted liner bedding soil. This type of configuration is commonly used in areas that experience low hydraulic head (typically less than 1 m). Double composite liner system con-sists of two geomembrane liners separated by a leak collection/drainage layer. The lower second-ary geomembrane is placed over a compacted liner bedding soil. A double composite liner system is normally only used where high hydraulic heads (several meters) may occur, such as in valley leach facilities.

The geomembrane liner materials commonly consist of High Density Polyethylene (HDPE), Linear Low Density Polyethylene (LLDPE), Polyvinylchloride (PVC), and Polypropylene (PP), although the primary materials used in modern HLPs are HDPE and LLDPE. The extensive use of HDPE and LLDPE geosynthetics in heap leach operations has demonstrated that they are suit-able for containment of acidic solutions and metal leaching products, for periods of over 50 years (Renken et al., 2007).

As illustrated in Figure 1, the design elements in HLP liner systems may include several of the following:

• Foundation,• Liner bedding soil or material between the geomembrane and the foundation,• A geomembrane liner. Two may be used for double-composite systems,


179

• A leak detection layer for double-composite liner systems, which often consists of gravel or sand. The leak detection layer may also include a geotextile, depending on the gradation,

• A protection layer directly over the geomembrane liner, either for permanent use or for use dur-ing construction,

• A drainage or solution collection layer (either above the protection layer or directly on the geomembrane liner),

• Solution collection pipes (geopipes) within the drainage layer,• Air injection layer (generally used in copper leach pads) placed directly above the drainage

layer, and• Air injection piping (geopipes) within the air injection layer.

Figure 2 presents a photograph showing the various components of an HLP liner system under construction.

For HLPs the liner design process is a balance between grading (earthworks), selection and test-ing of geosynthetic materials, and selection and testing of liner bedding and overliner materials. The stresses and shear traction induced by the ore loading and foundation slopes require rigorous testing and analyses to support the design effort. In addition, both safety and construction issues, which need to be integrated into the liner system design.

As noted in Lupo and Morrison (2005 and 2007), Lupo (2008), Breitenbach (1995), and Breitenbach & Smith (2006), there is a close inter-dependency between the components of the liner system. Table 1 presents a summary of inter-dependent relationships between liner compo-nents as they pertain to liner system design issues for HLPs. Figures 3 and 4 present design flow charts (after Lupo, 2008) that illustrate the inter-dependency of the liner components as part of the design process.

Figure 1. HLP liner system configurations.


180

As shown in Figures 3 and 4, there are a number of tests that must be conducted in order to assess the compatibility of materials. These tests include:

• Hydraulic (permeability) testing for liner bedding and drainage layer materials. These tests are primarily used to define the hydraulic conductivity of the materials under saturated conditions.

• Interface shear testing between each primary interface for the liner system (e.g. liner bedding to geomembrane and protection layer to geomembrane). Interface shear testing will define both the peak and residual shear strength, which is required to assess the stability of the HLP. Inter-face shear testing is particularly important when GCLs or geotextiles are used in the liner sys-tem, as these interfaces may be critical in terms of stability.

• Internal shear strength testing (triaxial shear and/or direct shear) of the liner bedding and pro-tection layer. Internal shear testing should also be conducted on GCLs if they are used within the liner system. Internal shear strength is as important as interface shear strength as failures may occur within layers not only interfaces.

• Liner puncture testing of the geomembrane liners. Liner puncture tests are used to verify the geomembrane liner selection is compatible with the liner bedding and protection layer materials under the anticipated loading conditions. These tests should be conducted using the cylinder test method as described in Environmental Agency (2006), Brachman et al. (2000), Lupo & Mor-rison (2007), Shercliff (1998) and Thiel & Smith (2004). The cylinder test method is preferred as it more accurately replicates the conditions present beneath an HLP.

• Compressibility testing of the subgrade, liner bedding, protection layer, drainage layer, and air injection layer materials. Compressibility testing is often conducted using a simple one-dimensional compression frame with the material confined in a steel vessel. The compressibility of the materials within the liner system is often overlooked, but can provide critical information for the design of geopipes under high loads. Studies by Adams et al. (1988), Lefebvre et al. (1976), Reeve et al. (1981), Sargand et al. (1993), Selig (1990), Valsangkar & Britto (1978), Watkins (1990), Watkins et al. (1987) and Watkins & Reeve (1979) have shows a strong cor-relation between geopipe performance and the compressibility of the material surrounding the

Figure 2. HLP liner system under construction.


181

geopipe. These studies have shown that when loads are applied to buried geopipes, a stress arch develops in the material surrounding the geopipe, thereby reducing the load transferred to the geopipe. The application of stress arching concepts to geopipe design is presented in Brachman (2001) and Lupo (2001). Once the compressibility of materials are known, analytical methods, such as those presented in Burns & Richard (1964) and Höeg (1968), may be used to assess geopipe performance under load.

2.2 Tailings storage facilities

Tailings storage facilities (TSFs) are engineered structures that are constructed to impound slurried, thickened, paste, or dry-stacked materials (tailings) resulting from mineral processing activities. Geosynthetic liner systems for TSFs are commonly designed and constructed as single composite liner systems. Double composite liner systems are less common because the potential for leakage is lower. A single composite liner system can be effectively used at TSFs even though the hydrau-lic head within the facility can exceed 100 meters. This is because the tailings often will form a

Table 1. Liner system inter-dependent parameters (after Lupo, 2008).

Design issue Inter-dependent parameters

Environmental Containment - Hydraulic properties of the liner bedding soil (hydraulic conductivity)

- Compaction characteristics of liner bedding soil (density and moisture content)

- Physical characteristics of the liner bedding soil (grain size distribution)

- Foundation settlement (elastic and plastic deformation)

- Mechanical properties of the geomembrane liner (tensile strength, yield strength, elongation)

- Physical characteristics of overliner material in direct contact with geomembrane liner (grain size and angularity of particles)

Solution Collection - Hydraulic properties of the overliner materials under load (hydraulic conductivity)

- Mechanical properties of overliner materials to support solution collec-tion pipes(compaction and shear strength)

- Geochemical properties of the overliner material (acid consumption/generation and decrepitation)

- Characteristics of the solution collection pipes (size, spacing, mechani-cal properties, hydraulic properties)

- Grade/slope of lined surface

- Mechanical and hydraulic properties of geotextile materials

Heap Stability - Interface friction between liner and liner bedding soil

- Mechanical properties of the liner bedding soil (internal shear strength, grain size, plastic/creep behavior)

- Interface friction between liner and overliner material in direct contact with geomembrane liner

- Mechanical properties of the overliner material (internal shear strength and grain size)

- Solution level within overliner material

- Applied ore loading

- Grade/slope of lined surface


182

low permeability layer at the base of the TSF (above the liner system). The hydraulic conductivity of consolidated whole (uncycloned) tailings can range from 10-6 to less than 10–10 m/s (Vick 1983); thereby minimizing seepage from the facility should it occur.

Liner system design for TSFs is similar to HLPs with a few different considerations. For many TSFs, a single composite liner system consists of a geomembrane liner placed over a compacted liner bedding soil (see Figure 1). A GCL may be also considered as a liner bedding layer for TSFs, provided that bentonite migration (Stark et al, 2004) and internal shear (Lupo & Morrison 2007) considerations are addressed.

Geomembrane liner materials typically consist of HDPE, LLDPE, PVC, or Reinforced Poly-ethylene (RPE), although the most common materials are HDPE and LLDPE. A continuous or intermittent drainage layer may also be placed over the liner to enhance tailings consolidation or provide internal drainage for the TSF. A decant or reclaim structure may also be included, which could affect liner system design with the addition of anchor blocks or pipes placed on or near the liner.

As with HLPs, liner system designs for TSFs need to consider compatibility of the various components of the system. Important factors that influence the design of the liner system include, foundation settlements, expansion considerations for the facility (e.g. upstream, centreline, down-stream construction), and environmental issues relating to the size and extent of the reclaim water pond.

Other key liner system design issues include:

• Loading conditions

• Load on liner due to tailings (static and earthquake).• Type of tailings to be stored (slurry, thickened, paste, dry stack).• Spigot/central discharge/truck haul or conveyor (protection layers).

Figure 3. HLP liner bedding design flow chart.


183

• Hydraulic conditions

• Maximum solution hydraulic head.• Solution chemistry which could affect geomembrane selection.• Tailings geochemistry which could affect geomembrane selection.

• Operational Considerations

• Reclaim/decant configuration.• Solution recovery (piping and drainage layers).• Leak detection.• Monitoring.

Since TSFs may cover large areas, the liner system design needs to address staged construction which may span over several wet and dry seasons. These types of designs may include the use of temporary construction berms to prevent liner damage during wet seasons. In addition, tempo-rary protective layers may need to be used during construction for placement of pipes and con-crete structures. Recommendations for construction protection layers are presented by Reddy et al (1996). Large lined area may also be subject to potential damage from wind uplift. Wind uplift may be addressed using ballast (or in combination with protection layers). The amount of ballast to address wind uplift can be assessed using the methods presented in Zornberg & Giroud (1997).

3 GEOMEMBRANE LINER SELECTION

An important element of any liner system design is the selection of the geomembrane liner. As presented in Section 2, geomembrane liners used for HLPs and TSFs typically consist of HDPE

Figure 4. HLP overliner material design flow chart.


184

or LLDPE. This is partially due to the strength and elongation characteristics of the materials. In addition, polyethylene geomembranes are seamed together by hot wedge fusion. Hot wedge fusion welding involves heating the geomembrane while rolling the geomembrane between rollers. The rollers apply pressure, fusing the geomembrane. A double fusion weld is most common (Mollard et al. 1996) and has the advantage of using two sets of rollers to form an air gap in the central portion of the seam. The air gap formed by the double rollers can then be used to nondestructively test the seam using an air pressure test.

In the selection between HDPE and LLDPE geomembranes for liner system designs, the fol-lowing characteristics should be considered:

• LLDPE is low crystalline polyethylene (about 5% crystalline) in comparison to HDPE geomem-branes, which are about 50% crystalline. As a result, LLDPE is more flexible than HDPE which can beneficial if differential foundation settlements or soft foundation conditions are anticipated;

• HDPE has a higher tensile strength than LLDPE, which gives HDPE an advantage for high load applications; and

• While HDPE has a higher tensile strength, LLDPE performs better than HDPE in applications where the liner may be exposed to elongation (high stress and deformation). HDPE has a defi-nite yield point at about 13% strain, while LLDPE can sustain larger strains. LLDPE also has a higher resistance to stress cracking, which could occur in some HDPE applications.

As a general guide, an LLDPE geomembrane should be selected in applications where set-tlements that could result in liner strains in excess of 10% are anticipated. Otherwise, HDPE geomembrane should be selected.

Once a geomembrane type has been selected, the next consideration is liner thickness. As refer-enced in Section 2, liner system designs often include liner-load tests to assess the performance of the geomembrane under load. For HLPs and TSFs, liner load tests use a version of the “cylinder” test (see Environmental Agency, 2006; Brachman et al., 2000; Lupo & Morrison, 2007; Reimers. 1997; Shercliff, 1998; and Thiel & Smith, 2004). A schematic of this type of testing frame is pre-sented in Figure 5. Liner load tests consist of constructing a representative section of the liner sys-tem (liner bedding—geomembrane—overliner) within a rigid-wall cylinder or box. A hydraulic

Figure 5. Liner load testing frame.


185

loading ram is used to apply a load to the liner system. The load is held for a period of time, then the frame is disassembled and the geomembrane liner inspected for failure. Figure 6 presents a photograph of a geomembrane liner after testing showing significant dimpling, but no failure (puncture) of the geomembrane. Liner load tests allow a direct method to evaluate geomembrane liner type and thickness under load and with other elements of the liner system.

While liner-load tests provide a direct method to test the performance of the geomembrane liner under load, it is not practical to run these tests for tens of years, the operational timeframe for most HLPs and TSFs. For long-term geomembrane liner performance, empirical methods may be used based on experience from actual projects. A general guideline for geomembrane liner type and thickness is presented in Table 2. This guideline provides a correlation between the geomembrane liner to foundation conditions, liner bedding, overliner, and load. The correlation presented in Table 2 is meant to be a general guide, not an absolute condition, but can be used as the starting point for liner system design.

4 CLOSURE

The purpose of this paper was to present general guidelines on the state of practice for the design of liner systems for heap leach pads and tailings storage facilities. The liner system design meth-ods stress compatibility between the various elements used within the liner system. Compatibility between the elements is critical to the overall performance of the liner system and facility. Incom-patibility between elements could lead to liner system failure, loss of containment, and poor facility performance.

Figure 6. Geomembrane liner after high loading.


186

By focusing on compatibility between the liner system elements, the overall design will per-form as a single, coherent unit. It is important that sufficient testing be conducted to define the pertinent material properties for design. It is particularly critical that interface shear strength, material compressibility (for geopipe design), permeability (for liner bedding and overliner mate-rials), and liner load performance be tested as part of the design process. These tests provide important design parameters to assess compatibility under load.

REFERENCES

Adams, D.N., Muindi, T. & Selig, E.T. 1988. Performance of High Density Polyethylene Pipe Under High Fill. Geotechnical Report No. ADS88–351F Department of Civil Engineering, Univ. Mass., Amherst.

Brachman, R.W.I., Moore, I.D., & Rowe, R.K., 2000. The design of a laboratory facility for evaluating the structural response of small diameter buried pipes. Canadian Geotechnical Journal 37, 281–295.

Brachman, R.W.I. 2001. Tensile stresses and the durability of PE pipes. Tailings and Mine Waste 2001, Balkema.

Breitenbach, A.J. & Smith, M.E. 2006. Overview of geomembrane history in the mining industry. Proceed-ings 8th International Conference on Geosynthetics, ISBN 90 5966 044 7, 345–349.

Breitenbach, A.J. 1995. Lessons learned from geomembrane liner failures under high fill loads, Geosynthetics: Lessons Learned from Failures, Industrial Fabrics Association International (IFAI). Draft Manuscript.

Burns, J.Q. & Richard, R.M. 1964. Attenuation of Stresses for Buried Conduits. Proceedings Symposium on Soil-Structure Interaction, University of Arizona. 378–392.

Environmental Agency, 2006. Methodology for cylinder testing of protectors for geomembranes on landfill sites, Environmental Agency U.K.

Höeg, K. 1968. Stresses Against Underground Structural Cylinders. Journal of Soil Mechanics and Founda-tion Division, ASCE, 94(SM4), 833–858.

Table 2. General geomembrane liner selection guide after Lupo & Morrison (2007).

Foundation conditions*

Liner bedding soil†

Overliner material‡

Effective normal stress (MPa)§

<0.5 0.5< <1.2 >1.2

Firm or high stiffness

Coarse-grained Coarse-grained 2 mm LLDPE or HDPE

2 mm LLDPE or HDPE

2.5 mm LLDPE or HDPE

Fine-grained 1.5 mm LLDPE or HDPE

2 mm LLDPE or HDPE


Fine-grained Coarse-grained 1.5 mm LLDPE or HDPE


2 mm LLDPE or HDPE

Fine-grained 1 mm LLDPE or HDPE


2 mm LLDPE or HDPE

Soft or low stiffness

Coarse-grained Coarse-grained 2 mm LLDPE 2 mm LLDPE 2.5 mm LLDPE

Fine-grained 1.5 mm LLDPE 2 mm LLDPE 2.5 mm LLDPE

Fine-grained Coarse-grained 2 mm LLDPE 2 mm LLDPE 2.5 mm LLDPE

Fine-grained 1.5 mm LLDPE 2 mm LLDPE 2.5 mm LLDPE

*Description of foundation conditions is a relative measure of stiffness. The foundation conditions need to be investigated and tested to determine compatibility with the geomembrane.

†Liner bedding soil refers to the soil in direct contact with the underside of the geomembrane. Testing and design calculations are required to assess compatibility with the geomembrane.

‡Overliner refers to the material placed directly onto the geomembrane. Testing and design calculations are required to assess compatibility with the geomembrane.

§Effective normal stress is the maximum stress onto the geomembrane.


187

Lefebvre, G., Laliberté, M., Lefebvre, L.M., LaFleur, J. & Fisher, C.L., 1976. Measurement of soil arching above a large diameter flexible culvert. Canadian Geotechnical Journal, 13(58), 58–71.

Lupo, J.F. & Morrison, K.F. 2005. Innovative geosynthetic liner design approaches and construction in the mining industry. In: Proceedings of the ASCE Geo-Frontiers, Austin, TX, 24–26 January, 16p.

Lupo, J.F. & Morrison, K.F. 2007. Geosynthetic design and construction approaches in the mining industry, Geotextiles and Geomembranes, 25, 96–108.

Lupo, J.F. 2001. Stability of HDPE pipes under high heap loads. Society of Mining Engineers Annual Meet-ing, Denver, Colorado, Preprint 01–102, 8p.

Lupo, J.F., 2008. Steep terrain heap leach pad liner system design, First Pan American Geosynthetics Confer-ence, GeoAmericas 2008, Cancun, Mexico.

Mollard, S.J., Jefford, C. E. Staff, M. G. & Browning, G. R. J. 1996. Geomembrane landfill liners in the real world. Geological Society, London, Engineering Geology Special Publications 1996, 11;, 165–170.

Stark, T.D., Choi, H, & Akhtarshad, R., 2004. Occurrence and effects of bentonite migration in geosynthetic clay liners, Geosynthetics International, Vol. 11, No. 4.

Reddy, K.R., Bandi, S.R., Rohr, J.J., Finy, M. & Siebken, J., 1996, Field Evaluation of Protective Covers for Landfill Geomembrane Liners Under Construction Loading, Geosynthetics International, Vol. 3, No. 6, pp. 679–700.

Reeve, R.C., Slicker, R.E. & Lang, T.J. 1981. Corrugated Plastic Tubing. Proceedings International Confer-ence Underground Plastic Pipe, ASCE, New Orleans, Louisiana, 227–242.

Renken, K., Mchaina, D. & Yanful, E. 2007. Use of geosynthetics in the mining and mineral process industry, Nags award of merit. Geosynthetics, 25(4), 38–42.

Shercliff, D.A., 1998. Designing with the cylinder test. In: Proceedings of the Polluted and Marginal Land Conference, Brunel University, London.

Sargand, S.M., Hazen, G.A., Fernando, M.E.R. & Hurd, J.O. 1993. Field Performance of a Corrugated HDPE Pipe. Transportation Research Board, 72nd Annual Meeting, Washington, D.C., Paper 93–0514, 15p.

Selig, E.T. 1990. Soil properties for plastic pipe installations. In. Proceedings of the Symposium on Buried Plastic Pipe Technology, ASTM, STP 1093, 281–293.

Thiel, R. & M. Smith, 2004. State of the practice reviewof heap leach pad design issues, Geotextiles and Geomembranes Vol. 22, pp. 555–568.

Valsangkar, A.J. & Britto, A.M. 1978. The validity of ring compression theory in the design of flexible buried pipes. Supplementary Report 440, Earthworks and Underground Pipes Division, Structures Department, Transport and Road Research Laboratory, Crowthorne, Berkshire ISSN 0305–1315.

Vick, S.G. (1983) Planing, Design, and Analysis of Tailings Dams. John Wiley & Sons, New York, New York.

Watkins, R.K. 1990. Plastic Pipes Under High Landfills. Buried Plastic Pipe Technology, ASTM STP 1093, George S. Buczala and Michael J. Cassady, Eds.

Watkins, R.K., Dwiggins, J.M. & Altermatt, W.E. 1987. Structural Design of Buried Corrugated Polyethyl-ene Pipes. In Transportation Research Record 1129, TRB, National Research Council, Washington, D.C., 12–20.

Watkins, R.K. & Reeve, R.C. 1979. Structural Performance of Buried Corrugated Polyethylene Tubing. Thir-tieth Annual Highway Geology Symposium, FHA, Portland Oregon, 12p.

Zornberg, J.G. & Giroud, J.P., 1997, Uplift of Geomembranes by Wind—Extension of Equations, Geosynthet-ics International, Vol. 4, No. 2, pp. 187–207.



189


Evaluation of geomembrane puncture potential and hydraulic performance in mining applications

Chris AthanassopoulosCETCO, Arlington Heights, IL, USA

Alyssa Kohlman & Michael HendersonTetra Tech, Golden, CO, USA

Joseph KaulKaul Corporation, Lakewood, CO, USA

ABSTRACT: Lining systems in mining applications often consist of a geomembrane under-lain by either a soil liner or a Geosynthetic Clay Liner (GCL). Geomembranes are vulnerable to damage from large stones both in the soil subgrade and in the overlying drainage layer. Although guidance has been developed for minimizing geomembrane puncture, this past work has focused on subgrade protrusions in municipal solid waste applications. There has been limited information regarding puncture performance in mining applications, where extreme loads are encountered and angular, large-diameter crushed ore is often used as the drainage medium above the geomem-brane. This paper discusses a laboratory puncture testing program involving various geomem-branes placed in direct contact with different drainage media under high loads, both with and without underlying GCLs. Variables being examined include: geomembrane type and thickness, GCL type, normal load, and drainage stone size. Preliminary test results have shown that geo-membrane/GCL composite liners are subject to less puncture damage (i.e., lower defect frequency and/or smaller puncture sizes) than geomembrane liners alone. This paper also presents a fea-sibility study of two lining alternatives, geomembrane/compacted soil and geomembrane/GCL composites. The feasibility study compares technical effectiveness and cost effectiveness based on cost savings associated with improved metal recovery rates afforded by improved containment. This information is intended for mining companies and engineers in evaluating lining options and allowable stone sizes.

1 INTRODUCTION

Geomembranes have been used in the mining industry since the early 1970s in solution and evaporation ponds, tailings impoundments, and heap leach pads. Traditionally, heap leach pad lining systems have consisted of a single geomembrane liner placed directly over a prepared sub-grade of locally available soil. Heap fills are constructed by placing a layer of highly-permeable drainage stone (overliner) over the geomembrane. Crushed ore is then placed on the leach pad in 15- to 30-foot (3- to 10-m) thick lifts, sometimes reaching final heights of several hundred feet. The crushed ore is irrigated with a chemical solution which dissolves the precious metals from the ore. The nature of the chemical leaching solution depends on the metal being targeted. Low pH sulfuric acid solutions are generally used to leach copper and nickel; high pH cyanide solutions are used to leach gold and silver. The metal-laden pregnant leach solution (PLS) passes down through the ore pile and is captured in a drainage system. Metals are extracted from the leach solution and the solution is then recycled back onto the leach pile.


190

When under load, geomembranes are vulnerable to damage from large stones both in the soil subgrade and in the overlying drainage layer. Although intact geomembranes are virtu-ally impermeable, installed geomembranes will have a small number of holes due to imperfect seams or damage during construction and filling operations. These holes serve as open path-ways for leakage into the soil below. The leakage rate through each hole increases as the hole size and hydraulic head on the hole increase, and as the permeability of the layer under the geomembrane increases.

A low-permeability layer is often used beneath the geomembrane to form a composite liner. The low-permeability material beneath the geomembrane can be either a compacted soil (clay or silt) liner or a geosynthetic clay liner (GCL). Compacted soil liners are typically constructed within a specific range of water contents and dry unit weights to achieve a maximum hydraulic conductiv-ity of either 1 × 10−6 or 1 × 10−7 cm/sec, depending on performance and regulatory requirements. GCLs are factory-manufactured liners consisting of sodium bentonite, with a laboratory-certified hydraulic conductivity of 5 × 10−9 cm/sec. Several factors affect the rate of leakage through com-posite systems, including the number of holes in the overlying geomembrane, the hydraulic con-ductivity of the underlying soil layer, and the contact quality between the geomembrane and the low-permeability layer (Giroud, 1997). Based on liner leakage measurements collected by the USEPA at 287 landfill cells spanning 91 sites (Bonaparte and Daniel, 2002), GCL-based compos-ite liner systems have been shown to allow less leakage than clay-based composite liner systems.

2 LITERATURE REVIEW

Narejo et al (1996), Koerner et al (1996), and Wilson-Fahmy et al (1996) developed guidance for addressing puncture damage due to subgrade protrusions below the geomembrane in municipal solid waste applications. Their design guidance involves selection of a cushioning geotextile to limit elongation of the geomembrane past the yield point, which helps avoid short-term punctur-ing of the geomembrane. European environmental agencies employ a more stringent approach, where local strains in the geomembrane are restricted to less than 0.25 percent, to not only avoid short-term puncturing, but to also avoid stress cracking of the geomembrane over long periods of time. Cylinder tests were first developed in Germany for 100-mil (2.5-mm) thick smooth HDPE geomembranes in contact with 0.6 to 1.25-inch (16 mm to 32 mm) drainage aggregate. The cyl-inder tests are used as site-specific performance tests to assess the effectiveness of geotextile pro-tection layers over geomembranes in several European countries (Seeger and Muller, 1996) and (UK Environmental Agency, 2006). A load is applied to simulate the waste loading for a particular landfill with safety factors applied to account for different temperatures and test durations. They used safety factors ranging from 1.50 for a test at 40°C for 1000 hrs to 2.50 for a test at 20°C for 100 hrs (Seeger and Muller, 1996).

There has been limited information published related to puncture performance in mining appli-cations, such as heap fills, where extreme loads are encountered and crushed rock is often used as the drainage medium above the geomembrane. For the particle sizes and high loads involved in heap leach applications, the design approaches discussed above for solid waste applications would result in unrealistically heavy geotextile layers. For example, assuming a 500-foot (152 m) high heap, and 1-inch (25 mm) diameter angular drainage stone over the geomembrane, the design guidance developed by Koerner et al (1996) would require a 130 oz/yd2 (4.4 kg/m2) cushioning nonwoven geotextile over the geomembrane to maintain a factor of safety of 3.0 against puncture damage. Unfortunately, the heaviest weight nonwoven geotextile that is readily available in the U.S. marketplace is perhaps only 32 oz/yd2 (1.1 kg/m2). The reality is that cushioning geotextiles are rarely, if ever, used in leach pad applications due to cost and stability considerations (Thiel and Smith, 2003). Additionally, heap fills typically operate over shorter periods of time (5 to 10 years) compared to solid waste landfills (more than 30 years), and are commonly built in less environ-mentally sensitive areas, so maintenance of a defect-free geomembrane over the long-term may not be a critical design priority.


191

Considering the recent price increases in precious and commodity metals, there may now be a stronger incentive to limit geomembrane punctures and PLS loss through liner systems in min-ing applications. Higher metals prices are also driving mining companies to design facilities that may be closer to populated or environmentally sensitive areas. As a result, there is a trend toward improving the containment capabilities of lining systems installed in mines.

Narejo et al (2007) and Heerten (1994), have found that GCLs can serve as effective cushions, minimizing the puncture damage in geomembranes. Compacted soil liners are not expected to offer the same protection; in fact, under the high normal loads seen at heap leach pads, any coarse particles in the compacted soil subgrade present increased puncture risks. Also, since the rate of leakage through defects in a composite liner system decreases with decreasing hydraulic conduc-tivity of the underlying soil layer, GCL-based composite liner systems are expected to allow less leakage than soil-based composite liner systems should a puncture occur (Narejo et al, 2002). A comparison of expected leakage rates through both geomembrane/soil and geomembrane/GCL composite liner systems will be presented later in this paper.

3 PREVIOUS GEOMEMBRANE PUNCTURE TESTING

Two of the authors have overseen geomembrane testing programs for several large heap leach pad projects throughout the world. A summary of these studies is shown in Table 1. Many of these tests involved single HDPE or LLDPE geomembrane liners placed over compacted site soils, cov-ered with different drainage media, and then subjected to normal loads between 180 and 625 psi (1241 and 4309 kPa). For one project, a copper heap leach pad in the southwestern United States, single geomembrane and geomembrane/GCL composite liners were tested in contact with 0.5 to 1.5-inch (13 to 38 mm) drainage stone at normal loads as high as 585 psi (4033 kPa). The majority of the tests resulted in “major” to “severe” yielding and puncturing of 60-mil and 80-mil (1.5-mm and 2.0-mm) LLDPE geomembranes, with the exception of the layer tested with a GCL between the 60-mil LLDPE and the bedding layer. That test resulted in only “minor” to “moderate” yield-ing of the LLDPE. The results indicated that with a protective GCL layer between the LLDPE and the bedding layer, a 60-mil LLDPE geomembrane could be specified rather than a bulkier and less cost effective 80-mil geomembrane, and provide improved hydraulic performance. Photographs of the geomembrane samples are shown in Figure 1.

4 PROPOSED GEOMEMBRANE AND GCL PUNCTURE TESTING PROGRAM

Based on the test results discussed above and the authors’ experience, it appears that geomem-brane/GCL composite liners for heap leach pads may have less severe puncture damage from overlying drainage media than geomembrane liners used alone. For the purposes of the research presented in this paper, “less severe puncture damage” is defined as lower defect frequency and/or smaller puncture sizes. In order to investigate the puncture behavior of geomembranes and composite liners further, a high-load static puncture testing program has been initiated. The test-ing program involves various geomembranes, both with and without underlying GCLs, placed in contact with different drainage aggregate and tested at normal loads up to 750 psi (5171 kPa). Variables examined include: geomembrane type and thickness, GCL type, normal load, drainage stone size, and test duration. The intent of these tests is to more rigorously evaluate the puncture potential of geomembranes used in heap leach pad applications with ultimate ore heights of up to 600 feet (183 meters), assuming a factor of safety of 1.50.

Geomembrane samples will be placed over a standardized sand bedding layer (such as Ottawa sand), covered with the coarse-grained drainage aggregate, and loaded with a Material Testing System (MTS) equipped with Linear Variable Distance Transducers (LVDTs) to monitor displace-ment. A conceptual diagram of the load system is shown in Figure 2. The typical test duration will be 48 hours, with some tests up to two weeks possible. After loading, the geomembrane sample


192

Table 1. Summary of previous geomembrane puncture testing.

Upper materialGeomem-brane

Lower material

Normal stress

Yielding Puncture?

None A B C D N Y

Mexico Heap LeachGravel (1") 60-mil

LLDPESite silt 180 psi X X

Gravel (2") 60-mil LLDPE

Site silt 180 psi X X X

Gravel (minus 2") 60-mil LLDPE

Stockpiles 300 psi X X X

Ore (3/4") 60-mil LLDPE


Ore (1–1/2") 60-mil LLDPE


Ore (1–1/2") 60-mil LLDPE

CL/CH 300 psi X X X X

Gravel 60-mil LLDPE

Site silt 180 psi X X

Turkey Heap LeachOre (minus 1") 60-mil

LLDPESandy CH 214 psi X X

Southwest USA Heap Leach

Ore (1.5" to 0.5") 60-mil LLDPE

Clay 417 psi X X

Ore (1.5" to 0.5") 60-mil LLDPE

Clay 625 psi X X

Ore (1.5" to 0.5")

80-mil LLDPE

Clay 625 psi X X X

Mongolia Heap LeachLean Clay 60-mil

LLDPEMinus 2" crushed rock

256 psi X X X X

Southwest USA Heap LeachCrushed Ore

(1.5" to 0.5")60-mil LLDPE

Minus 0.5" Bedding

312 psi X X

Crushed Ore (1.5" to 0.5")

60-mil LLDPE

Minus 0.5" Bedding

585 psi X X X


60-mil LLDPE and GCL

Minus 0.5" Bedding

585 psi X X X


80-mil LLDPE

Minus 0.5" Bedding

585 psi X X

Crushed Ore (minus 1")

60-mil LLDPE

Minus 0.5" Bedding

312 psi X X

Ore (minus 1") 60-mil LLDPE

Minus 0.5" Bedding

585 psi X X

Geomembrane Yielding Descriptions, based on size and number of yield points:

A = Minor; B = Moderate; C = Major; D = Severe.


193

Figure 1. Comparison of 60-mil (1.5 mm) LLDPE Geomembrane tested in contact with 1/2" to 1" (12 to 25 mm) rock at 585 psi (4033 kPa) for 48 hours, both with (left) and without (right) underlying GCL.

Drainage Stone

Sand subgrade

GCL

Geomembrane

Loading Plate

Figure 2. Conceptual diagram of geomembrane/GCL loading system.

will be removed and examined for punctures, both visually and with a vacuum test. In addition to punctures, other signs of distress, including yielding in the geomembrane (defined as indentations in the geomembrane which do not recover after removal of the pressure) will also be recorded. Yield deformation will be reported as “None”, “Minor”, “Moderate”, “Major” or “Severe”, which are specific terms defined by the number of yield points observed as well as the size of each yield point.

Geomembrane sample with underlying GCL experienced less severe (minor to moderate) yield-ing, and no punctures. Geomembrane tested without GCL experienced severe yielding and one puncture (circled).


194

A separate testing program, involving long-term compatibility/hydraulic conductivity testing of GCL samples in contact with an aggressive, low-pH synthetic copper leach solution is also being per-formed. The tests will follow a modified version of ASTM D6766, the Standard Test Method for Eval-uation of Hydraulic Properties of Geosynthetic Clay Liners Permeated with Potentially Incompatible Liquids. The GCL samples will be hydrated with synthetic leachate under low effective stress and then subjected to a hydraulic head of 2 psi (13.8 kPa) to drive the flow of leach solution through the sam-ples. Testing will be performed at confining pressures ranging from 5 to 500 psi (34.5 to 3447 kPa), to simulate the range of typical operational stages of a copper heap leach facility. The method rec-ommends that testing continue until specific termination criteria (steady-state flow and chemical equilibrium) be established between the effluent and influent. Accordingly, flow and water quality measurements will be collected daily to monitor termination criteria during the testing period.

5 HEAP LEACH PAD LINER FEASIBILITY STUDY

A comparison of expected hydraulic performance and metal recovery was performed for potential leach pad liner options at an example copper heap leach project. Two scenarios were analyzed: (1) a 60-mil HDPE geomembrane overlying a GCL; and (2) a 60-mil HDPE geomembrane over-lying a 1-foot thick layer of compacted soil with a permeability of 1 × 10−6 cm/sec. (State min-ing regulatory agencies in the western United States commonly require the low-permeability soil layer beneath the geomembrane to have a maximum hydraulic conductivity of 1 × 10−6 cm/sec). A copper heap leach has been selected as a “worst-case” example due to potential GCL chemical compatibility concerns between the acidic PLS and the bentonite in the underlying GCL. A gold heap leach, which employs a high-pH dilute cyanide solution, has been shown to be compatible with sodium bentonite (CETCO, 2000), and is therefore expected to result in a low long-term GCL hydraulic conductivity (on the order of 5 × 10−9 cm/sec).

5.1 Flow through geomembrane defects

Theoretical leakage calculations were performed using the semi-empirical Giroud equations (1997). These equations are similar to the equations used in the Hydrologic Evaluation of Land-fill Performance (HELP) model, which were also developed by Giroud (Schroeder et al, 1994). Since geomembranes are virtually impermeable, the only significant liquid migration through the composite liner system will occur through geomembrane defects. HELP provides estimates for the number of installation defects (caused by installation quality, equipment, and surface prepara-tion) that can be expected when the geomembrane is placed using good, fair, and poor installation practice and QA/QC. HELP recommends an installation defect diameter of 1 cm2.

At this time, defect frequencies corresponding to “fair” installation quality (4 to 10 per acre, or 10 to 25 per hectare) will be used for both liner options presented in this paper. The authors feel that this is a reasonable assumption, considering the high loads involved, the common use of crushed rock for overliner, and the fact that heap leach construction projects may not follow the same strict construction quality assurance (CQA) as landfill liner projects. In addition, considering that GCLs have been shown to be effective geomembrane cushions, allowing less puncture damage (Figure 1), the geomembrane/GCL liner option will be assumed to have fewer installation defects (4 defects/acre) compared to the geomembrane/soil liner option (10 defects/acre). Each installation defect will be assumed to be circular, with an area of 1 cm2. A list of the assumptions used in this example fea-sibility study is presented in Table 2. The defect assumptions will be re-examined after the ongoing laboratory puncture testing program discussed above has been completed.

5.2 Interface flow

Where defects are present, the liquid will pass through the defect, and then flow laterally in the space between the geomembrane and low-permeability soil layer before infiltrating through the


195

soil (Giroud, 1997). The radius of this “interface flow” is dependant upon the quality of contact between the geomembrane and the low-permeability soil. Composite liner components in good contact (no geomembrane wrinkles, well-prepared, smooth subgrade) will permit less interface flow (and therefore, less overall leakage) than those components in poor contact. The contact qual-ity factor is a coefficient introduced to account for the effects of interface flow. Giroud provides estimates of 0.21 and 1.15 for good and poor contact quality, respectively. Giroud states that good contact should be assumed with GCLs, since they are usually installed flat and, when under pres-sure, bentonite will exude through the surrounding geotextiles, forming a hydraulic seal with the geomembrane.

5.3 GCL hydraulic conductivity

Sulfuric acid solutions are typically used in copper heap leach pads to leach copper from the ore. This results in an acidic PLS containing high levels of sulfates, dissolved metals, and total dis-solved solids (TDS). Jo et al (2001) found that sodium bentonite samples exhibited approximately a 50 percent decrease in swell at pH values less than 3. As part of the same study, GCL perme-ability values on the order of 10−6 to 10−5 cm/sec were measured at pH values less than 2. However, Ruhl and Daniel (1997) found that when exposed to strong acid, a GCL’s buffering capacity was not exhausted until after 15 pore volumes of flow. At the low water flow rates expected in a liner, it may take months or years for the first 15 pore volumes to flow through liner. By this time, the liner will likely be covered and compressed by several hundred feet of ore.

In addition to exhibiting low pH values, copper PLS contains high levels of dissolved sulfates and metals. High ionic strength solutions may be incompatible with sodium bentonite and can decrease a GCL’s hydraulic performance. Many researchers have observed decreasing swell and increasing hydraulic conductivity in GCLs exposed to high strength leachates and liquids containing high divalent cation concentrations (Jo et al, 2001, Kolstad et al, 2004, Shackelford et al, 2000).

The hydraulic conductivity of bentonite is dictated by not only the pore water chemistry, but also the compressive stress acting on the GCL. Daniel (2000) permeated GCLs with concentrated calcium chloride (5000 ppm) solutions at various confining pressures. At low compressive stress, the calcium solution had a dramatic effect on GCL performance. But as the pressure increased to 58 psi (400 kPa), the hydraulic conductivity to distilled water and concentrated calcium solution was virtually identical. These results are consistent with the findings of Thiel and Criley (2005), who found that at effective stresses greater than 58 to 72 psi (400 to 500 kPa), the hydraulic conductivity of a GCL is independent of the leachate chemistry. Since modern heap leach piles

Table 2. Liner leakage calculations.

60-mil LLDPE/ compacted soil

60-mil LLDPE/GCL (1 × 10−7 cm/sec)

60-mil LLDPE/GCL (5x10−9 cm/sec)

Soil hydraulic conductivity

1 × 10−6 cm/sec 1 × 10−7 cm/sec 5 × 10−9 cm/sec

Soil thickness 1 ft (0.3048 m) 0.02 ft (0.006 m) 0.02 ft (0.006 m)Hydraulic head 1 ft (0.3048 m) 1 ft (0.3048 m) 1 ft (0.3048 m)Contact quality factor 1.15 0.21 0.21Number of defects 10 per acre

(25 per hectare)4 per acre (10 per hectare)

4 per acre (10 per hectare)

Size of each defect 1 cm2 1 cm2 1 cm2

Liner leakage 47 gpad (442 lphd)

3 gpad (28 lphd) 0.3 gpad (2.9 lphd)

Note: gpad = gallons per acre per day. lphd = liters per hectare per day.Calculations performed using the methodology in Giroud (1997).


196

are typically several hundred feet high, the GCL will be under a very high confining pressure, and is therefore expected to maintain a relatively low hydraulic conductivity.

Shackelford et al (2000) and Jo et al (2004) have shown that prehydration of a GCL with clean water prior to exposure to high strength liquids can significantly improve the GCL’s hydraulic con-ductivity. Considering that a GCL typically achieves hydration through moisture in the subgrade within weeks or months of placement, it is likely that a GCL used to line our example copper heap leach pad will be at least partially hydrated with subgrade moisture before it is exposed to any aggressive acidic PLS.

Considering the combined effects of low pH, high ionic strength, prehydration, and confining pressure, a GCL in this example application could be conservatively expected to exhibit a hydrau-lic conductivity less than 1 × 10−7 cm/sec, or an increase of almost two orders of magnitude from the value expected with clean water. As discussed above, to confirm this assumption, long-term compatibility/hydraulic conductivity testing of a GCL in contact with a synthetic copper PLS is currently underway, in accordance with a modified version of ASTM D6766. If the testing indicates that a GCL under high confining pressure is not impacted as strongly by the PLS, the disparity in leakage rates will be even greater. Accordingly, the calculations in Tables 2 and 3 were performed for two different GCL hydraulic conductivity values: 1 × 10−7 cm/sec (significant nega-tive impact) and 5 × 10−9 cm/sec (little or no impact).

5.4 Estimated liner leakage rates and recoverable copper

Giroud’s equation requires knowledge of the hydraulic head on the liner system. For purposes of this calculation, it is assumed that the head is 1 foot (0.3 m), the regulatory requirement in many states. It should be noted that head levels can vary depending on annual rainfall, leach solution application/collection rates, and the type of fill (e.g., valley or heap). The calculations in Table 2 show that, even if the GCL’s hydraulic conductivity increases to 10−7 cm/sec due to chemical inter-actions with the PLS, a geomembrane/GCL composite liner would be expected to allow less than one-tenth as much leakage as a geomembrane/one-foot thick compacted soil composite.

Based on a review of the literature (Drummond et al 2003, Thiel and Smith 2003, and Jergensen, 1999), copper PLS concentrations may range from 3000 to 7000 ppm. By multiplying the leakage rates with 3000 ppm of copper, estimates of the mass of copper escaping through each type of liner to the environment can also be made. These calculations, which are shown in Table 3, indicate that significantly more copper can be captured when using a GCL composite liner. Assuming an average copper price of $3.60 per pound ($7.92 per kg), and a recovery efficiency of 90 percent, the improved recovery rate afforded by a GCL represents an additional $1300 per acre per year of revenue ($3200 per hectare per year). For a large heap leach site of 200 acres (80 hectares), this represents several hundred thousands of dollars per year of added revenue.

Table 3. Copper recovery calculations.

60-mil LLDPE/ compacted soil

60-mil LLDPE/GCL (10−7 cm/sec)

60-mil LLDPE/GCL (5 × 10−9 cm/sec)

Copper in PLS 3000 ppm 3000 ppm 3000 ppmCopper lost due to leakage 433 lb/acre/yr 27 lb/acre/yr 3 lb/acre/yr

(486 kg/ha/yr) (30 kg/ha/yr) (3.3 kg/ha/yr)Copper price (June 2008) $3.60/lb $3.60/lb $3.60/lb

($7.92/kg) ($7.92/kg) ($7.92/kg)Copper recovery 90% 90% 90%Cost of recoverable copper lost

$1401/acre/yr ($3462/ha/yr)

$88/acre/yr ($217/ha/yr)

$10/acre/yr ($25/ha/yr)

Gain in Revenue – $1314/acre/yr $1392/acre/yr ($3246/ha/yr) ($3438/ha/yr)


197

Additional factors not discussed above include a comparison of the installed costs of GCLs and compacted soil liners, as this is a highly variable, strongly site-specific consideration. The authors’ experience at past sites, including a recent mine site in Nevada, has shown that the installed cost of a GCL is roughly equivalent to or lower than the installed cost of a compacted soil liner when the soil is transported from an off-site location, or when soil amendments such as bentonite are required. Another factor is the revenue gained through faster heap leach pad construction when using GCLs. GCLs can often be deployed at a faster rate than compacted low-permeability soil liners can be con-structed, and offer a preferable working surface for deploying and welding the overlying geomem-brane. Additionally, GCLs are factory-controlled materials, with consistent bentonite distribution and hydraulic performance. As such, GCLs are less likely than compacted soil liners to yield failing CQA test results. These factors suggest that GCLs allow for a shorter construction schedule and an earlier start to leaching operations. A final factor to consider when evaluating installed costs of GCLs and compacted soil liners is the potential for reduced screening operations when using a GCL-based composite liner system. If the laboratory puncture testing program described above confirms that a GCL will reduce puncture damage to the geomembrane from coarse-grained overliner materi-als, then a larger stone size may be allowable, resulting in fewer screening operations.

6 CONCLUSIONS

Lining systems in mining applications can consist of a geomembrane underlain by either a soil liner or a GCL. When under load, geomembranes are vulnerable to damage from large stones both in the soil subgrade and in the overlying drainage layer. There has been limited information published regarding geomembrane puncture in mining applications, where extreme loads are encountered and angular, large-diameter crushed ore is often used as the drainage medium above the geomembrane. Considering the recent price increases in precious and commodity metals, and the increased environ-mental sensitivity of the mining industry, there may now be even stronger incentive to limit geomem-brane punctures and PLS loss through the liner system in mining applications.

The author’s experience and preliminary results of high-load static puncture tests have shown that geomembrane/GCL composite liners may be subject to less puncture damage than geomem-brane liners alone over compacted soil subgrades. A feasibility study of two lining alternatives for an example copper heap leach pad was performed. Theoretical liner leakage calculations revealed that, for a reasonable set of assumptions at a typical copper heap leach, a geomembrane/GCL composite liner would be expected to allow only a third as much leakage as a geomembrane/one-foot thick compacted soil composite. The resulting improvement in PLS capture is expected to result in a significant increase in copper recovery and increased revenue (potentially hundreds of thousands of dollars per year for large sites).

Ongoing and future work includes a laboratory puncture testing program involving various geomembranes placed in direct contact with different drainage media under high loads, both with and without underlying GCLs, and long-term compatibility/hydraulic conductivity testing of GCL sam-ples in contact with an aggressive, low-pH synthetic copper leach solution, and shear strength testing under high loads. The results of this laboratory testing will be used to refine the calculations presented in this paper, with the end goal of providing mining companies and engineers with information to assist in their evaluations of potential lining options and optimizing allowable drainage stone sizes.

REFERENCES

CETCO. 2000. Bentomat Compatibility Testing with Dilute Sodium Cyanide. Technical Reference 105.Bonaparte, R., Daniel, D.E. & Koerner, R.M. 2002. Assessment and Recommendations for Optimal Perform-

ance of Waste Containment Systems, EPA/600/R-02/099. U. S. EPA, ORD, Cincinnati, OH, http://www.epa.gov/nrmrl/pubs/600r02099/600R02099.pdf

Daniel, D. 2000. Hydraulic Durability of Geosynthetic Clay Liners. GRI-14, Conference on Hot Topics in Geosynthetics.


198

Drummond, S., Robison, N.E. & Smith, M.S. 2003. Equatorial Tonopah Mine Copper Heap Leach Closure—A Project of Firsts. Heap Leach Closure Workshop, Sponsored by the Mining Life-Cycle Center (MLC) Mackay School of Mines, University of Nevada-Reno.

Giroud, J.P. 1997. Equations for Calculating the Rate of Liquid Migration Through Composite Liners Due to Geomembrane Defects. Geosynthetics International, 4(3–4): 335–348.

Heerten, G. 1994. Geotextile and/or GCL Protection Systems for Geomembranes. GRI-7, Geosynthetic Liner Systems: Innovations, Concerns and Designs.

Jergensen, G., Ed. 1999. Copper Leaching, Solvent Extraction, and Electrowinning Technology. Society for Mining Metallurgy & Exploration.

Jo, H.Y., Katsumi, K., Benson, C.H. & Edil, T. 2001. Hydraulic Conductivity and Swelling of Nonprehydrated GCLs Permeated with Single-Species Salt Solutions. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 127(7): 557–567.

Jo, H.Y., Benson, C.H. & Edil, T. 2004. Hydraulic Conductivity and Cation Exchange in Nonprehydrated and Prehydrated Bentonite Permeated with Weak Inorganic Salt Solutions. Clays and Clay Minerals, 52(6): 661–679.

Koerner, R.M., Wilson-Fahmy, R.F. & Narejo, D. 1996. Puncture Protection of Geomembranes Part III: Examples. Geosynthetics International, 3(5): 655–675.

Kolstad, D.C., Benson, C.H. & Edil, T.B. 2004. Hydraulic conductivity and swell of non-prehydrated geosyn-thetic clay liners permeated with multispecies inorganic solutions. Journal of Geotechnical and Geoenvi-ronmental Engineering, ASCE, 130(12): 1236–1249.

Narejo, D., Corcoran, C. & Zunker, R. 2002. An evaluation of geosynthetic clay liners to minimize leakage caused by protrusions in subgrades and compacted clay liners. In Zanzinger, Koerner, and Gartung (eds), Clay Geosynthetic Barriers, 61–69. Rotterdam: Balkema.

Narejo, D., Koerner, R. & Wilson-Fahmy, R. 1996. Puncture Protection of Geomembranes, Part II: Experi-mental. Geosynthetics International, 3(5): 605–627.

Narejo, D., Kavazanjian, E. & Erickson, R. 2007. Maximum Protrusion Size under Geomembrane/GCL Com-posite Liners. Geosynthetics 2007 Conference, Washington, D.C.

Ruhl, J.L. & Daniel, D.E. 1997. Geosynthetic Clay Liners Permeated with Chemical Solutions and Leachates. Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 123(4): 369–381.

Seeger, S. & Muller, W. 1996. Requirements and Testing of Protective Layer Systems for Geomembranes. Geotextiles and Geomembranes, 14: 365–376.

Shackelford, C.D, Benson, C.H., Katsumi, K., Edil, T., & Lin, L. 2000. Evaluating the Hydraulic Conductivity of GCLs Permeated with Non-Standard Liquids. Geotextiles and Geomembranes, 18:133–161.

Schroeder, P.R., Dozier, T.S., Zappi, P.A., McEnroe, B.M., Sjostrom, J.W. & Peton, R.L. 1994. The Hydrologic Evaluation of Landfill Performance (HELP) Model: Engineering Documentation for Version 3, EPA/600/R-94/168b, US. Environmental Protection Agency, Risk Reduction Engineering Laboratory, Cincinnati, OH.

Thiel, R. & Criley, K. 2005. Hydraulic Conductivity of a GCL Under Various High Effective Confining Stresses for Three Different Leachates. Geofrontiers 2005, Waste Containment and Remediation.

Thiel, R., Beck, A. & Smith, M.E. 2005. The Value of Geoelectric Leak Detection Services for the Mining Industry. Geofrontiers 2005, Waste Containment and Remediation.

Thiel, R. & Smith, M.E. 2003. State of the Practice Review of Heap Leach Pad Design Issues. 17th Annual GRI Conference Hot Topics in Geosynthetics.

UK Environmental Agency. 2006. A Methodology for Cylinder Testing of Protectors for Geomembranes on Landfill Sites.

Wilson-Fahmy, R.F., Narejo, D. & Koerner, R.M. 1996. Puncture Protection of Geomembranes Part I: Theory. Geosynthetics International, 3(5): 605–628.


199


Field performance of cover systems

N.R. AmorimCivil Engineering Department, UFV, BrazilGeoestrutural Consultoria and Projetos Ltda., Brazil

R.F. AzevedoFederal University of Vicosa, Vicosa, Brazil

O.R. FerreiraRio Paracatu Mineração, Brazil

A.G.C. RibeiroCivil Engineering Department, UFV, Brazil

I.D. AzevedoFederal University of Vicosa, Vicosa, Brazil

ABSTRACT: Twenty years ago, Rio Paracatu Mineração (RPM) started the production of gold, in the district of Paracatu, Minas Gerais state, Brazil. Gold is present in the leached ore and is also associated with pyrite, arsenopyrite and chalcopyrite and represents great concern related to the design of cover systems for final decommissioning and reclamation of the mine areas, the main tailings dam impoundment and the specific tanks, to avoid formation of acid drainage. An experi-ment was built at a place where the tailings pilot plant existed. Two store-and-release monitored cover systems were implanted in the area. The monitoring system consists of a complete weather station, soil moisture and suction measurements in each soil layer, runoff evaluation and two lysimeters constructed below each cover to measure percolation. This paper describes the experi-ment construction and the performance of the cover system during the first year of monitoring.

1 INTRODUCTION

Annually the mining industry explores hundreds of millions of tons of soil and rock to extract minerals that, after improvement, are used to produce a large amount of fundamental products for human life (Carrier III et alli 1983).

Frequently, most of the explored material is residue or tailings. In some cases, as in copper or gold mining, these tailings may represent more than 99% of the ore material.

One of the most serious environmental impact associated with mining activities is the oxida-tion of sulphide minerals in presence of oxygen and water, generating an acid aqueous solution denominated acid drainage of mines (ADM).

Rio Paracatu Mineração (RPM) company, a Rio Tinto subsidiary in Brazil, has been working in the gold production for at least twenty years, in the district of Paracatu, state of Minas Gerais, Brazil. The production presents the lowest excavated material/ore relationship of the world with an average of 0,45 g of ore for 1,0 t of dug material. RPM operates a non-conventional open cast gold mine that essentially entails the removal of Morro do Ouro (“Hill of Gold”) without the production of waste rock. The mining started in 1987 producing around six million tons/year and has undergone major expansions in 1995, 1997 and 1999 to achieve the current production rate of 20 million tons/year. The entire reserve is considered ore, and is sent to the processing plant. The ore can be mined using dozers without the use of drilling and blasting.


200

There are two main ore types, the oxide and the sulphide one. Normally, there are two mine fronts in operation, producing an ore blend of 50/50 oxide and sulfide that are sent through a crushing and grinding circuit and, to prevent acid formation in the tailings, crushed limestone is added in the grinding. The final product consists of a grind with 80% passing a 74 microns mesh. Gold (and sulfide minerals) are concentrated in three flotation stages. The solid residues (at 30% sol-ids) from the final concentration process are pumped to “Specific Tanks”. The final flotation tails are partly diverted to two tailings thickeners for water recovery. Thickener underflows and the remaining tails stream gravitate to the “Main Tailings Impoundment” at a solids content of less than 30%. The “SpecificTanks” contain 30 to 40% sulfide and the tailings impoundment less than 0.4% sulfide.

The gold is present in the leached ore and is also associated with arsenopyrite (FeAsS), pyrite (FeS

2) and chalcopyrite (CuFeS

2). Therefore, a great concern exists in relation to the final decom-

missioning and reclamation of the mine areas to avoid formation of ADM.Presently, an expansion project is being concluded to increase the production to 80 Mt/year

extending the useful life of the mine for more 20 years. With this expansion, the so-called “Hill of Gold” will become an open pit with approximately 350 m of depth. The open pit will demand the excavation of, at least, 200 m of soil, generating a considerable volume of waste rock. The closing mine plan will be significantly changed, introducing a final decommissioning need, not only for the mine areas, the main tailings dam impoundment and the specific tanks, but also for the open pit and the 200 Mt waste rock pile that will be generated.

In order to minimize oxygen ingress and precipitation infiltration in these areas, soil cover sys-tems are necessary. Regarding RPM needs and the scientific interest of studying cover systems, two experimental store-and-release covers were constructed with a monitoring system consisting of a complete weather station, soil moisture and suction measurements in each soil layer used in both covers, runoff evaluation and two lysimeters constructed below each cover to measure percolation.

This paper describes the cover system construction and the instrumentation results during the first year of monitoring.

2 BIBLIOGRAPHICAL REVIEW

Cover systems are used mainly to reduce the amount of water that infiltrates in the residues, to control the migration of gases and to isolate the residues of the environment (Abichou et alli 2004; O’Kane and Barbour 2003; Koerner and Daniel 1997; etc.).

These systems are, usually, divided in two types: prescriptives and evapotranspiratives. The prescriptive covers use low hydraulic conductivity layers (compacted clayey soils with or without geomembrane or GCL) to minimize infiltration and to maximize runoff and evapotranspiration. The basic components are a soil layer with high organic matter content appropriate for planting and a barrier layer with low hydraulic conductivity compacted soil. Vegetation, besides the aes-thetic function, guarantees protection to the barrier layer against erosion and increases evapotran-spiration. The barrier layer minimizes the percolation of liquids.

The first layer of evapotranspiratives covers also consists of a high organic matter content soil appropriate for planting, underlain by a low compacted soil layer that has the function of storing infiltration during rain period and releases it back to atmosphere through evapotranspiration during the dry season. During the rain period that layer progressively saturates without allowing a signifi-cant amount of liquid to reach the tailings. As precipitation ceases or decreases, evapotranspiration starts to prevail and progressively reduces soil moisture of the storage layer until next rainy period, when the storage and release processes are resumed. Therefore, the soil layer, instead of “preventing” the passage of liquids, works as a “water tank” which fills and empties depending on season. These store-and-release layers consist of silty sands and/or clayey silts and should be sufficiently thick so that the humidity increment does not reach the base, where the material to be protected is placed. The needed thickness depends on the climatic conditions (evaporation), the vegetation type used


201

in the topsoil layer (transpiration), and the hydraulic properties of the store-and-release soil layer (hydraulic conductivity and water retention functions). The tailings to be covered can be below that layer of liquid storage.

Basically, there are two types of evapotranspirative covers: monolithic and capillary barriers. The monolithic are the ones described previously. Evapotranspirative covers with capillary barriers use a system in which the store-and-release layer overlies a coarser soil layer that increases the water stor-age capacity of the former due to the non-saturated hydraulic conductivity contrast between these two adjacent layers.

In Figure 1b, it can be observed that for the same high suction values, the storage-release and barrier layers have soil moisture equal to Ac and Af, respectively. Due to the differences between the water retention curves of the two soils, Af is much larger than Ac. Consequently, the coarse soil will have a significantly smaller hydraulic conductivity than the one of the fine soil (Figure 1c) and will work almost as an impermeable boundary for the store-and-release layer soil, therefore augmenting its storage capacity.

According to Dwyer (2003) and Carlsson (2002), two main problems exist regarding capillary barriers. One is clogging of the coarse material by the fine soil. In this case, the use of a geotextile as a separation element is advisable. The second problem is related with long and high precipita-tion periods. In such circumstances, the capillary barrier can stop working, since when the coarse soil saturates or is close to saturation, its hydraulic conductivity is much larger than that of the fine soil (“capillary barrier break”).

Nyhan et al. (1990) and Khire et al. (1994) affirm that capillary barriers have been more effec-tive than the conventional ones, are more easily built and cost less than prescritive covers.

However, there is some controversy regarding the effectiveness of the capillary barriers and climatic conditions. According to Benson and Khire (1995), field studies have showed that capil-lary barriers with two layers are effective in arid areas and arid semi-arid regions. Besides, more complex projects, with more than two layers, have worked well in humid areas. On the other hand, Morris and Stormont (1997) comment that capillary barriers are not efficient in regions where moderate to high precipitations occur.

3 MATERIALS AND METHODS

The experiment described in this paper was accomplished in the area denominated “Barrag-inha” used to release tailings of the mine pilot plant for 10 years. The tailings deposited formed

Figure 1. Figure1 (a) Capillary barrier, (b) Water retention curves, (c) Hydraulic conductivity curves (Qian et. Alli, 2002).


202

a layer with thickness varying from 1,0 to 2,5 m that did not have enough bearing capacity to support the traffic of machines to construct the experiment. For this reason, it was necessary to build a layer of compacted soil, herein denominated as foundation layer, approximately 1,0 m thick, overlying the tailings. The two cover systems described below were built on the top of the

Figure 2. General arrangement of the experiment.

Figure 3. Lysimeters construction.


203

Figure 4. Location of the TDR and the HDU used in the cover systems.

Figure 5. Overview of the area, (a) when the experiments began and, (b) nowadays.

Figure 6. Runoff variation with the time.


204

Cover system 1 Cover system 2

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Figure 7. Instrumentation performance of cover systems 1 and 2. (a) and (b) topsoil and precipitation. (c) and (d) Silty soil layer. (e) and (f) clayey soil layer and capillary barrier layer. (g) and (h) foundation layer.


205

foundation layer, each one occupying approximately half of the area, presenting the following characteristics:

• Cover system 1 consisted of a 15 cm thick layer of organic soil that overlies a 50 cm silty soil layer placed over a 50 cm layer of a compacted clayey soil (hydraulic barrier).

• Cover system 2 was composed by the same 15 cm thick organic soil and 50 cm silty soil layers but placed over a 50 cm layer of gravel (capillary barrier).

On the top of the foundation layer and below the cover systems, two large lysimeters were constructed. Figure 2 presents a drawing of the general arrangement of the area, showing the lysimeters, the instrumentation location, and the systems to measure runoff and lysimeter percola-tions. Figure 3 presents an overview during the lysimeter construction.

To measure soil moisture and suction in each soil layer, time domain deflectometers (TDR) and heating dissipation units (HDU) were used. Locations of these instruments are shown in Figure 2 and 4

A complete weather station that measures daily maximum and minimum temperatures, pre-cipitations and relative humidity, as well as wind velocity and net radiation is available at the RPM site.

An overview of the area at the beginning of the experiment and after the rainy season is shown in Figure 5.

4 RESULTS AND DISCUSSIONS

Measured runoff values are presented in Figure 6. It can be observed that, in average, runoff cor-responded to around 30% of precipitation.

Soil moisture variation measured by the TDR located on the top soil layer of lysimeters 1 and 2, during the first year, are shown together with precipitation in Figure 7a and b, respectively. It can be seen that, in general, as expected, soil moisture and precipitation distributions follow a similar trend, although there are some discrepancies. Also, TDR results present large variation in short time periods in the topmost layer.

Soil moisture and suction measured in the different layers of lysimeters 1 and 2, during the first year, are shown in Figure 7c, d, e, f, g and h. Variations are in most of the time consistent, i.e., when soil moisture decreases, suction increases and vice-versa. Unexpected very high suction values were measured in most of the time in all soil layers. This behavior has to be checked in the continuation of the work. It can also be observed that the capillary barrier layer augmented the storage capacity of the store-and-release layer, giving rise to smaller values of soil moisture in the foundation layer. However, since the objective of the cover system is to prevent the ingress of both water and oxygen, cover system 1 seems to be more appropriate as the clayey layer stays almost saturated during the entire year.

5 CONCLUSIONS

This paper described the construction and performance of two cover systems during the first year of monitoring. The following conclusions can be drawn.

Soil moisture and precipitation variations in the top soil layer of lysimeters 1 and 2 followed similar trends during the first year, although some discrepancies could be noticed.

Soil moisture and suction measured in different layers of lysimeters 1 and 2 were in most of the time consistent, i.e., when soil moisture decreased, suction increased and vice-versa.

Unexpected very high suction values were measured in most of the time in all soil layers. This probably occurred due to HDU calibration problems.

Both cover systems prevented significant increase of the water content in the foundation layer and, moreover, in the tailings.


206

The capillary barrier layer augmented the water storage capacity of the store-and-release silty layer.

In spite of that, since the objective of the cover systems is to prevent the passage of both water and oxygen, cover system 1 seemed to be more appropriate as the clayey layer stayed almost satu-rated during the entire year therefore preventing passage of water and oxygen.

Finally, it is expected that this comprehensive monitoring integrated with numerical modeling will assess total system performance and will allow the development of a predictive understanding of ET covers.

ACKNOWLEDGEMENTS

The authors would like to acknowledge Rio Paracatu Mineração (RPM) company for the finan-cial support that made possible the development of the experiment described in the paper and to FAPEMIG for the scholarship provided to the fourth author.

REFERENCES

Abichou, T. et alli (2004) Design of Cost Effective Lysimeters goes Alternative Landfill Cover Demonstra-tions Projects FAMU—FSU College of Engineering State University System of Florida, Florida Center it goes Solid and Hazardous Waste Management, University of Florida, 88 p.

Benson, C.H. & Khire, M.V. 1995. “Earthen Covers goes Semiarid and Arid Climates, “ Landfill Closures—Enviromental Protection and Land Recovery, ASCE, Geothecnical Special Publication, No. 53, R. Jeffrey Dunn and Udai P. Singh, Eds., New York, NY, p. 201–217.

Carlsson, E. 2002. Sulphide-Rich Tailings Remediated by Soil Cover, Doctoral Thesis, Lulea University of Technology, Sweden.

Carrier, W.D., Bromwell, L.G. & Somogyi, F. 1983. Design Capacity of Slurried Mineral Waste Ponds, Jour-nal of the Geotechnical Engineering Division, ASCE, Vol. 109, no. GT5, p. 699–716.

Dwyer, F.S. 2003. Water Balances Measurements and Computer Simulations of Landfill Covers Doctoral Thesis, The University of New Mexico, USES.

Khire, M., Benson, C & Bosscber, P. 1994. Final Cover Hydrologic Evaluation-PhaseIll,Environmental Geo-tecbnics, University of Wisconsin-Madison.

Koerner, R.M. & Daniel, D.E. 1997. Final Covers goes Solid Waste Landfills and Abandoned Dumps. ASCE Press, Reston, Va.

Morris and Stormont 1997. Capillarity Barries and Subtitle D Covers: Estimating Equivalency, Journal of Environmental Engineering, ASCE, Vol. 123, N. 1, p. 3–10.

Nyhan, J.W., AND Hakonson, T. & Drennon, B.J. 1990. The water balances study of two landfill cover designs goes semiarid regions. J. Environmental Quality, 19:281–288.

O’Kane, M. & Barbour, S.L. 2003. Predicting Field Performance of Lysimeters Used to evaluate Cover Sys-tems goes Waste. 6th ICARD, Cairns, QLD, p. 327–339.

Qian, Xuede, Koerner, Robert M. & Gray, Donald H. 2002. “Geotechnical Aspects of Landfill Design and Construction”, in Michigan Department of Environmental Quality, Waste Management Division, USES, p. 399–437.


207


Capping the tailings impoundment at the Jamestown Mine

Julian C. IshamShaw Environmental, Inc, Concord, CA, USA

ABSTRACT: The Jamestown Mine is a gold mine that was operated by Sonora Mining Com-pany from 1986 to 1994. The mine facility consists of three mine pits (including the Harvard Pit), the Tailings Management Facility (TMF), a Waste Rock Storage Area (RSA), and several deten-tion ponds. The TMF consists of a 120 acre lined tailings impoundment, a containment dam, and a leachate collection system.

The California Regional Water Quality Control Board has determined that waste from the former gold mine had impacted the groundwater. Under a guaranteed fixed price contract, Shaw Environmental Inc. was retained to cap the TMF, investigate releases at the mine facility, perform closure of the Detention Pond 5, and evaluate of the RSA and the Harvard Pit. The work described in this paper included:

• Dewatering of the TMF and transferring the water to the Harvard Pit,• Designing of the final cover for the TMF,• Permitting with the Division of Safety of Dams to modify and de-list the Jamestown Mine

Tailings Dam,• Construction and quality assurance monitoring of the final cover of the TMF in compliance

with California CCR Title 27 requirements.

1 INTRODUCTION

The Jamestown Mine is an inactive gold mine that operated most recently from 1986 to 1994 in Tuolumne County, California. On June 22, 2007, the Regional Water Quality Control Board, Central Valley Region (RWQCB) adopted Waste Discharge Requirements (WDRs) Order No. R5–2007–0083, which specifies conditions related to the Jamestown Mine closure. These WDRs were consistent with the Stipulated Judgment By and Between the RWQCB (“Plaintiff ”); current and past owners of the Jamestown Mine site (Defendants); Shaw Environmental Inc. (Shaw); and the Trustee of the Jamestown Trust (Trust) in the case entitled, People of the State of California, et al. v. Sonora Mining Corp., et al.

The purpose of the Trust is to create a mechanism to channel the funds Plaintiff had collected from the Defendants to pay Shaw for site work.

The Trust contracted with Shaw for the design and construction of the capping of the tailings management facility (TMF) (Figures 1 and 2). The capping of the TMF complied with the Tail-ings Management Facility Final Closure Plan prepared by Shaw. The Final Closure Plan included construction quality assurance (CQA) requirements, specifications, and construction drawings, which was approved by the RWQCB and conformed to Title 27 of the California Code of Regula-tions (CCR).

Shaw is also the CQA firm responsible for monitoring construction activities and verifying that the construction materials and the installation procedures were consistent with the technical specifications and drawings of the Final Closure Plan.

The contract between the Trust and Shaw that was competitively won was for a lump sum amount that also include ten years of maintenance of the TMF final cover.


208


209


210

2 PROJECT OVERVIEW

The TMF consisted of an approximate 130-acre tailings impoundment, containment dam, under drain and leachate collection system. The approach described in the Final Closure Plan was to grade the tailings to a 1 percent grade draining to a rock-lined spillway, lowering the top of the dam, grad-ing the benches and upper face of the dam to drain, placing a one-foot-thick low permeability soil layer and one-foot-thick erosion resistive layer over the tailings and graded areas of the dam.

The soil for the low permeability layer was excavated from existing overburden stockpiles, and was screened and crushed before placing. The soil for the erosion resistive layer was excavated from existing overburden stockpiles, a portion of the dam and on-site borrow pits.

An important cost consideration in the successful completion of the project was the availabil-ity of soil in sufficient quantity and quality for the final cover. Shaw was considered the most qualified candidate, but we also under bid the completion. The basis for the lower bid was our confidence in being able to convert on-site shale into a low-permeability soil capability of achiev-ing 1 × 10–6 centimeters per second (cm/sec). The project was successful from both technical and financial aspects as we were able to achieve our goals.

2.1 Construction

The TMF closure construction included the following:

• Dewatering the supernatant pond on the TMF.• Clearing and stripping the existing vegetation of the tailings.• Stabilizing the soft portions tailings.• Grading and compacting the existing tailings to act as a foundation layer for the low permeability

layer.• Grading the upper face and benches of the dam.• Excavating, crushing and screening material for the low-permeability soil layer.• Placing a low-permeability soil layer to limit infiltration of rain water into the tails.• Placing an erosion resistive soil layer to protect the low-permeability layer.• Constructing drainage structures.• Hydroseeding.

The foundation layer placement consisted of the placement of tailings to achieve the project drainage grades prior to placement of the low-permeability layer.

The low-permeability layer consisted of the placement of a 1-foot minimum thickness of processed low-permeability soil from existing overburden stockpiles with permeability less than 1 × 10–6 cm/sec.

The erosion resistant layer consisted of the placement of a minimum of 1-foot thickness of material over the low-permeability layer to provide protection.

2.2 CQA activities

Shaw used CQA procedures outlined in the Construction Quality Assurance Manual, prepared by Shaw to verify construction was consistent with the project specifications and drawings.

3 SOIL COMPONENTS

3.1 Clearing, Grubbing, and Stripping

Prior to the construction of the foundation layer, the existing vegetation was removed from the TMF closure area. The material was stripped using various equipment including low ground pressure D-5 dozers, agriculture scrapers, excavators, loaders, and hauled in trucks to a designated stock-pile area or placed on the rock face of the dam. In addition, approximately 100 million gallons of


211

water was drained from the supernatant pond on the top of the TMF. This was transferred to the Harvard Pit by gravity flow through an 8-inch diameter siphon line. An additional 100 million gal-lons of water was also transferred to the Harvard Pit from the under drain and leachate collection system of the TMF.

3.2 Foundation layer construction

Potholes were excavated through the tailings to assess the stability of the TMF prior to construc-tion. It was discovered that the south end of the TMF would require stabilization. In general, the following procedure was used to stabilize the TMF:

• Rock was excavated from the face and benches of the dam using a Hyundai 450 Excavator.• The rock was loaded and transported by either a ten-wheel dump truck or a Terex articulated off

road dump truck to the placement area.• The rock was dumped and placed with a D-8 dozer.

This procedure was repeated until the area was stabilized. In the area of the central drainage ditch, the tailings were excavated so that the final elevation of the rock would match subgrade elevations. In other areas, the grading plan was modified so that no excavation of tailings would be required.

Concurrent with the stabilization effort the foundation layer was constructed by cutting and filling the tailings until the 1 percent grades were established.

The following procedure was utilized to place the earthfill and foundation layer:

• Laser-controlled John Deere 1814 agricultural scrapers towed by agricultural tractors provided cut and fill within the TMF (mine tailings) in thin lifts until the top of the Foundation Layer grades were achieved.

• The tailings were moisture-conditioned by adding water as needed using a water truck.• Compaction within the TMF was achieved by wheel rolling the foundation layer with loaded

scrapers and other rubber-tired equipment.• The D-5 dozer with a tow behind sheep foot compactor, D-6, and D-8 dozers were used to place

and compact the Foundation Layer along the northwestern perimeter of the TMF.• Motor graders and dozers provided support during grading, compaction, and finish grading of

the top of the TMF.

Grade control was established and maintained using a global positioning system (GPS) and software from Agteck. Lasers were also used to control the agriculture scrapers and the ditch grades.

A CAT D-8 was used to grade the rock face and benches of the dam. The D-8 pushed and compacted the rocks to form a smooth surface. Because of the natural slope of the benches and the amount of very large rocks, the grading plan was modified to minimize the amount of grading needed.

3.3 Foundation layer CQA

Shaw’s CQA monitor observed the construction procedures to evaluate the suitability of meth-ods and compliance with specification requirements. Consistent with the CQA manual, labora-tory tests were conducted on the foundation layer at the required frequency and compared to the specification requirements. The samples were tested for particle size analysis, moisture content, moisture density relationship, and Atterberg Limits.

Shaw’s monitor examined the subgrade prior to placement. The suitability of subgrade, soil moisture conditioning, compactive effort, and grading were some of the items observed. Approxi-mately 178,000 cubic yards (CY) of foundation soil was placed in the TMF and dam.

The CQA monitor conducted field moisture (American Society of Testing and Materials [ASTM D 2216]) and density (ASTM D 2922) tests and compared the results to laboratory


212

compaction curves. If the material did not meet specification requirements for moisture or in-place density, the area was reworked and then retested until passing results were obtained. The results indicated that the material met the requirements set forth in the Final Closure Plan.

3.3.1 TMFBecause of the difficulty of compacting, a portion of the foundation layer material on the TMF, a layer of crushed rock was spread over the softer tailings and used for the foundation layer. This crushed rock bridged over and stabilized the uncompactable tails. The thickness of the crushed rock placed on the foundation layer varied in depth depending on the degree of softness of the encountered tails. Standard nuclear density gage testing of the crushed rock was not required, because the large diameter of the rock is beyond the limit of this type of test. Engineering judg-ment in the field by a Registered Civil Engineer or a Certified Engineering Geologist was used to ascertain when a suitable thickness of crushed rock was placed over the soft tails for stabilization. In addition, densities of less than 90% relative compaction were allowed on the foundation layer in the southern portion of the TMF where compaction was unachievable due to saturated subgrade conditions. Confirmation of these procedures was determined by a passing nuclear density test (90%) in the low-permeability layer over the crushed rock or the soft tails.

The greatest thickness of the crushed rock was placed beneath the former location of the super-natant pond, which was at the lowest point of the TMF adjacent to the spillway of the dam. The presence of the 15-acre plus supernatant pond caused the tailings in this area to be soft, and dif-ficult to compact and regrade. Because of this condition, a significant quantity of tailings were removed from this area and replaced with crushed rock to provide a suitable foundation layer for the low-permeability layer.

Due to the softness of the tailings in this area, an even greater amount of crushed rock had to be placed in former supernatant pond area. Because of this overfilling of crushed rock to provide a suitable foundation layer, this portion of the TMF has a final grade of less than 1%. However, the entire gradient of the TMF achieves an approximate 1% grade due to higher than 1% grades upstream of the former supernatant pond location.

The spillway through the dam was enlarged to compensate for this lower grade adjacent to the dam. The Title 27 closure regulations require the drainage features to be designed to handle the 100-year storm event. The spillway was redesigned to handle the 1,000-year storm event. It is our professional opinion that the enlargement of the spillway to discharge water at a higher rate will compensate for the lower gradient leading into the spillway, thus limit the ponding of water on the TMF during storm events.

3.3.2 Top of damThe benches and the face of the dam were constructed with large diameter crushed rock. Standard nuclear density gage testing of the crushed rock was not required, because the large diameter of the rock is beyond the limit of this type of test. Confirmation of this procedure was determined by a passing nuclear density test (90%) in the low-permeability layer on the dam.

3.4 Low-permeability layer construction

Shaw constructed a test fill that was constructed using material screened from the on-site stock-piles prior to the construction of the final cover. A series of two-stage borehole infiltrometer (Boutwell Permeameters) were used to measure the field permeability of a test pad constructed of the stockpiled clay. The results of the field permeability test indicated that the screened stockpiled clay met the CCR Title 27, Code of Federal Regulations (CFR) Title 40, and the CQA specifica-tions for the low-permeability soil barrier layer. Based this test fill result the quality of the soil was suitable for placement.

The material in was excavated using CAT D-5 and D-8 dozers. The material was pushed to a cen-tral point where it was loaded into two screening plants using a CAT 621 and Kobello SK 210 LC excavators. The screened material was placed into stockpiles using conveyors and stackers.


213

Oversized material rejected by the screening plants was picked up by a CAT 950 loader and transported to a crushing plant that was part of one of the screening plants. The majority of the over-sized material was loaded into the crusher where it was processed and sent through the screening plant with the material from the stockpiles. Some oversized material was later used to construct the erosion resistant layer. Approximately 232,000 CY of soil was screened and stockpiled.

A CAT 980 loader was used to load the screened soil into trucks (various types of trucks were used including semi end dumps, bottom dumps, and side dump trucks) for transportation to the area of placement. Approximately 203,000 CY of low-permeability soil was needed to cover the TMF and benches of the dam with a thickness of 1-foot.

The following procedure was used to place the low-permeability soil layer:

• Wood lath was installed on the approximate 100-foot grid maintained by the grade setter to control low-permeability layer placement on the TMF and the top of the dam.

• The soil was deposited in eight-inch-thick loose lifts with trucks and spread using a CAT D-6 dozer, CAT 163 or JD 570 motor grader and/or CAT 815 compactor.

• The soil was moisture-conditioned by adding water as needed using a water truck.• A self-propelled sheep’s foot (CAT 815) compactor made several passes to compact the soil to

the minimum required density.• The lift process continued until a minimum 1-foot-thick layer of low-permeability soil covered

the TMF and the top of the dam.• The top of foundation layer was surveyed using the GPS system and the top of low-permeability

survey were compared to document the as-built low-permeability soil thickness. The thickness was also verified in the field by the CQA monitor by both checking the markings on the wood lath and checking test pits excavated through the cover.

• The water truck continued to place a fine mist of water on the barrier layer to prevent desicca-tion of the low-permeability soil until the erosion resistant layer was constructed.

Sludge from the bottom of Detention Pond-5 (DP-5) was also placed under the low-permeability layer. The closure of DP-5 was also a component of the project.

3.5 Low-permeability soil CQA

Shaw’s CQA monitor observed the construction procedures to evaluate the suitability of methods and compliance with specification requirements. Consistent with the CQA Manual, laboratory tests were conducted on the low-permeability barrier layer at the required frequency and compared to the specification requirements. The samples were tested for particle size analysis, moisture content, Atterberg limits and moisture density relationship.

Shaw’s CQA monitor observed the subgrade for any loose or soft areas prior to the contractor placing the low-permeability layer soil on the TMF and the top of the dam. Any loose or soft areas were corrected by scarifying and recompacting, or filled with rock as described above prior to placing the low-permeability soil. During construction of the low-permeability soil barrier layer, the CQA monitor observed the construction procedures to evaluate the suitability of methods and compliance with specification requirements. The suitability of subgrade, fill placement, lift thick-ness, soil moisture conditioning, lift bonding, compactive effort, and finish grading were some of the items observed for compliance with the contract requirements.

The CQA monitor conducted field moisture (ASTM D 2216) and density (ASTM D 2922) tests and compared the results to laboratory compaction curves. If the material did not meet specifica-tion requirements for moisture or in-place density, the area was reworked, and then retested until passing results were obtained.

Once the field tests indicated compliance with the density and moisture requirements, labora-tory samples were taken for tests that included particle size analysis, Atterberg limits, and mois-ture density relationships. The results indicated that the material met the requirements set forth in the Final Closure Plan. The survey elevations of the top of foundation layer and the top of low-permeability layer were compared to document the as-built low-permeability soil minimum


214

thickness of one foot. The one foot minimum thickness was also verified in the field by both checking the markings on the wood lath and checking test pits excavated through the cover.

3.6 Erosion resistant layer construction

The erosion resistant layer was constructed with material from various sources including stock-piles, the on top of the dam, and from oversize material from the screening operation. Approxi-mately 215,000 CY of erosion resistant layer soil was needed to cover the 130-acre area with a thickness of 1-foot.

The following procedures were used to construct the erosion resistant layer on the TMF and the top of the dam:

• Soil was excavated from the various stockpiles using a variety of equipment including CAT D-6 and D-8 dozers and various excavators.

• The soil was loaded into dump trucks and hauled to the site.• The soil was placed in an approximated 12 inch lose lifts and spread and compacted by either a

CAT D-6 dozer, or a CAT 815 compactor.• A sheep foot type (CAT 815) compactor made several passes to compact the soil.• This process was repeated until a minimum 1-foot-thick layer of erosion resistant soil was

placed.

3.7 Erosion resistant layer CQA

Shaw’s CQA monitor observed the construction procedures to evaluate the suitability of methods and compliance with specification requirements. Consistent with the CQA Manual, laboratory tests were conducted on the erosion resistant soil layer at the required frequency. The samples were tested for particle size analysis, moisture content, Atterberg limits, and moisture density relationships.

The CQA monitor observed placement operations, monitored lift thickness, and gathered sam-ples for laboratory testing. The placement of the erosion resistant layer was carefully monitored to reduce the possibility of undetected damage to the underlying low-permeability layer on the TMF and the top of the dam. The CQA Manual did not have any material specification requirements for the erosion resistant layer.

The survey elevations of the top of the low-permeability layer and the top of erosion resistant layer were compared to document the as-built erosion resistant soil minimum thickness of one foot. The one-foot minimum thickness was also verified in the field by the CQA monitor by both checking the markings on the wood lath and checking test pits excavated through the cover.

An aerial topographic survey was flown to document the completion of the closure process. The top of erosion resistant layer are shown on the TMF Final Grades drawing.

4 EROSION CONTROL AND DRAINAGE

4.1 Drainage

The TMF was graded to drain to a rock-lined spillway constructed in the southwest portion of the site. To assist the drainage, a central ditch was constructed to the spillway. The spillway was constructed to meet the 1,000-year storm requirements. The design of the spillway was approved by Department of Safety of Dams (DSOD). The DSOD required that the spillway be founded in bedrock to prevent erosion during the 1,000-year storm. The soil and rock that was excavated was used for the erosion resistive layer. Rock from the dam was used to line the spillway, as well as, constructing a rock lining around the lower portion of the TMF. This rock was then grouted in place to stabilize the spillway.


215

The benches on the dam were graded to flow either into the spillway or into a 48-inch corrugated steel pipe drain located on the north end of the lower bench. The pipe was installed by excavating through the rock. The pipe was bedded with screened soil from the stockpiles. After the pipe was installed, it was backfilled with additional screened soil, compacted and the surface covered with rocks to prevent erosion of the soil.

4.2 Erosion control

The TMF and the dam were hydroseeded to prevent erosion damage. In addition to being hydro-seeded, the slopes of the borrow areas and the stockpiles had straw placed on them to help the seed establish itself.

5 CONCLUSIONS

The construction of the TMF cap proceeded in accordance with the design criteria, plans, and specifications pursuant to Title 27 of the California Code of Regulations as verified by the CQA testing. The final cover was approved by the RWQCB.

The project was completed on time and with budget. Issues that commonly plague tailings impoundment closures that were overcome on this project included, the draining of the super-natant pond, stabilization of the tails, finding suitable quality soil on site, and placement of low permeability layer. Photos of the TMF before and after capping, and the final closure grades are shown below.


216

TMF before capping.

TMF after capping.


217

TMF final closure grades.



Water management and geochemistry



221

Applying numerical hydrogeochemical models as decision support tools for mine closure planning

L.E. EaryMWH Americas, Inc., Fort Collins, CO, USA

R. JakubowskiMWH Americas, Inc., Steamboat Springs, CO, USA

J. Eshleman & A. WatsonMWH Americas, Inc., Denver, CO, USA

ABSTRACT: Numerical models of hydrogeochemical processes are commonly used prior to mine startup or expansion to forecast conditions at closure. While forecasting is an important use of numerical models prior to mine startup and for assessments of mine expansion, the models can also be restructured as investigatory tools for guiding decisions about the potential efficacy of implementing various reclamation options prior to closure. In fact, models developed and refined during mine operation may be expected to be more reliable than those used for pre-mine forecast-ing because they are based on updated conceptual models and can incorporate data measured dur-ing mining that are characteristic of important site-specific hydrogeochemical processes.

This paper describes a numerical hydrogeochemical model that has been restructured as a deci-sion support tool. The tool is used to evaluate different options for closing a pit lake located at a hypothetical copper-molybdenum mine in a semiarid environment. This pit lake is expected to become mildly acidic (pH∼4 to 5) with elevated levels of metals if left to fill naturally after cessation of mining due to sulfide oxidation and wallrock runoff. The numerical components of the decision support tool are developed on the GoldSim® dynamic systems modeling platform, and utilize data from current dewatering and environmental monitoring systems from a variety of mine sites to represent hydrogeologic processes. Geochemical processes are represented by direct linkages between GoldSim® and the PHREEQC geochemical model through customized software. The details of these integrated numerical models are presented to the user through a set of dashboards (GoldSim® specific user interfaces) that allow modification of input data and selec-tion of management options for closure of the pit lake, such as no action, rapid filling of the pit with groundwater and/or drainage from heap leach and waste rock piles, and/or addition of lime as a neutralizing agent. The net effects of the options on the long-term filling rate and chemical composition of the pit lake are presented as time trend charts. The management options also allow definition of unit costs for specific aspects of the closure plan, allowing for a comparison of costs associated with different management alternatives. The dashboards provide a convenient and efficient method for evaluating “what if ” types of calculations involving different combinations of management options. Once feasible options are identified, data collection may be directed at refining model reliability while the mine is still in operation.

Management of water quality is an important aspect of most mine plans because it often repre-sents a major liability that has to be considered for successful mine closure. Decision support tools based on reconfiguration of numerical hydrogeochemical models, such as the one described in this paper, can provide a means to assess the potential effectiveness of management options prior to closure, and minimize future liabilities.



222

1 INTRODUCTION

Prior to the startup of a new mine or expansion of an existing mine, a series of detailed hydrologic and geochemical models are typically developed as part of the permitting processes. These models are used to characterize the hydrogeochemical characteristics of the mine site and to base manage-ment decisions regarding the need for mitigation programs to minimize environmental impacts during operation and after mine closure. However, after the initial permit process is completed, these detailed models are often lost or at least rarely revisited to check their veracity, update their concepts and data, or assess their utility to help plan for future steps in the mine life cycle, such as mine closure. This is particularly true for hydrogeochemical models used for water quality predic-tions, such as for waste rock drainage, groundwater contaminant transport, tailings discharge, and pit lake formation. It is probably less true for hydrologic models because they can provide infor-mation for designing and running groundwater dewatering systems and surface runoff controls that are part of the day-to-day mine operation.

The purpose of this paper is to demonstrate that the types of detailed hydrogeochemical models usually developed only for water quality aspects of mine permitting can be reconfigured and used as decision support tools to evaluate potential alternatives for mine closure while the mine is still in operation. In some cases, the original models developed for permitting may be used as a starting point, whereas in other cases when the mine plan has changed significantly, new models may need to be developed. This use of numerical models as decision support tools is not a new concept. For example, Voss and Letient (2006) describe a dynamic system model used for decision support in the operation of water management systems at the Antamina Mine in Peru. Also, the use of predic-tive models for management decision support and closure planning is consistent with the guiding principles for assessing the potential for acid rock drainage and metal leaching at hardrock mines through the linked studies of prediction, prevention, mitigation, and contingency for the purpose of developing environmentally sound, economically viable management practices (MEND, 2005; Price and Errington, 1998; USEPA, 1994). The example used in this paper is a pit lake model that has been reconfigured as a decision support tool.

2 MODEL DEVELOPMENT APPROACH

The typical sequence involved in developing a numerical model of water quality that takes into account complex independent and interrelated processes is as follows:

• Purpose• Conceptualization• Data collection/analysis• Model construction• Model testing, calibration, refinement, and revisiting the conceptualization through construc-

tion steps as needed.

The details for most of these steps are well known and need no further explanation with the exception of the purpose for the model. In developing decision support models, the purpose of that model is different from the normal practice. More specifically, the purpose of decision support for mine closure is to make “what if ” types of calculations based on the set of potential alternatives that may become apparent as potential mine closure strategies during mining operation, which might or might not have been considered in the original mine permitting stage of the mine life cycle. This purpose is different from that for a permit support model, and a comparison of these two purposes may explain this concept better.

Generally, the purpose for building an environmental permit support model is to make predic-tions of water quality for some future point in time. These predictions are compared to regulatory criteria to assess potential impacts, and decisions are made whether or not to modify the mine plan to mitigate environmental risks and preliminarily aid in the design of reclamation plans for mine


223

closure. This is usually a one-time modeling effort based on the best available theoretical, labora-tory, and estimated information at the time of permitting on processes, such as groundwater flow, geochemical reactivity, climate, geologic structure, etc. In many cases, the mine does not exist at the time of applying the permit support models, severely limiting the quality of data on which to base the models.

In comparison, the decision support model described in this paper is developed after mines have been in operation, allowing them to take into account the potential options available for mine clo-sure that may result from changes to the mine plan that make the original permit-support models obsolete. In addition, environmental monitoring data that have been collected over time allow much better assessment of the primary factors affecting water quality compared to the informa-tion available in the permitting stage. The decision support model must be carefully constructed to allow the user to consider the various new options resulting from changes to the mine plan over time and assess closure options based on inclusion of those changes. This requires that the decision support model be constructed in a manner that allows the user to run it numerous times in a what-if fashion rather than a one-time prediction effort. Numerous models are needed to determine the effects of different options for improving future water quality based on up-to-date knowledge of the mine plan and recognition of what are likely to be the greatest environmental liabilities after closure.

3 SET-UP OF A DECISION SUPPORT MODEL FOR PIT LAKE CLOSURE

The example that we are presenting here is a hydrochemical model of a hypothetical pit lake. This model example does not represent any particular pit lake but is based on an amalgamation of a number of different pit lakes that we have worked on in the past and combined here to illustrate the concepts and approach.

3.1 Model development framework

We commonly use the GoldSim® (http://www.goldsim.com/) modeling framework to develop models of water and chemical mass balance. GoldSim® is a well known simulation framework for developing dynamic systems models that includes options for conducting probabilistic analy-ses based on stochastic representations of key input variables. It provides tools for the creation of license-free runtime versions that allow users to operate, control, and inspect model inputs and outputs through dashboards linked to various internal parameters in the model. GoldSim® also allows linkage to supporting software through the use of dynamic link libraries (DLL). It is impor-tant to note that GoldSim® is not the only framework that can be used to develop complex models of water and chemical mass balance, but having all of these types of capabilities in one package makes it a convenient system for constructing models of dynamic systems.

3.2 Model structure

The GoldSim® framework includes the capabilities to simulate chemical mass balances as well as water mass balance in the same model. In practice, a model of combined hydrochemistry will contain one submodel for the water balance and a second submodel for the chemical balance (Figure 1). On the water balance side of the model, flow rates are the primary types of data entered as inputs and simulated. Flow rates are integrated over time in Reservoir elements in GoldSim®, which keep track of cumulative volumes of water in static bodies of water, such as a pit lake.

On the chemical balance side of the model, chemistry is usually input as concentrations (Figure 1). However, GoldSim® works on a mass balance basis rather than concentrations. Hence, the usual approach is to multiply concentrations by flow rates to obtain the rate of chemical mass transfer in units of mass per unit time. Cumulative masses moving through the system for any particular flow route and time are calculated by Integrator elements. At this point, the cumulative


224

masses are converted back to concentrations for the static water body (e.g., pit lake) in Cell Path-way elements by dividing the accumulated masses by the accumulated water volume. This opera-tion is done by the Cell Pathway element (Figure 1).

At this point, the model has a representation of the bulk solution composition, that is, the composition prior to any chemical reactions. GoldSim® has the capability to model simple chemical reactions. However, to represent complex sets of nonlinear reactions that occur in environmental systems, such as aqueous speciation, gas-phase equilibria, redox, and mineral solubility equilibria, a thermodynamic-based geochemical model is required. A GoldSim® model can be directly linked to a geochemical model through a GoldSim® modeling element called an External DLL. This element controls data flow to and from the geochemical model.

Figure 1. Parallel model structure used for combining chemical and water mass balances.

Water Balance

Flow Rates

(Volume Water/time)

Chemical Balance

Cumulative Volume(Flow Rate) X (Time) = Volume

GoldSim® “Reservoir” ElementKeeps track of cumulative volume

Concentrations

(mass/volume)

Mass Transfer Rate(Conc) X (Flow Rates) = mass/time

Cell_Pathway

GoldSim® “Cell Pathway” Element Keeps track of Concentration(Concentration=mass/volume)

Parallel Models

Reservoir

DLL

DLL_for_PHREEQC

GoldSim ® “External Dynamic Link Library Element (DLL)” - Controls data flow to and from PHREEQC

GoldSim® “Integrator” Element Keeps track of cumulative mass

dtIntegrator


225

For the model described in this paper, the bulk concentrations at each time step are sent to the geochemical computer program PHREEQC through the External DLL element and equilibrated solution compositions are returned from PHREEQC back to the GoldSim® model. For our appli-cations, we use PHREEQC (Parkhurst and Appelo, 1999) because of its flexibility in simulating a wide range of chemical processes.

Some explanation of the use of the External DLL element is warranted here. The DLL is a sepa-rate piece of software that is created by the model developer in a programming language, such as C++ or Visual Basic. It is not part of the GoldSim® software package but resides on the file system as a separate piece of customized software. The DLL contains a series of functions that perform different duties based on inputs and provide a set of defined outputs. The External DLL element in GoldSim® uses the functions contained in the DLL by appropriate referencing to the function arguments and data types.

For the model discussed here, the data flow and functions inside the DLL are illustrated in Figure 2. The sequence of operations controlled by the DLL functions is as follows: GoldSim® sends bulk concentrations calculated from the chemical and water mass balance to the External DLL function. The external DLL takes these data and combines them with a template file to create an input file to PHREEQC with the desired set of chemical parameters. The DLL then runs PHREEQC using the newly created input file. It waits until the PHREEQC calculation has finished and then retrieves the PHREEQC outputs of equilibrated solution composition and returns it to the GoldSim® model. This sequence of operations is performed at every time step (Figure 2).

3.3 Model interface and operation

The approach of combining chemical and water mass balances and using PHREEQC for reaction chemistry into a single model provides a flexible and powerful means to simulate complex systems of solution mixing and reaction. In the case of planning for the eventual closure of a mine pit lake discussed in this paper, these capabilities allow the effectiveness of different management strategies for closure planning to be investigated with the same model.

Figure 2. Data flow to and from the External DLL software and DLL functions used to link to the PHREEQC geochemical model.

GoldSim Model

Receive and

format input data

Template

PHREEQC

input file(s)Run PHREEQC

Get output from

PHREEQC and

return it to

GoldSim

External

DLL

Repeat each time step

PHREEQC.exe

DLL

DLL_for_PHREEQC

Functions

GoldSim Model

Receive and

format input data

Template

PHREEQC

input file(s)Run PHREEQC

Get output from

PHREEQC and

return it to

GoldSim

External

DLL

PHREEQC.exe

DLL

DLL_for_PHREEQC

Functions


226

For example, if a pit lake is predicted to be acidic, then potential management strategies for mitigating the effect of acidic conditions might include rapid filling with groundwater, addition of lime, or addition of organic compounds. As a result, the model interface may include inputs to control how these options may be implemented as shown in Figure 3 for the example discussed in this paper. This interface has input controls for adding groundwater at a specified rate for a specified period of time, starting from the time of termination for the dewatering system. The interface also allows the user to investigate the alternative of adding lime to improve water quality by specification of rates of addition over specified periods of time (Figure 3).

The dashboard also contains a checkbox to turn the geochemistry DLL linkage on or off (Figure 3). The reason for this checkbox is that the model runs in less than a second in the hydrol-ogy-only mode but takes about four minutes when the geochemistry is included because of the need to run PHREEQC at each time step. Thus, hydrology runs can be made quickly if necessary. The benchmarks for these run times are monthly time steps for 150 years or 1800 total time steps running on a laptop computer with 2GB RAM and 2.2GHz CPU. Example calculations using the options presented in the dashboard in Figure 3 are discussed below.

Figure 3. Example of a GoldSim® dashboard that serves as a model interface to a pit lake closure alternative investigation model that includes both hydrologic and geochemical effects on water quality.


227

3.4 Input data

The model discussed in this paper is a relatively simple representation of a water and chemical balance for a pit lake. The inflows to the pit lake are groundwater, surface runoff, and direct pre-cipitation. The lake is modeled as a terminal hydrologic system, and the only outflow component is evaporation from the free water surface. The rate of groundwater inflow is represented as a function of water elevation in the pit lake such that the rate of inflow decreases as the lake fills due to decline in the hydrologic gradient between the lake and surrounding groundwater system. The function for groundwater inflow is obtained from a numerical model of the hydrologic system. The rate of direct precipitation is represented as a function of the surface area of the pit lake mul-tiplied by the average, site-specific monthly precipitation. The rate of evaporation is represented as a function of the surface area of the pit lake, the total monthly average pan evaporation, and a pan evaporation coefficient. Monthly climate data are repeated for each calendar year simulated. The rate of surface runoff is modeled as a simple function of the catchment area, which includes the pit wall rock area, monthly precipitation rate, and fractional runoff coefficient.

The chemical compositions of the inflows to the pit lake are given in Table 1. These data are obtained from a combination of onsite monitoring data and estimated leachate concentra-tions developed by scaling of laboratory test data and geochemical modeling. The groundwater composition is based on monitoring data and has a neutral pH and high alkalinity. The surface runoff is based on geochemical modeling of leachate chemistry and is acidic due to oxidation of sulfides in the pit walls. The composition of precipitation is an estimated composition rep-resented as very dilute, slightly acidic water in equilibrium with atmospheric CO

2(g) (Table 1).

The representations of geochemical processes in the model caused by mixing and equilibration of the inflows in the pit lake are obtained from literature information on pit lake geochemistry (e.g., Castendyk et al. 2005; Davis et al. 1996; Eary, 1999; Price et al. 1995; Shevenell, et al. 1999; Tempel et al. 2000).

Table 1. Solution compositions used as inputs to the pit lake model.

Groundwater Pit wall runoff PrecipitationAnalyte mg/L mg/L mg/L

Alkalinity 457 0 1.1pH (s.u.) 6.88 2.00 5.50Aluminum – 151 –Calcium 197 355 2Chloride 57 0.1 0.4Fluoride 1.2 25.8 –Nitrate 2.5 1.5 0.1Iron 0.13 161.1 –Potassium 3.6 9.4 –Magnesium 88 322.1 –Manganese 0.25 96.3 –Sodium 81 456 0.4Sulfate 473 5750 4.5Antimony 0.0006 0.032 –Arsenic 0.0016 0.0009 –Cadmium 0.0001 0.183 –Chromium 0.005 0.213 –Copper 0.020 147 –Lead 0.0015 0.020 –Mercury 0.0001 0.0001 –Selenium 0.00005 0.112 –Zinc 0.15 13.2 –Dissolved O2

8.1 8.1 8.0


228

4 RESULTS OF DECISION SUPPORT MODEL FOR PIT LAKE CLOSURE

The structure of the model allows comparison of hydrochemical results for a base case condition to results from changes to the system due to the effects of the different management options shown in the dashboard in Figure 3. These options include the base case where no mitigation is included, rapid fill with groundwater, and addition of lime at a prescribed rate.

For example, the rate of filling of the pit lake is shown in Figure 4 compared to the rate cal-culated for the management option of pumping additional groundwater into the pit at a rate of 500 gpm (31.5 L/s) for the first five years of filling. The results show that the addition of ground-water at this rate has a negligible effect on the overall rate of pit lake filling.

However, the groundwater has a high alkalinity, so the effect on the chemical composition of the pit lake from pumping in additional groundwater is large. This large effect is shown in Figure 5 where it can be seen that the pH trend for the base case is predicted to decrease over time and eventually become progressively more acidic after approximately 44 years. This trend is due to the change in relative amounts of inflow to the lake over time. Early in the filling history, the vast majority of inflow water is from groundwater, which is alkaline. Over time as the pit fills to a higher elevation, the rate of groundwater inflow decreases, whereas the proportion of acidic inflow from wallrock runoff remains relatively constant even though the total rate of runoff decreases slightly as the pit walls become partially submerged. As a result, the alkalinity originally built up in the lake is increasingly consumed by the continual addition of acidic runoff and the pH and alkalinity decrease proportionately (Figure 5). In comparison, for the scenario where 500 gpm (31.5 L/s) of the alkaline groundwater is pumped to the pit lake for the initial 5 years of filling, the alkalinity becomes high enough to buffer the acidic runoff and the lake water remains near neutral for the duration of filling (Figure 5).

The effects of the high alkalinity of the groundwater addition can also be seen in trends for metals. Metal concentrations continually increase over time as the pH decreases for the base case (Figure 6), where we use the sum of Al, Fe, and Mn as a proxy to represent metal chemical behav-ior in general. In comparison, metal concentrations remain low throughout the simulation period for the case of adding groundwater due to the maintenance of the pH at a near-neutral level.

Figure 4. Rates of pit lake filling for the base case and rapid fill with groundwater option.

1900

2000

2100

2200

2300

2400

2500

2600

2700

2800

2900

0 20 40 60 80 100 120 140 160

Years

Pit

La

ke

Ele

va

tio

n (

ft)

0.0E+00

5.0E+03

1.0E+04

1.5E+04

2.0E+04

2.5E+04

3.0E+04P

it L

ak

e V

olu

me

(m

illi

on

s o

f g

al)

Base Case

Rapid Fill, 500 gpm for First 5 Years

Vol.

Vol.Elev.

Elev.


229

2

3

4

5

6

7

8

0 20 40 60 80 100 120 140 160

Years

pH

0

50

100

150

200

Alk

alin

ity

(m

g/L

as

Ca

CO

3)


Base Case

pH

pH

Alkalinity

Alkalinity

Figure 5. Alkalinity and pH trends for the base case and rapid fill with groundwater option.

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100 120 140 160

Years

Al+

Fe+

Mn

, m

g/L


Base Case

Figure 6. Trends in metals (Al+Fe+Mn) for the base case and rapid fill with groundwater option.

The next scenario that we examine is the addition of lime compared to the base case as shown in Figures 7 and 8. For a scenario where lime is added at a rate of 5000 tonne/yr for 50 years, the pH remains at or above about 6 for approximately 90 years, which is significantly higher than for the base case, in which the pH drops below 6 after approximately 5 years. But, thereafter the pH decreases fairly rapidly due to the continual addition of acidity from the wallrock runoff and consequent consumption of the lime (Figure 7). For a second scenario, where lime is added at a


230

2

3

4

5

6

7

8

0 20 40 60 80 100 120 140 160

Years

pH

Base Case

Lime at 5000 t/y for First 50 Years

Lime at 10,000 t/y for First 50 Years

0

10

20

30

40

50

60

70

80

0 20 40 60 80 100 120 140 160

Years

Al+

Fe

+M

n,

mg

/L

Base Case

Lime at 5000 t/y for First 50 Years

Lime at 10,000 t/y for First 50 Years

Figure 7. Trends in pH for the base case and options of two rates and time lengths of lime addition.

Figure 8. Trends in metals (Al+Fe+Mn) for the base case and options of two rates and time lengths of lime addition.

rate of 10,000 tonne/yr for 50 years, the pH remains at or above about 6 for a longer period of approximately 140 years before starting to decrease (Figure 7). Metal concentrations for these two scenarios remain low for the initial periods when the pH is maintained at 6 or above but then


231

increase in response to lowering of the pH as would generally be expected for most metals in acid rock drainage affected systems (Figure 8).

The data shown in Figures 4 through 8 are examples of what-if calculations that can be made with the type of model described in this paper. For the particular hydrogeochemical conditions of this pit lake, the option of adding groundwater appears to be more effective from an environmental and economic perspective over the long run as compared to long-term lime addition. The ground-water addition option takes advantage of the high alkalinity of the natural groundwater to mitigate the effects of acid drainage from the wallrock runoff. At other pit lakes where groundwater com-positions are less alkaline, a different result may be obtained, but in either situation, the decision support type model provides a means to investigate potential options for managing closure.

5 CONCLUSIONS

The hydrogeochemical model discussed in this paper is a simple representation of water and chemical balances for a pit lake in which many hydrochemical processes are lumped together. This simple structure is consistent with the purpose of the model, which is to provide a tool for rapidly investigating the relative merits, effort required, and costs of different alternatives for improving water quality in the pit lake after mine closure. In summary, the idea here is to leverage the use of already existent data and numerical models to the extent possible to construct a relatively simple model that is focused on determining the efficacies of different closure options.

REFERENCES

Castendyk, D.N., Mauk, J.L. & Webster, J.G. 2005. A mineral quantification method for wall rocks at open pit mines, and application to the Martha Au-Ag mine, Waihi, New Zealand. Appl. Geoch. 20: 135–156.

Davis, A. & Eary, L.E. 1996. Pit lake water quality in the western U.S.: Unifying concepts. In: 1996 SME Annual Meeting, Phoenix. Soc. Min. Metall. Expl, Littleton, CO, Preprint 96–136.

Eary, L.E. 1999. Geochemical and equilibrium trends in mine pit lakes. Appl. Geoch. 14: 963–987.MEND. 2005. List of Potential Information Requirements in Metal Leaching and Acid Rock Drainage

Assessment and Mitigation Work. Mine Environment Neutral Drainage (MEND) Program, MEND Report 5.10E, Natural Resources Canada (available at: http://www.nrcan.gc.ca/mms/canmet-mtb/mmsl-lmsm/ mend/reports/report510-e.pdf).

Parkhurst, D.L. & Appelo. C.A.J. 1999. User’s guide to PHREEQC (version 2.14): A computer program for speciation, Batch-reaction, one-dimensional transport, and inverse geochemical calculations. Water-Res. Invest. Rep. 99–4259 U.S. Geol. Surv., Denver, Colorado.

Price, W.A. & Errington, J.C. 1998. Guidelines For Metal Leaching and Acid Rock Drainage at Minesites in British Columbia. Ministry of Energy and Mines, British Columbia, Canada (available at: http://www.em.gov.bc.ca/Mining/MinePer/ardguide.htm).

Price, J.G., Shevenell, L., Henry, C.D., Rigby, J.G., Christensen, L.G., Lechler, P.J. & Desilets, M.O. 1995. Water quality at inactive and abandoned mines in Nevada. Open-File Rep. 95–4, Nevada Bur. Mines. Geol., Reno, Nevada.

Shevenell, L. & Connors, K.A. Henry, C.D. 1999. Controls on pit lake water quality at sixteen open-pit mines in Nevada. Appl. Geoch. 14: 669–687.

Tempel, R.N., Shevenell, L.A., Lechler, P. & Price, J. 2000. Geochemical modeling approach to predicting arsenic concentrations in a mine pit lake. Appl. Geoch. 15: 475–492.

USEPA. 1994. Acid Mine Drainage Prevention, Technical Document. Report EPA530-R-94–036 (NTIS PB94–201829), U.S. Environmental Protection Agency, Office of Solid Waste, Washington, DC.

Voss, C. & Letient, H. 2006. Peruvian mine operation using dynamic system modeling. Southwest Hydr. July/August: 22–23, 31.



233

Arsenic species & its binding forms in tailing sediments

T. NaamounDepartment for Geology, Sfax, Tunisia

B. MerkelTU Bergakademie Freiberg, Department for Geology, Freiberg/Sa. Germany

ABSTRACT: The considerable amount of hazardous arsenic detected in the seepage water of the uranium tailings of Schneckentein suggest that high attention to that metalloid is necessary A hydrochemical study was conducted to define the different chemical parameters of pore waters in the tailing system. Also, to determine the fate of arsenic contained within the solid matter, it is important to identify its forms present and assess its mobility. Thus, a seven steps sequen-tial extraction was conducted. The measured E

h–pH values favour the solubility of arsenic. This

result is also confirmed by the model PHREEQC. It shows that AsO4

3–, H2 AsO

3–, H

3 AsO

3 and

HAsO4

2– are the most frequent arsenic species with the dominance of the last mentioned one with a share exceeding 60%. Also, the geochemical procedure confirms its high mobility and most of the non residual arsenic is in association with the nodular hydrogenous fraction (between 75 and 98%) with the dominance of the moderately reducible phase.

1 INTRODUCTION

Arsenic is of increasing concern due to its high toxicity and widespread occurrence in the environ-ment (Wang and Mulligan, 2006). Indeed, many cases of acute and chronic arsenic poisoning have been reported in various part of the world (Gray et al., 1989; WHO, 1993; Gorby, 1994; Senesse et al., 1999; Saha, 2003). It is also demonstrated that arsenic can cause toxic effects for plants or may accumulate in plants and thereby enter the animal and human food chain (Sheppard, 1992).

Arsenic exists essentially in four oxidation states (–III, 0, +III, and +V), as both inorganic and organometallic species in the environment (Cullen and Reimer, 1989; Nriagu, J.O., 1994; Frankenberger Jr. W.T., 2002; Cai and Braids, 2002; Le et al., 2004). The most common forms of arsenic present in groundwater are the inorganic forms of arsenate (As(V)) and arsenite (As(III)) (Cullen and Reimer, 1989; Welch A.H., 2000; Welch A.H. et al., 1988; NRC, 1999; Watt and Le, 2002). The forms of arsenic present are dependent on the type and amounts of sorbents, pH, redox potential (E

h), and microbial activity (Yong and Mulligan, 2004). Arsenic metal rarely

occurs and the (–III) oxidation state is found only in extremely reduced environment. Arsenate ions (As(V)) are most prevalent in oxic conditions whereas arsenite ions (As(III)) are found in anaerobic conditions(Challenger, 1945; Braman, 1975; Cullen and Reimer, 1989). The toxicity and mobility of arsenic species differ with their chemical forms and oxidation states (NRC, 1999; Thomas et al., 2001). Generally, inorganic forms are more toxic and mobile than organoarsenic species, while arsenite is considered to be more toxic and mobile than arsenate (Gulens et al., 1979; Squibb and Fowler, 1983; Xu et al., 1988; Lamble and Hill, 1996). For instance, according to Gulens et al. (1979), As(III) is 5 to 8 times more mobile than As(V) in a non adsorbing sandy loam. Squibb and Fowler (1983) reported that As(III) is 10 times more toxic than As(V). Despite the fact that knowledge of the total arsenic concentration is necessary, speciation analysis is also very important as the toxicity, bioavailability, and treatment and removal options depend on the species of arsenic present (Cullen and Reimer, 1989; Nriagu J.O., 1994; Frankenberger Jr. W.T.,



234

2002; Cai and Braids, 2002; Le et al., 2004; NRC, 1999; Watt and Le, 2002). For example, the particulate arsenic and As(V) are generally easier to remove than As(III) from water (NRC, 1999, Chen and Frey, 1999; Wang et al., 2002; Clifford et al., 2003; Chwirka et al., 2004).

Mining residues usually contain high concentrations of arsenic and are of concern as potential sources of environmental contamination. The purpose of this paper is to give a comprehensive analysis of the distribution and mobility of arsenic in the uranium tailings of Schneckenstein; Germany.

2 A SHORT DESCRIPTION OF THE SITE

The area of investigation, the uranium tailing Schneckenstein , is located in the Boda valley north of the village of Tannenbergsthal/county of Vogtland, southwest of Saxony (Figure 1).

The Boda valley is boardered by the Runder Hübel (837 m above sea level) and the Kiel (943 m a.s.l) to the Northwest and Southeast respectively.

The area is situated within the southern branch of the Boda valley at an altitude of 740 to 815 m above sea level.

On the other hand, the uranium tailings Scheckenstein lie in the area of the watershed between the Zwickauer basin and the Eger. It belongs to the precipitation-richest of the entire Erzgebirge.

Depending on the altitude, 960 to 1160 mm precipitation fall annually (I.W. 1958) and the study area receives an average annual precipitation of 1053 mm. The mean annual temperature is 5.5°C. The calculated evaporation is of 413 mm.

From the geological point of view, the area of investigation is located on the Southwest border of the Eibenstock granite. The Eibenstock granite could be considered a biotite-syenogranite and belongs to the group of the younger granites (JG1). It is medium to coarse-grained, serialporphyric and tourmaline bearing. Its intrusion is related to the Upper Carboniferous series. The Eibenstock granite is covered with a weathered surface layer. Southwest of the investigation area follows the contact zone with quartz-schist.

Figure 1. The area of investigation—Uranium tailings Schneckenstein.


235

3 EXPERIMENTAL DETAILS

The investigations consist of two aspects, the field work and laboratory analyses.

3.1 Field activity and sampling

Four sediment cores were taken at the tailing site by drilling to different depths. Two borings were located in each tailing (Figure 2). The first and the second borings (GWM 1/96; GWM 2 /96) terminated at the granite formation in Tailing 2 (IAA I). The third boring (RKS 1/96, Tailing 1(IAA II) was sunk to a depth of about eight meters. The granite formation was not encountered due to technical problems. The fourth boring (RKS 2/98, Tailing1) is 12 m deep. The cores (diameter 50 mm) were cut into slices of 1 m length and transported in argon filled plastic cylinders to avoid contact with air. A summary of the location of borings and materials used are given in Table 1.

3.2 Analyses in the laboratories

3.2.1 Pore water analysesTwo different procedures were used for the chemical water analysis.

3.2.1.1 Pore water extraction was conducted by mean of a high pressure deviceDue to the low permeability of the material on the one hand and the limited water content on the other, this procedure was not effective and only three samples were analysed. Directly after the extraction process the electric conductivity, the redox potential and the pH values were measured. For the trace elements determination including arsenic, the water samples were filtered with 0, 2μm membrane filter then stabilised with diluted nitric acid (1 ml acid/100 ml water) until a pH ∼2 and finally filled in polyethylene bottles then stored in a refrigerator. For the measurement process, the ICP-MS equipment was used. The detection limit of arsenic is 10 μg/l.

3.2.1.2 Mixed pore water extractionThe procedure is similar to the first step of the Salomon and Forstner extraction procedure described in Salomon and Forstner, (1984) with the difference that in this time the analysed mate-rial was not crashed. The arsenic concentration was determined using ICP AES equipment. Its detection limit is 321 μg/l.

Figure 2. Location of the boreholes in the tailings sites IAA I & IAA II.


236

3.2.2 Sequential extraction procedureSediment samples characterised with their difference in mineral contents and metal concentrations were collected from the second borehole at different intervals (0,5–1,5 m; 3,5–4,5 m; 5,5–6,5 m; 6,5–7,5 m; 8,5–9,5 m; 11,5–12,5 m; 14,5–15,5 m; 17,5–18,5 m; 19,5–20,2). In order to avoid con-tamination, the soil sample was taken with a polyethylene spoon from the middle of each sample container. The samples were freeze-dried. They were subsequently ground in agate mortar until a grain-size ≤ 63 μm, homogenised and stored until needed.

The leaching procedure and reagents are shown schematically above in Figure 3. After the same DIN (DIN 38414/7) as in the last fraction presented in the diagram sited above, an aqua-regia unlocks for the entire samples from each depth interval was accomplished. Due to inaccurate uranium determination with X-ray fluorescence analysis, a fluoric acid unlock was accomplished for the entire samples. The procedure is as the following:

Into a well cleaned beaker, from each freeze dried sample, between 0.1 and 0.2 g of well pulver-ised soil material was filled. 2.5 ml of HNO

3 with 5 ml HF were added to it. Without exhausting,

the sample was two times heated for one hour at temperature of 50°C, then for half an hour at 100°C. Thereafter, it was evaporated to the dry with exhausting. After the cooling of the sample, the same volume of both acids was added to it and the same cited procedure was repeated. Then the sample was lifted by adding 2 ml of concentrated HNO

3 and shortly heated. Finally, by adding

an amount of approximately 10 ml of deionised water and heating, the concentrated precipitation was diluted. The resulting solution was filled in special bulb afterwards conserved.

To minimise losses of solid material, the selective extraction was conducted in centrifuge tubes polypropylene. Between each successive extraction, separation was effected by centrifuging at 4000 rpm. The supernatant was removed with a pipette and filled in polyethylene-bottles then analysed for trace metals, whereas the residue was washed with ∼10 ml of deionised water. After centrifugation for 45 min, this second supernatant was discarded. The volume of rinse water used was kept to a minimum to avoid excessive solubilization of solid material, particularly organic matter. All reagents used in this work were of analytical grade. All glassware used for the analysis was previously soaked in weak nitric acid and rinsed with deionised water.

The arsenic concentrations in the aqua-regia extracts were measured by means of ICP-AES, whereas ICP-MS was used for other fractions.

4 RESULTS AND DISCUSSION

4.1 Hydrochemical analysis and hydrochemical model

4.1.1 Chemistry of the porewaterThe arsenic concentration lies between 49 and 105 μg/l with a mean value close to 82 μg/l and a standard deviation of 29 μg/l. The analysed tailing sediments have pH value near 7 which increases slightly with depth. The E

h value ranges from 406 to 430 mV with a mean value around 420 mV

and a standard deviation of 12 mV. According to Wagman et al., (1982), the measured Eh –pH

values favour the mobility of arsenic in its HAsO4

2– ionic form.

Table 1. The location of borings in the tailing sites.

Location Tailing I Tailing I Tailing II Tailing II

Co-ordinate 45 3256 45 3258 45 3265 55 8723 55 8727 55 8703 Depth under the deposit (m) 9,5 20,2 8 12Filter-tube 1,00 m HDPE 1,00 m HDPE 1,00 m HDPE 1,00 m HDPEFull-tube 9,00 m PVC 19,00 m PVC 8,00 m PVC 12.00 m PVC


237

1. + 100ml H2O deionised; shake 24 h then centrifuge

2. + 100ml 1M NH4ac/pH=7; shake 2 h then centrifuge

3. + 100ml 1M Naacetate-Puffer/pH=5; shake 5 h then centrifuge

4. + 500ml 0.1M NHOH*HCl + 0.01M HNO/pH=2; shake 12 h then centrifuge

5. + 500ml 0.2M NH4ox + 0.2M oxal acid/pH=3; shake 24 h then centrifuge

6. + 15ml 0.02M HNO3 + 25ml 35% H

2O

2+ 2h at 80°C

+ 25ml 35% H2O

2+ 3h at 80°C

+ 500ml 6% HNO3+ 1M NH

4ac; shake 24 h then centrifuge

Dried at 105°C

7. aqua regia (unlock) after DIN 38414/7

5.0 g Sample

Residue

moderately reducible

organics/sulfides

silicates

easily reducible

carbonates

exchangeable-cations

water-soluble share

Residue

Residue

Residue

Residue

Residue

1.0 g Sample

Figure 3. Sequential extraction—Scheme.

4.1.2 Chemistry of the mixed porewaterAs said by Wagman et al., (1982), by the recorded pH values (Figure 4), arsenic tends to be highly mobile mostly in its ionic form HAsO

42−.

Its content is tremendously high in almost all the analysed samples. The extracted water sam-ples from the first two boreholes show a total arsenic concentration varying from 13 to 980 μg/l and from 13 to 2810 μg/l with mean values of 366 and 1305 μg/l and standard deviations of 347 and 1168 μg/l respectively. For those extracted from the third and the fourth boreholes demon-strate an arsenic content ranging from 80 to 590 μg/l and from 0.2 to 427 μg/l with mean values of 373 and 234.6 μg/l respectively. Their standard deviations are 177 and 210 μg/l respectively. The arsenic species measurements show the dominance of the valence +5 in almost all analysed


238

samples. Therefore with regard to arsenic, the arsenic-bearing minerals are in contact with water under oxidizable conditions.

4.1.3 Hydrochemical modelThe computer program PHREEQC (Parkhurst, 1995) windows version was used to calculate the equilibrium speciation and saturation indices. In addition, it estimates the E

h of the chemical

medium. Additionally, it is able to simulate a wide range of geochemical reactions including mixing of water, dissolving and precipitating phases to achieve equilibrium with the aqueous phase and effects of changing temperature. Besides, it provides estimates of element concentrations that had not been determined analytically as well as of molalities and activities of aqueous species, pH, pe, and saturation indices. Also, it indicates mineral species.

As illustrated in the Figue 5, for the three extracted pore water samples, the calculated Eh values

are considerably lower than the measured ones. In addition, the tailings environment becomes under post aerobic or reducing conditions with depth mainly at the interval of depth that coincide with the beginning of the processed material. This is due probably to the presence of an appre-ciable amount of humic substances in the processed material on one hand, and to the absence of any supplying of oxygen on the other. The considerable difference between the recorded and the calculated values is probably due mainly to the influence of the climatic factors onto such measurements.

Comparing the recorded values for the extracted pure and mixed pore water samples of the mentioned three sediments as demonstrated in the Figure 6, the calculated E

h values by means of

the arsenic species are considerably lower. This is due to the influence of the change of the used species for modelling on the one hand and of the added de-ionised water that affect the chemistry of the analysed sediments on the other hand. In addition, for all extracted mixed pore water sam-ples, the recorded values illustrate the change of the chemical conditions between the heap materi-als and the tailing sediments characterised by the considerable decrease of the mentioned factor with depth that indicates the change of the medium to post aerobic or anaerobic conditions. Since

0

1

2

3

4

5

6

7

8

9

10

6 7 8 9 10

pH (Borehole N°1) D

epth

(m

)

0

5

10

15

20

6 8 10 12

pH (Borehole N°2)

Dep

th (

m)

Figure 4. The change with depth of the pH value.


239

large amounts of trace metals are bound to nodules, the anoxic conditions of the treated material favour the dissociation of such compounds i.e. the supplying of the pore water by such contami-nants. The recorded results demonstrate the usefulness of the used procedure and the PHREEQC model for qualitative resolving geochemical problems.

Although most phases containing arsenic are undersaturated for almost all analysed samples, the high saturation of all extracted waters with the phase BaHAsO

4H

2O (Planer-Friedrich et al., 2001)

may explain its relative high content in these extracts. In addition, AsO4

3–, H2 AsO

3–, H

3 AsO

3 and

HAsO4

2– are the most recorded arsenic species with the dominance of the last mentioned one with a share exceeding 60%. In spite of the ability of As to be exchanged with clay minerals in its anion form AsO

43– replacing OH– ion, the low amount of the mentioned anion that not exceed 4% will

not significantly reduce the mobile arsenic. Moreover, the chemical conditions of the area as well as the low amount of sulphur in the tailings material will not reduce the aqueous arsenic despite its chalcophilic tendency to form sulphide complexes. Therefore, its adsorption to nodules mainly amorphous iron oxide is the only fact that may disable the migration of the aqueous arsenic.

4.2 Sequential extraction procedure

In the treated sediments, the arsenic content varies between 170 and 690 ppm. 29 to 66% of the total arsenic is held in the crystal lattice (Figures 7; 8) with only a minor part in association with clays as is demonstrated by the different methods of analysis (total dissolution together with X-ray fluorescence analysis). In addition, according to the work of Onishi & Sandell, (1955), the most of the arsenic in mentioned phase is bound to silicate minerals such as quartz, feldspars and hornblende (amphibole). Moreover, the relative high As concentration in these minerals results from the ease with which this element substitutes for Si, Al, or Fe in their crystal lattices (Onishi & Sandell, 1955).

Besides, although due to the chalcophilic tendencies of this element to form sulphide com-plexes under anaerobic conditions (Mc Bride, 1994) as well as the presence of many As-sulphide minerals such as arsenopyrite, realgar, orpiment and arsenolite in the most uranium deposits of the Ore Mountains; no association of this element with the sulphide-organic phase was detected. This may be due to the dissolution of most As-sulphide compounds by the influence of the

0

1

2

3

4

5

6

7

8

9

10

-400 -300 -200 -100 0

Eh (mV) Borehole N°2

Dep

th (

m)

Figure 5. The change with depth of Eh; calculated and

measured values.

Figure 6. The change with depth of calculated Eh.

3.0

3.5

4.0

4.5

5.0

100 200 300 400 500

Measured value Calculated value

Eh (mV)

Dep

th (

m)


240

mineral processing mainly under oxygen attack for some minerals. In addition As demonstrates a low affinity to organic species in comparison to other chemical components (Gluskoter et al., 1977).

Further, since As has a high affinity for Mn oxides (Sadiq et al., 1983; Galba & Polacek, 1973) as well as to Fe oxides and hydroxides (Gile, 1936; Jacobs et al., 1970; Woolson et al., 1971; Deuel & Swoboda, 1972; Fordham & Norrish, 1979), the most important of its non residual amount is found in association with the nodular hydrogenous fraction ranging between 75 and 98% with the dominance of the moderately reducible phase. These findings are in agreement with those of Jacobs et al., (1970) which confirms the strong sorption of this element mainly to the amorphous iron oxide.

Furthermore, only a trace amount of As from its non residual fraction is in association with the carbonatic phase which varies between ∼0 and 2% and these results support those reported by

1 2 3 4 5 6 7 8 90

20

40

60

80

100

Soil samples in different depth intervals: 1. 0.5-1.5m 2. 3.5-4.5m 3. 5.5-6.5m 4. 6.5-7.5m 5. 8.5-9.5m 6. 11.5-12.5m 7. 14.5-15.5m 8. 17.5-18.5m 9. 19.5-20.2m.

Port

ion

(%)

Pore water Exchangeable Carbonatic Easily reducible Moderately reducible Sulfid./org.

Uranium Tailings SchneckensteinSelective extraction procedure (As element)

Figure 8. Partition of arsenic in 6 different phases.

1 2 3 4 5 6 7 8 90

20

40

60

80

100

Soil samples in different depth intervals: 1. 0.5-1.5m 2. 3.5-4.5m 3. 5.5-6.5m 4. 6.5-7.5m 5. 8.5-9.5m 6. 11.5-12.5m 7. 14.5-15.5m 8. 17.5-18.5m 9. 19.5-20.2m.

Port

ion

(%)

Pore water Exchangeable Carbonatic Easily reducible Moderately reducible Sulfid./org. Residual/ aqua-regia Residual/ X-ray.F.A

Uranium Tailings SchneckensteinSelective extraction procedure (As element)

Figure 7. Partition of arsenic in 8 different phases.


241

Yan-Chu, (1994). This phenomenon may be due to the very low affinity of As to carbonates with respect to other components mainly at pH under pH 11.

Since As is exchangeable in its ionic form onto clays only at low pHs and due to its high affinity to be sorbed by Fe-Hydroxides, only a negligible amount of As is associated with the exchangeable phase. However, on the other hand, a reasonable amount of the non residual fraction of arsenic is found in association with the pore water phase which ranges from ∼0.6 to 14%. This is due to the increase in the solubility of As with the increase of pH under alkaline conditions (Brookins, 1988; Yan-Chu, 1994).

5 CONCLUSIONS

The hydro geochemical model PHREEQC offers valuable information about the arsenic species in the interstitial waters. Also it summarises the chemical conditions affecting the behaviour of metals and metalloids including arsenic. Of its share, the sequential extraction provides significant information concerning the compound forms of arsenic in the study area. Therefore, these two procedures complete each other.

Basing on the results, independent of the chemical conditions of the study areas, arsenic seems to be highly soluble since a considerable amount of its non residual fraction is in association with the pore water phase. Moreover, the oxygen consumption enhance its solubility since most of its non residual fraction is in association with nodules and only high amounts of sulphates may reduce its soluble contents in pore waters.

REFERENCES

Braman, R.S., (1975): Arsenic in the environment. In: Woolson EA, editor. Arsenical pesticides. Washington, DC7, Am. Chem. Soc. 108–23.

Brookins, D.G., (1988): Eh-pH diagrams for geochemistry. Springer-Verlag, 175 pp.Challenger, F., (1945): Biological methylation. Chem. Rev. 36, 315–61.Chen, H.W. et al., (1999): J. Am. Water Works Assoc. 91, 74–85.Chwirka, J.D. et al., (2004): J. Am. Water Works Assoc. 96 (3), 106–114.Clifford, D.A. et al., (2003): J. Am. Water Works Assoc. 95 (6), 119–130.Cullen, W.R. & Reimer K.J., (1989): Arsenic speciation in the environment. Chem. Rev. 89, 713–64.Deuel, L.E. & Swoboda, A.R., (1972): Arsenic solubility in a reduced environment. Soil Sci. Soc. Am. Proc.

36, 276–278.DIN 38414, Part 7: Deutsche Einheitsverfahren zur Wasser-, Abwasser- und Schlammunter suchung. Auf-

schluß mit Königswasser... Beuth- Verlag Berlin Köln. Germany.Fordham, A.W. & Norrish, K., (1979): Arsenate- 74 uptake by components of several acidic soils and its

implications for phosphate retention. Aust. J. Soil Res. 17, 307–316.Frankenberger Jr., W.T. (Ed.), (2002): Environmental Chemistry of Arsenic, Marcel Dekker, New York.Galba, J. & Polacek, S., (1973): Sorption of arsenates under kinetic conditions in selected soil types. Acta

Fytotec. 28, 187–197.Gile, P. L., (1936): The effect of different colloidal soil materials on the toxicity of calcium arsenate to millet.

J. Sagric. Res. 52, 477–491.Gluskoter, H.J. et al., (1977): Trace elements in coal: Occurrences and distribution. Circ._Ill. State Geol. Surv.

499, 153.Gorby, M.S., (1994): Arsenic in human medicine. In: Nriagu J.O., editor. Arsenic in the environment, part II,

human health and ecosystem effects, NY7 John Wiley & Sons, INC, 1–16.Gray, J.R. et al., (1989): Acute arsenic toxicity: an opaque poison. Can Assoc Radiol J; 40, 226–227.Gulens, J. et al., (1979) : Champ DR, Jackson RE. Influence of redox environments on the mobility of arsenic

in ground water. In: Jenne EA, editor. Chemical modeling in aqueous systems, ACS Symposium Series 93. Washington, DC7, Am. Chem. Soc. 81–95.

Institut Fur Wasserwirtschaft, (1958): Abflußhöhen- u. Niederschlagskarte f. das 20jährige Mittel (1921–1940), Karl-Markx Stadt, M 1: 200 000 Institut f. Wasserwirtschaft Berlin.


242

Jacobs, L.W. et al., (1970): Arsenic sorption by soil. Soil Sci. Soc. Am. J. 34, 750–754.Lamble, K.J. & Hill S.J., (1996): Arsenic speciation in biological samples by online high performance liquid

chromatography-microwave digestion-hydride generation-atomic absorption spectrometry. Anal. Chim. Acta 334, 261–70.

Le, X.C. et al., (2004): Anal. Chem. 76, 26 A–33 A.McBride, M.B., (1994): Environmental chemistry of soils. Oxford University Press, New York, USA,

406 pp.National Research Council (NRC), (1999): Arsenic in Drinking Water, National Academy Press, Washington,

DC, 305 pp.Nriagu, J.O. (Ed.), (1994): Arsenic in the Environment. Part I: Cycling and Characterization, Wiley, New

York.Onishi, H. & Sandell, E.B., (1955): Geochemistry of arsenic. Geochimica & cosmochimica acta, 7, 1–33.Parkhurst, D.L., (1995): PHREEQC, a computer program for speciation, reaction-path, advective-trans-

port, and inverse geochemical calculations. Water-resources Investigations report 95–4227. Lakewood, Colorado.

Planer-Friedrich et al., (2001): Investigations on arsenic in the groundwater of the Rioverde basin, Mexico.- Environmental Geology, 40 (10), 1290–1298.

Saha, K.C., (2003): Review of arsenicosis in West Bengal, India: a clinical perspective. Crit Rev. Environ. Sci. Technol. 30, 127–63.

Sadiq, M. et al., (1983): Environmental behaviour of arsenic in soils: Theoretical. Water, Air, Soil Pollt. 20 (4), 369–377.

Salomons, W. & Forstner, U., (1984): Metals in the hydrocycle, Springer Verlag, Berlin-Heidelberg, 349 pp.Senesse, P. et al., (1999): Cronkhite–Canada syndrome and arsenic poisoning: fortuitous association or new

etiological hypothesis? Gastroenterol Clin. Biol. 23, 399–402.Sheppard, S.C., (1992): Summary of phytotoxic levels of soil arsenic. Water Air Soil Pollut. 64, 539–50.Squibb, K.S. & Fowler B.A., (1983): The toxicity of arsenic and its compounds. In: Fowler B.A., editor. Bio-

logical and environmental effects of arsenic. Amsterdam7 Elsevier, 233–69.Thomas, D.J. et al., (2001): The cellular metabolism and systemic toxicity of arsenic. Appl. Pharmacol. 176,

127–44.Wagman, D.D. et al., (1982): The NBS tables of chemical thermodynamic properties. Selected values for

inorganic and C1 and C2 organic substances in SI units. J. Phys. Chem., 11 (2), 392.Wang, S. & Mulligan, C.N., (2006): Occurrence of arsenic contamination in Canada: Sources, behavior and

distribution, Science of the Total Environment 366, 701–721.Wang, L. et al., (2002): J. Am. Water Works Assoc. 94, 161–173.Watt, C. & Le, X.C., (2002): Biogeochemistry of Trace Elements, in: Y. Cai, O.C. Braids (Eds.), American

Chemical Society Symposium Series, ACS, Washington, DC, pp. 11–32.Welch, A.H., (2000): Ground Water 38, 589–604.Welch, A.H. et al., (1988): Ground Water 26, 333–347.WHO, (1993): Guidelines for drinking water quality, 2nd ed. Recommendations. World Health

Organization.Woolson, E.A. et al., (1971): Correlation between available soil arsenic , estimated by six methods, and

response of corn. Soil Sci. Soc. Am. Proc. 35 (1), 101–105.Xu, H. et al., (1988): Influence of pH and organic substance on adsorption of As(V) on geologic materials.

Water Air Soil Pollut. 40, 293–305.Yan-Chu, H., (1994): Arsenic distribution in soils, in: Nriagu, J. O. (ed): Arsenic in the environment. Part I:

Cycling and characterisation. John Wiley & Sons, Inc.Yong R.N. & Mulligan C.N., (2004): Natural attenuation of contaminants in soils. Boca Raton7 CRC Press.


243

Geochemical characterization of proposed waste dumps over time and space

Larry Breckenridge & Amy HudsonTetra Tech Inc., Littleton, MA, USA

Susan PoosMarston and Marston Inc., Lakewood, CO, USA

Dan ThompsonGold Reserve Inc., Spokane, WA, USA

ABSTRACT: Most mining operations have geochemically-heterogeneous waste rock charac-teristics that are a challenge for mine planning and environmental management. Any attempt to blend potentially acid generating and carbonate-rich waste, or any effort to predict environmen-tal impacts of mine waste leachate must incorporate a method to characterize the geochemical characteristics of the waste over time and space as the dump is being constructed. The Brisas del Cuyuni open pit gold and copper mine, a project run by Gold Reserve Inc., plans to construct two waste rock dumps to accommodate 952 million tonnes of waste. The project team integrated the geologic block-model, mine planning, and geochemistry to determine the geochemical char-acteristics of the waste over time and space in the waste rock dumps. The results were used for geotechnical stability modeling, unsaturated flow modeling, and geochemical modeling of leach-ate water quality.

1 INTRODUCTION

The Mine is a proposed 70,000 tonnes per day open pit copper-gold project situated in a tropical environment. The mine plan currently involves producing an open pit 2.5 kilometers long, 1 kilometer wide, and over 500 meters deep. After its 20 year forecasted mine life, the mine will have built two Waste Rock Deposition Areas (WRDA) containing a total of 952 million tonnes of waste rock. Due to the presence of sulphides in the waste, acid rock drainage (ARD) must be mitigated throughout the mine life and upon closure.

The Mine has a challenging environment for mitigating ARD from mine facilities. Approxi-mately 3.5 meters of rain a year falls at the mine, primarily during the wet season stretch-ing from May to October. However, moderate rains persist throughout the year. The Mine is located in an impacted tropical rainforest environment characterized by flat topography and dense vegetation.

Not all of the waste rock produced is acid-generating. Significant calcite concentrations, typi-cally in the form of calcite fracture-filling and calcite veins, are present in the rock. Calcite con-centrations increase with depth. As a result, the mine plans to construct WRDAs with potentially acid generating (PAG) waste encapsulation, and non-potentially acid generating (Non-PAG) waste blending. Waste blending is becoming more and more common for large mining operations, and because of the quantity of materials involved, the blending plan must be optimized as part of



244

the mine final design. Proper planning of a waste blending and management plan requires the following:

− Characterizing the waste types into specific waste categories based on geochemical behavior and material type;

− Quantifying the tonnes of each waste category in-situ in the pit;− Determining how much of each waste category will be produced each year based on the mine

plan;− Determining the construction requirements of the WRDAs throughout the mine life;− Balancing the tonnes excavated, the WRDA construction requirements, and the rock geochem-

istry to minimize waste re-handling; and− Determining the final geochemical behavior of the waste in the WRDA over time and space to

provide a key input to geochemical modeling.

Developing a successful waste handling plan involves a significant amount of spatial geochemi-cal data and the integration of the fields of mine planning and mine environmental management.

2 GEOCHEMICAL CHARACTERIZATION

The Mine lies in the granite greenstone belt of the Pastora province of the Guyana Shield. Table 1 presents the geologic units expected to be encountered at the Mine.

The geochemical characterization involved testing core from the exploratory drilling program. The initial geochemical characterization of the units described in Table 1 was accomplished using a variety of static geochemical tests, including:

− Acid Base Accounting (ABA) testing;− Whole Rock Analysis (by Aqua Regia);− Net Acid Generation pH test (NAG pH); and− Leach Extraction.

Static testing defined the potential geochemical behavior of the material. These tests are designed to characterize the acid generating and acid neutralizing potential, the overall material composition, and the metals that are readily leachable.

The acid neutralization potential (ANP) and the acid generating potential (AGP) are generally expressed as equivalent tonnes of calcium carbonate (CaCO

3) generated or neutralized per kilo-

tonne of material (T/KT). A sample would be defined as acid generating if the net neutralization potential (NNP = ANP-AGP) was less than 0 T/KT. In practice, the risk of ARD has been found to be highest for samples with NNP values less than −20 T/KT, and low when the NNP is greater than +20 T/KT (Price et al, 1997). NNP values were used as the primary geochemical analysis for the study. The material between −20 T/KT and 20T/KT has moderate/uncertain acid generation potential and generally requires additional characterization to fully define.

Further characterization of the geologic units at the Mine was accomplished through kinetic testing of the material. Select samples were tested using laboratory humidity cells and on-site leach columns. These tests are designed to characterize the rate of acid generation or neutralization.

Table 1. Geologic units at the mine and generalized geochemical behavior.

Depth start Depth end

Material type Meters below surface Physical description ARD potential

Oxide saprolite 0 20–30 Fully-leached saprolitic soil NoneSulphide saprolite 20–30 40–70 Sulphide-rich saprolite soil HighWeathered bedrock 40–70 60–100 Heavily weathered and altered VariableUnweathered bedrock 60–100 +500 Unweathered and unfractured Net neutralizing


245

Laboratory humidity cell testing is designed to simulate ideal acid generation under controlled conditions, while the on-site leach columns are designed to consider site specific climate and microbial activity. These tests are intended to be run over a long period of time (20 weeks or greater) to allow for changes in geochemical behavior to be observed. Some of the on-site kinetic tests at the Mine have been running for nearly three years.

Testing has confirmed the ARD potential listed in Table 1. Kinetic testing has proven that although deeper samples may have pyrite concentrations as high as 3%, fresh pyrite crystals offer significant resistance to weathering and these samples have been resistant to acid production even after three years of kinetic testing. As a result, it appears that a combination of a well-designed waste cover, waste blending, and waste encapsulation will prevent a majority of the waste from generating acid.

3 IN-SITU CHARACTERIZATION OF GEOCHEMICAL WASTE

The first step in determining the geochemical behavior of the WRDA, is to determine the geo-chemical behavior of the waste in-situ prior to mining.

3.1 Defining waste categories

All mine waste was categorized by material type and geochemical behavior. Geochemically, the waste was broken into five groups:

− Oxide saprolite: geochemically inert− Sulphide saprolite: Acid generating− Acid generating rock: NNP<−20 T/KT− Rock of uncertain geochemical behavior (Uncertain Rock: −20<NNP<20 T/KT).

Weathered and unweathered bedrock could be grouped if they had similar geochemical behav-ior because they also had similar geotechnical properties once blasted.

3.2 Determination of the in-situ quantities of the waste categories

Most modern mining operations use three-dimensional computer modeling to determine ore grades within the pit. Complex geologic models are created by applying geostatistics to bore hole assay data with the resulting model providing a reasonable forecast of grades and tonnages for mine planning. This same technique was applied to geochemical properties.

Marston used the same general approach to determining neutralization potential as would be used to determine gold and copper values. As a first step in modeling, various statistics were calculated for the NP, AP, and sulphide values in the databases including minimum, maximum, and average values and log probability graphs were generated. These variables were evaluated because they represent the base data from which all the other parameters were calculated. The statistical analysis indicated that the NP and AP values needed to be capped to exclude outlying samples.

Waste characterization modeling was limited to the rock types weathered and hard rock. Oxide and sulphide saprolite were not included in the modeling because their neutralization potentials could be defined and understood without modeling. Marston developed interpolated models for NP, AP, and sulphide from which values for NNP were calculated. Additionally, Marston devel-oped a model to define ore and waste and four different models to categorize the waste according to different criteria.

Interpolation of NP, AP, and sulphide block values was done using inverse distance cubed meth-ods. Search ellipses were oriented the same as those used in modeling the gold in order to follow the mineralization. Distances used in the search were extended in order to fill the block model. There are 335 data points for waste characterization.


246

Figure 1. Proportions of the waste categories in the pit.

Figure 2. Ultimate pit geochemical characterization colored by waste category: plan view.


247

Using the interpolated block values for NP and AP, the values for the NNP model was calculated. Block values for the model Ore and Waste were set to 1 for ore and 2 for waste. Using a combination of these criteria the waste rock was categorized into the five categories based on a combination of rock type and the four waste categories. Figure 1 shows the resulting proportions of these five categories.

Marston reported the annual waste production by category. By color coding the waste catego-ries Marston could determine where the different types of waste were being generated so that the production plans could take this into account so that the material was placed in the appropriate location in the WRDA. Figure 2 shows the ultimate pit colored by waste category in plan view, Figure 3 shows it in a side view (looking East), and Figure 4 shows it in a three-dimensional (3-D) isometric view looking North.

Figure 3. Ultimate pit geochemical waste characterization, showing the west pit wall.

Figure 4. Ultimate pit geochemical waste characterization, north-looking 3-D isometric view.


248

Additionally, in the production planning haulage routes to the WRDA were adjusted to account for material placement in order to balance truck requirements through out the mine life. By track-ing the waste production on an annual basis the costs associated with the handling and storage of PAG material could be minimized.

4 WASTE ROCK DEPOSITION AREA CONSTRUCTION

In order to accommodate the flow path of a major site drainage, the waste must be placed in two separate WRDAs, a smaller cell northwest of the pit (North WRDA), and a larger cell to the southwest of the pit (South WRDA). Both cells are designed to encapsulate PAG waste, and in order to build a geotechnically-stable WRDA, the following construction criteria will be applied to the WRDAs:

− WRDAs will be buttressed with Non-PAG hard rock;− Sulphide saprolite, the greatest acid-producing waste category, will be under laid with com-

pacted oxide saprolite, buttressed by compacted oxide saprolite, and covered with oxide sapro-lite to create a full-closed containment cell;

− Hard rock waste from later in mining will be stacked on top of the cell, with appropriate but-tresses for stability; and

− Upon closure, the WRDA will be covered with an evapotranspiration soil cover constructed with compacted oxide saprolite and topsoil. It will also be vegetated with a dense community of native plants.

Figure 5 is a cross-section of the North WRDA once it is closed and capped. The South WRD has a similar design, but is larger. A cross-section of the completed South WRDA (end of mine life) is shown in Figure 6.

As can be seen in Figures 5 and 6, the WRDAs are engineered structures requiring signifi-cant effort and planning to construct. These are not haphazard piles created from uncoordinated

Figure 5. North WRDA construction.

Figure 6. South WRDA construction.


249

truck dumping, instead, these are engineered waste management facilities designed to minimize environmental impacts.

One of the largest challenges of constructing the WRDAs is that the construction requirements of the WRDAs must be managed in the context of the mine plan. Based on this plan, different waste is removed each mining year. Integrating the spatial geochemical model of mine waste with the mining plan, the project team defined the quantity of each waste category removed each year during the mine life (see Figure 7).

In the early years of mining, there is little hard rock extracted, particularly Non-PAG hard rock because saprolite materials in the upper stratigraphic column are being removed (see Table 1). Later in the mine life, it becomes necessary to stockpile oxide saprolite to form the encapsulation buttresses and for the cover because only hard rock is available. Production rates for these materi-als and WRDA requirements were taken into account in the mine planning in order to minimize the re-handling and haulage costs associated with WRDA. This was done by strategically selecting stockpile locations and timing the placement of waste in the north and south WRDA.

5 CHARACTERIZATION OF MINE WASTE TO BE PLACED IN THE WRDA

Once the geochemical properties of the stream of waste rock were defined, the next step was to give each category a quantifiable geochemical property that could be used to determine how acid-generating or how acid neutralizing the WRDA material would be over time and space.

The parameter selected to define waste was NNP in tonnes of CaCO3 equivalent neutralization

capacity per kiloton of waste (see Section 2). This value was selected because it provides a single easy-to-evaluate number to quantify waste.

Oxide and sulphide saprolite were quantified using a single NNP value. Oxide saprolite has a zero NNP, and sulphide saprolite was found to be consistently on the order of −70 T/KT.

Figure 7. Cumulative quantities of each waste category by year of mining life.

0

50

100

150

200

250

300

350

400

0 5 10 15 20 25

Mining Year

Cu

mu

lati

ve

s T

on

ne

s o

f W

as

te (

in m

illi

on

s o

f to

nn

es

)

OX-SAP (Non-PAG)

SULF-SAP (PAG)

Total PAG Rock

Total Non-PAG Rock

Mixed Waste


250

Hard rock waste, a mixture of Non-PAG, PAG, and Mixed Rock waste categories makes up the majority of the material in the North and South WRDAs. Because this material is a mixture of the waste types defined in Section 2.3, it is important to determine its geochemical properties over time and space. The steps below describe the accounting techniques used to determine the geochemical characteristics of the mixed rock.

− Step 1: Determine the quantities of Non-PAG, Mixed, and PAG material (both weathered and unweathered) coming from the pit in a given year.

− Step 2: Determine the quantity of Non-PAG material placed in the WRDA buttresses (available from current mine planning).

− Step 3: Subtract the quantity of Non-PAG material used in WRDA buttressing from the total quantity of Non-PAG material. The remaining quantity of Non-PAG becomes part of the Mixed Rock waste.

− Step 4: Calculate the year-by-year weighted average of NNP values considering the quantities of PAG and Mixed material determined in Step 1 and Non-PAG material determined in Step 3.

This results in a NNP value for the hard rock waste by year, which is presented in Figure 8.Values above the bold line at 20 T/KT signify waste that is net neutralizing, values below the bold

line at −20 T/KT signify acid generating, and values in the middle are “transition” material that could be either acid neutralizing or acid generating. One can see that earlier in the mine life, the material is net acid generating. As the mine encounters more calcite, this acid-generating potential is mitigated, but the cumulative NNP of the hard rock waste at the end of the mine life is above 0 NNP.

6 WRDA WASTE CHARACTERIZATION OVER TIME AND SPACE

With the hard rock parameters defined, all of the waste material can be characterized over space and time. Each year there is a corresponding increase in the construction of the WRDAs as the

-40

-30

-20

-10

0

10

20

30

40

0 2 4 6 8 10 12 14 16 18 20

Year

NN

P (

T/K

T C

aC

O3)

Mixed Waste Yearly NNPCumulative Mixed Waste NNP

Figure 8. Hard rock waste NNP by mine year.


251

waste is placed. Therefore, the year-by-year results of Figure 8 can be correlated to the year-by year WRDA construction schedules.

This results in one of the key goals of the project—the definition of quantifiable geochemical properties over time and space in the WRDA. Results of the post-closure WRDA geochemical characterization are shown in Figures 9 and 10.

7 PREDICTIVE MODELING OF WRDA LEACHATE QUALITY AND QUANTITY

Predictive modeling of the quantity and quality of WRDA leachate was the final step in the project. This was accomplished using unsaturated flow modeling and geochemical modeling. The predictive model utilized the waste characterization model as part of the following series of analysis:

− Representative vertical flow paths were selected through the WRDA;− These flow paths encountered material with the geochemical behavior as defined in Figures 9

and 10;− Unsaturated flow modeling determined leachate quantity and the travel time of leachate passing

through different material horizons;− Geochemical modeling determined the water quality based on dilution, equilibrium, and kinet-

ics. The reaction time with each material was determined by Step No. 3 above; and− Final leachate quantity and quality was determined by mixing different flow paths in the aquifer

beneath the WRDA.

This resulted in a predictive model of the leachate quality and quantity of the WRDAs. This information was critical to mine planning and environmental permitting.

Figure 10. South dump geochemical characterization.

Figure 9. North dump geochemical characterization.


252

8 CONCLUSIONS

This product of the waste characterization can give mine planners a powerful new tool. With these analyses, mine operators can optimize a mine plan that takes environmental considerations into account. If a WRDA has a zone that could be a serious ARD-producer, the mine plan can be adjusted to better blend the waste from this area with acid-neutralizing waste. In almost every cir-cumstance, prevention is far more cost-effective than treatment. In addition, a fully-characterized WRDA will show regulators that a mine is aware and proactive relative to issues of WRDA acid leachate generation risk.

REFERENCE

Price W.A. 1997. Draft: Guidelines and Recommended Methods for the Prediction of Metal Leaching and Acid Rock Drainage at Minesites in British Columbia. British Columbia Mine Reclamation Services (MRS). Smithers BC: British Columbia Ministry of Employment and Investment, Energy and Minerals Division.


253

Stochastic prediction of mine site water balance, Gilt Edge Mine Superfund Site, Lawrence County, South Dakota

M. Nelson, S. Fundingsland, G. Hazen & P. HightCDM Inc., Denver, CO, USA

V. KetellapperUS Environmental Protection Agency, Denver, CO, USA

ABSTRACT: Water balance management is a critical component of both environmental man-agement at active mines and environmental remediation of closed or abandoned mines. Failures in mine water management have led to environmental impacts across the globe ranging from rela-tively local impacts caused by short-term uncontrolled discharges to major environmental catas-trophes such as a January 30, 2000 release at Baia Mare, Romania, which affected aquatic life in the river Tisza for hundreds of kilometers. Managing the water balance at a large-scale mine is generally not problematic under average conditions. Water inflows and outflows are measured (or estimated) and a mass continuity expression is developed to describe inflows, outflows, and resulting changes in storage. Water balance problems and associated uncontrolled releases are commonly triggered by uncommon climatic conditions.

To reduce the risk of an uncontrolled release, it is advantageous to incorporate uncertainty analysis into water balance evaluations. This process facilitates improved site management deci-sions, because uncertainties in numerous site variables are propagated through the water balance evaluation and the results are presented in terms of the probability that a future event may or may not occur. Site management decisions can then be based on a risk level that is deemed acceptable based on site-specific factors such as the sensitivity of the environment to a potential release, chemical hazards present in mine water, and costs of achieving lower risk levels. A stochastic evaluation of the mine site water balance was completed at the Gilt Edge Mine Superfund Site. This evaluation is presented as a case study, which describes the application of uncertainty analy-sis in water balance management.

The Gilt Edge Mine is an abandoned heap leach gold mine, which produces approximately 378 million liters of Acid Rock Drainage (ARD) in an average year. ARD must be collected and treated prior to discharge to prevent adverse effects to downstream coldwater fisheries and pub-lic drinking water supplies. A forward-looking stochastic method was developed to evaluate the water balance at Gilt Edge mine. This method relies on development of mass continuity expres-sions from empirical data; development of probability distributions describing climatic condi-tions, ARD yield, and water treatment plant performance; and Monte Carlo simulation of potential future conditions. The method is used to predict the volume of ARD that may require collection and treatment in the future to prevent an uncontrolled release at the site.

1 INTRODUCTION

The Gilt Edge Mine is an abandoned heap leach gold mine, which operated intermittently over the last 130 years. It is located in a mountainous region of South Dakota called the Black Hills. Large-scale surface mining operations were conducted at the site during 1986 to 1997 by Brohm Mining Company (Brohm). In 1999, Brohm’s parent company, Dakota Mining Corporation, declared bankruptcy in Canada, and Brohm was forced to abandon the site. At that time, there



254

was an imminent threat of an uncontrolled release of ARD from a pit lake (Sunday pit) and one of the large waste rock dumps at the site (Ruby Gulch). The Sunday pit contained approximately 568 million liters of ARD stored behind a temporary rock dam, and ARD storage facilities at Ruby Gulch were only capable of preventing an uncontrolled release for several days without operation of the pumpback system. An uncontrolled release from the site could affect downstream drinking water resources and sensitive cold water fisheries.

After the site was abandoned in 1999, the State of South Dakota immediately responded and took over site operations. These operations included ARD treatment and operation and mainte-nance of ARD pumpback infrastructure. In early 2000, the US Environmental Protection Agency (EPA) assumed responsibility for site operations and maintenance under the emergency response program authorized by the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA). The site is currently managed by EPA and the State of South Dakota under the remedial program of CERCLA.

2 MANAGING RISK OF UNCONTROLLED RELEASE

An imminent threat of an uncontrolled release was clearly present at the Gilt Edge Mine in 1999 when the site was abandoned. At that time, the State of South Dakota decided that an immediate response was necessary to mitigate that risk. This emergency response required investment of sig-nificant capital, which demonstrates the difficult decisions that managers of active or abandoned mines face when evaluating potential allocation of limited funding. Understanding the probability of a potential release assists site decision-makers in determining an appropriate tolerance to risk. However, selecting an appropriate probability threshold (for example, a 95-percent probability that a release will not occur) is dependant on site-specific conditions such as the sensitivity of down-stream resources, the regulatory environment, the toxicity of mine water to human or eco-logical receptors, and realities of potential funding mechanisms. Although a 100 percent prob-ability that an uncontrolled release will not occur is the goal, it may be unrealistic from a funding, climatic, or operational perspective.

The Gilt Edge Mine is used as a case study to present a method that can assist site decision-makers with difficult decisions regarding allocation of capital for mine water management to reduce the risk of an uncontrolled release. In the case of the emergency response at Gilt Edge mine in 1999, the decision was relatively simple due to the imminent threat of an ARD release and the potential impacts to drinking water resources and sensitive ecological resources. At mines that are adequately managing the site water balance under average conditions, a decision to invest in mine water management is more difficult. However, water balance problems and associated uncontrolled releases are commonly triggered by uncommon climatic conditions. The stochastic approach presented in this paper is a means by which site decision-makers can understand the probability that an uncontrolled release may occur and can evaluate the potential reduction in this risk that may be realized by additional investments in mine water management.

Evaluations of mine water balances may be based on deterministic or stochastic methods. In a deterministic approach, the inputs to the mass continuity expression are fixed and based on factors such as average precipitation, average mine water yield, or average water treatment rates. Deter-ministic methods may also consider sensitivity analyses incorporating a range of inputs (for exam-ple, 75th- or 95th- percentile precipitation rather than average precipitation). The advantage of a stochastic approach is that the method does not require a set of deterministic (or fixed) inputs. A probability distribution is developed for each of the inputs to the mass continuity expression, and the uncertainties in each input are propagated through the evaluation by Monte Carlo simulation.

The risk of an uncontrolled release at the Gilt Edge Mine can be understood in terms of two potential mechanisms. The first is inadequate water treatment and discharge rates in relation to the rate of ARD inflows and available storage capacity. The second mechanism is a release related to a large precipitation event, which overwhelms the capacity of individual ARD collection systems to pump ARD into storage prior to treatment. The case study presented in this paper evaluates the


255

first mechanism and incorporates evaluation of the site water balance and Monte Carlo simulation of the future water balance. The future water balance is then compared with available ARD storage capacity to estimate the probability that the capacity will be exceeded in future years.

3 EVALUATION OF SITE WATER BALANCE

Understanding the site water balance is a critical component of ARD management. The water bal-ance is an evaluation of the volume of ARD generated at the site, the volume of stored ARD, and the performance of water management activities in maintaining this water balance at acceptable levels. The current estimate is based on evaluations of the site water balance for water years 1999 to 2007.

3.1 Data sources used to estimate water balance

3.1.1 Climate dataThe volume of ARD generated each year is directly related to the amount of precipitation received at the site. The current water balance estimate is based on precipitation data from a weather sta-tion located in the town of Lead, South Dakota. The Lead station is part of the National Weather Service Cooperative Station Network. Data from the station are validated prior to posting by the National Oceanographic and Atmospheric Administration. This station has an extensive period of record extending from May 1, 1948 to the present, which makes the data particularly useful for statistical analyses of precipitation variations. The Lead station is located at an elevation of 1631 meters (m). The Gilt Edge site is approximately 10 km southeast of the Lead station with elevations ranging from 1520 to 1740 m.

It is possible that mountainous topography causes orographic variations in precipitation bet -ween Lead, South Dakota, and the Gilt Edge Mine. Driscoll et al, (2000) described marked oro-graphic variations in precipitation in the Black Hills, with the highest precipitation occurring in the Lead area and lower precipitation occurring in other parts of the Black Hills. Dricoll et al., (2000) developed an isohyetal map, which shows contours of annual precipitation. The map is based on a geostatistical analysis of precipitation data from Black Hills weather stations. Stations in the area of Gilt Edge mine used in the Driscoll et al., (2000) analysis include a US Geological Survey (USGS) station along Bear Butte Creek approximately 3 kilometers (km) southwest of the site, a USGS station approximately 8 km north of the site near the town of Deadwood, and the Lead station. The Driscoll et al., (2000) analysis suggests that actual precipitation at the Gilt Edge Mine site may be about 5 centimeters (cm) lower than precipitation measured at Lead as a result of oro-graphic effects. However, the Lead data is used in this evaluation because of the extensive period of record and consistent correlation between the Lead data and ARD generation at the site.

3.1.2 Measurement of stored ARDARD at the site is stored in two major impoundments, Sunday Pit and Anchor Hill Pit. Additional impoundments collect and store ARD prior to conveyance into one of the larger impoundments. These impoundments include the stormwater pond, Pond E, Dakota Maid Pit, and subsurface collection galleries at Hoodoo Gulch and Ruby Gulch. The volume of ARD stored on the site is routinely measured by surveying the water level of the impoundments and comparing the sur-veyed elevation with site-specific tables that relate the elevation in the pit lake with the estimated volume. These tables were developed based on as-built surveys of the facilities and subsequent volumetric modeling. The frequency of surveys generally ranges from weekly to monthly, which provides adequate data for the annual water balance evaluations.

3.1.3 Measurement of ARD inflowsMeasurement of discrete ARD inflows is conducted where practicable. ARD inflows are measured at discrete ARD sources such as Ruby Gulch, the wet well, Hoodoo Gulch, and Pond C. These


256

inflows are monitored based on flow-rate measurements and/or measurements of the total volume of ARD pumped from the collection facility. However, dispersed inflows such as direct precipita-tion on pit lakes and mine highwalls are not currently measured.

3.2 Site water balance

The site water balance is evaluated based on a mass continuity approach. This method is a standard hydrological approach, which can be used to evaluate systems ranging from very simple to very complex (Younger et al., 2002). The approach is summarized as follows:

inflow outflow storage∑ ∑− = Δ (1)

Inflows include precipitation and groundwater inflows. Outflows include pit lake evapora-tion, evapotranspiration, water treatment plant discharges, and discharges from surface water to groundwater. Accurately estimating pit lake evaporation and evapotranspiration is difficult at the site. Therefore, a method has been developed to facilitate estimation of net inflows of ARD to the site without relying on estimates of pit lake evaporation and evapotranspiration. This method, based on net inflow, is described as follows:

net inflow inflow E ETpl= − − (2)

where Epl = Water lost to pit lake evaporation; and ET = Water lost to evapotranspiration.

The sum of the inflows can be expressed as:

inflow net inflow E ETpl= + +∑ ∑∑∑ (3)

The sum of the outflows can be expressed as:

outflow WTP E ETout pl∑ ∑∑ ∑= + + (4)

where WTP = Water treated and discharged by water treatment plant.By substituting the expanded expressions for inflows and outflows into Equation 1, the terms

related to pit lake evaporation and evapotranspiration cancel out.

net inflow E ET WTP E ET storagepl out pl∑ ∑ ∑∑∑∑+ + − + +( ) = Δ (5)

Based on these relationships, the mass balance expression shown in Equation 1 can be expressed as:

net inflow WTP storageout∑ ∑− = Δ (6)

In order to provide a meaningful value of net inflow, the summation is completed over the water year, which is defined as October 1 to September 30. This period is evaluated because it facilitates accounting for precipitation that falls as snow in the fall and subsequently melts in the spring during the same water year. It also provides for annual accounting of pit lake evaporation and evapotranspiration, which attenuates short-term variations in pit lake evaporation and eva-potranspiration rates.

Equation 6 can be rearranged as follows:

net inflow storage WTPout∑ ∑= +Δ (7)


257

Based on Equation 7, the sum of net inflows to the site can be estimated by adding the change in water storage to the sum of the water treated and discharged by the water treatment plant. Estimated net inflows for the water year based on the relationship shown in Equation 7 are presented in Table 1.

As shown in Table 1, the total volume of ARD stored on the site at the end of the 2007 water year was approximately 635 million liters, and net inflows during 2007 were approximately 424 mil-lion liters. Estimated net inflows have ranged from a high of approximately 444 million liters in the 2006 water year to a low of 158 million liters in the 2004 water year. The estimated net inflow shown in Table 1 is defined as the “annual ARD yield.”

The annual ARD yield is strongly influenced by the amount of precipitation received at the site. In order to evaluate relative changes from year to year, the “normalized ARD yield” for the site is defined as the ratio of annual ARD yield to total annual precipitation. Normalized ARD yield is expressed in units of million liters per centimeter precipitation (million liters/cm). The normalized ARD yield for the Gilt Edge site is shown in Table 2 for water years 2000–2007.

Several uncertainties are present in the normalized ARD yield. The primary uncertainties are the distance of the weather station from the site (about 10 km) and variations in the timing of precipita-tion. Site precipitation during the summer months is dominantly in the form of large thunderstorms. Variations in precipitation are common over relatively short distances during summer thunderstorms in the Black Hills (Driscoll et al., 2000). Variations in the timing of precipitation also affect the measured ARD yield. The ARD runoff produced for a given precipitation event (for example, 1 cm)

Table 1. Estimated net ARD inflow for water years 2000 to 2007 (millions of liters).

Water yearTotal stored volumeat end of water year Change in storage

Water treatment plant discharge Estimated net inflow

2007 635 −14.6 439 424

2006 650 74.1 370 444

2005 576 142 63.8 206

2004 434 −135 293 158

2003 569 178 2.65 181

2002 391 19.8 153 173

2001 370 −147 457 310

2000 518 −32.2 400 368

1999 550

Table 2. Normalized ARD yield for water years 2000 to 2007.

Water YearEstimated inflow(million liters)

Precipitation at lead,SD (cm)

Normalized ARD yield (million liters/cm)

2007 424 78.4 5.4

2006 444 95.4 4.7

2005 206 69.0 3.0

2004 158 57.3 2.8

2003 181 65.2 2.8

2002 173 60.2 2.9

2001 310 70.7 4.4

2000 368 72.9 5.0


258

varies in relation to the soil moisture conditions. Precipitation that occurs during the spring when soil is moist causes more ARD per centimeter than precipitation that occurs in very dry months such as August and September. Snowfall also affects seasonal variations in ARD runoff because evapora-tion, infiltration and sublimation reduce the snowpack during winter months. The resulting spring runoff represents only a portion of the water that fell as snow in previous weeks or months.

The normalized ARD yield varies in response to changes in site conditions and ARD manage-ment practices. Site conditions and ARD management practices remained relatively stable over the 2006 and 2007 water years. This is an advantage in terms of understanding the normalized ARD yield, because it allows averaging the normalized ARD yield for the 2006 and 2007 water years. The best estimate of the current normalized ARD yield is 5.1 million liters/cm, which is the aver-age of the 2006 and 2007 normalized ARD yields. The higher normalized ARD yield observed during the 2007 water year is thought to be a result of uncertainties in the estimate caused by the distance of the precipitation gauge from the site and variations in the timing of precipitation events throughout the year. The range in values for the 2006 and 2007 water years displays the magnitude of uncertainty in the estimates.

4 MONTE CARLO SIMULATION OF FUTURE WATER BALANCE

The normalized ARD yield can be used to estimate the volume of stored ARD that may be present on the site in the future based on current ARD management practices. This information is useful to assess water treatment plant performance, site operations and maintenance practices, potential risks of ARD release, and ARD management actions that may be required in order to conduct future remedial actions such as pit backfilling. The maximum capacity for ARD storage at the site is approximately 961 million liters, and it is necessary to maintain the volume of stored ARD below this level in order to prevent an uncontrolled release.

This analysis utilizes a spreadsheet-based model and Monte Carlo simulation to evaluate the volume of ARD that may require storage in future years. The spreadsheet-based model relates expected precipitation, the current estimate of normalized ARD yield, expected water treatment rates, and expected water treatment plant operational efficiency. The specific value of each of these variables in the future is unknown, although a general range of expected values can be esti-mated. The Monte Carlo simulation allows definition of a probability distribution for each vari-able based on the estimated range of future values. Random values for each variable are selected from the corresponding probability distributions and the spreadsheet model calculations are com-pleted. This process is repeated thousands of times to develop a probability distribution describing future ARD storage. This distribution incorporates the range of uncertainty in the input variables and propagates this uncertainty through the estimation.

Precipitation data used in the estimate are based on historical precipitation at Lead, South Dakota from the years 1948 to 2005. These data were compiled into 1-year totals, 3-year totals, 5-year totals and 7-year totals to provide data regarding expected precipitation over these time intervals. Table 3 shows summary statistics based on this analysis.

Table 3. Summary Statistics of 1-, 3-, 5- and 7-year precipitation totals at Lead, South Dakota (centimeters).

1-year 3-year 5-year 7

Mean 73.0 219.2 367.4 517.6

Standard deviation 16.1 34.4 46.3 53.6

Maximum 108.6 294.4 473.6 650.3

Minimum 44.4 158.8 290.3 414.3

Coefficient of variability 22.06% 15.70% 12.61% 10.35%

Skewness 0.57 0.56 0.67 0.58


259

The summary statistics show several pertinent features of the precipitation data set. The coef-ficient of variability (the standard deviation divided by the mean) allows relative comparison of the variability of the populations for the 1-year, 3-year, 5-year and 7-year totals. The coefficient of variation is highest for the 1-year total and it decreases successively for the 3-year, 5-year and 7-year totals. This shows that precipitation totals summed over longer periods are less variable and that precipitation summed over longer periods more closely approaches average conditions. The positive skewness indicates that the populations are not normal (bell-shaped) distributions; rather, they are positively skewed with a larger number of values lower than the mean and less frequent, very high values. A logarithmic distribution is an example of a positively skewed distribution. In the Monte Carlo simulation, the precipitation distribution is represented by a probability distribu-tion called a gamma distribution. This type of distribution is used in the simulation to estimate future precipitation at the site for 1-year, 3-year, 5-year and 7-year periods. A comparison of the modeled gamma distribution to actual annual precipitation is shown as Figure 1.

As discussed previously, the best current estimate of the normalized ARD yield is the mean of the 2006 and 2007 water year estimates. However, there are uncertainties in this estimate related to the distance of the precipitation gauge from the site, orographic variations in precipitation, and the timing of precipitation during each water year. Therefore, a probability distribution is developed to estimate the range of these uncertainties in the Monte Carlo simulation. This normalized ARD yield is assumed to be represented by a normal distribution with a mean of 5.1 million liters/cm and a standard deviation derived from the two estimates. Although this is a limited dataset to define the probability distribution of the normalized ARD yield, additional data are not available and the assumption results in consideration of a range of values that recognizes the uncertain-ties in the estimate. The probability distribution used in the Monte Carlo simulation is shown in Figure 2.

The average water treatment plant discharge rate increased during the 2007 water 12.6 liters/second (l/s) year from approximately 12.6 liters/second (l/s) to approximately 17.9 l/s. For the

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 20 40 60 80 100 120 140 160 180

1 Year Precipitation 1948-2005 (cm) 3 Parameter Gamma Distribution

Figure 1. Comparison of the modeled gamma distribution to actual annual precipitation at Lead, South Dakota 1948 to 2005.


260

purpose of forward-looking water balance estimates in the Monte Carlo simulation, treatment rates ranging from 15.8 to 18.9 l/s have been used. This range is represented by a triangular prob-ability distribution with a most likely value of 18.0 l/s. In addition, the simulation considers plant operational efficiencies ranging from 85-percent to 98-percent. This range is represented by a triangular probability distribution with a most likely value of 95-percent. This assumption is sup-ported by a measured operational efficiency of 95.4-percent for the period of January 1, 2007 to October 31, 2007. These probability distributions are shown in Figures 3 and 4.

The Monte Carlo simulation evaluated the spreadsheet model 10,000 times using random val-ues generated from each of the defined probability distributions. The results of this exercise are summarized in cumulative probability diagrams in Figures 5 through 8. These diagrams show the likelihood that stored ARD volumes will be below various levels at the end of the 2008, 2010, 2012, and 2014 water years. This estimate takes into account the potential variation in future pre-cipitation, uncertainties in the normalized ARD yield, and potential variations in water treatment rates and efficiencies described in the proceeding paragraphs.

0

500

1000

1500

2000

2500

2.9E+06

3.1E+06

3.3E+06

3.5E+06

3.7E+06

3.9E+06

4.1E+06

4.3E+06

4.5E+06

4.7E+06

4.9E+06

5.1E+06

5.3E+06

5.5E+06

5.7E+06

5.9E+06

6.1E+06

6.3E+06

6.5E+06

6.7E+06

6.9E+06

Bin (liters per centimeter)

Nu

mb

er

of

Sim

ula

tio

ns

Figure 2. Histogram showing simulated distribution used to represent the normalized ARD yield.

0

200

400

600

800

1000

1200

1400

15.2

015

.4015

.6015

.8016

.0016

.2016

.4016

.6016

.8017

.0017

.2017

.4017

.6017

.8018

.0018

.2018

.4018

.6018

.8019

.0019

.20

Bin (liters per second)

Nu

mb

er

of

Sim

ula

tio

ns

0

200

400

600

800

1000

1200

0.81

0.82

0.83

0.84

0.85

0.86

0.87

0.87

0.88

0.89

0.90

0.91

0.92

0.93

0.94

0.95

0.96

0.96

0.97

0.98

0.99

Bin (percent)

Nu

mb

er

of

Sim

ula

tio

ns

Figures 3 & 4. Histograms showing simulated distributions used to represent water treatment rate and water treatment plant operational efficeincy.


261

Figure 5. Cumulative probability diagram showing probability that stored ARD volume will be less than the volume shown at end of 2008 water year.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0.00E+00 5.00E+08 1.00E+09 1.50E+09 2.00E+09

Volume of Stored ARD (liters)

Cu

mu

lati

ve

Pro

ba

bil

ity


0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0.00E+00 2.00E+08 4.00E+08 6.00E+08 8.00E+08 1.00E+09 1.20E+09


Cu

mu

lati

ve P

rob

ab

ilit

y


262

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0.00E+00 5.00E+08 1.00E+09 1.50E+09 2.00E+09


Cu

mu

lati

ve P

rob

ab

ilit

y

Figure 7. Cumulative probability diagrams showing probability that stored ARD volume will be less than the volume shown at end of 2012 water year.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0.00E+00 5.00E+08 1.00E+09 1.50E+09 2.00E+09


Cu

mu

lati

ve

Pro

ba

bil

ity



263

The Monte Carlo simulation also provides information regarding the sensitivity of the model to the various assumptions for future precipitation, water treatment plant performance and normal-ized ARD yield. A sensitivity analysis based on a regression method is shown in Table 4.

The model is very sensitive to assumptions regarding future precipitation, as would be expected. Fortunately, an extensive precipitation record is available for the Lead, South Dakota weather sta-tion (1948 to present) and a rigorous statistical model incorporates these data into the Monte Carlo simulation. In later years, the model becomes relatively less sensitive to precipitation and relatively more sensitive to uncertainty in the normalized ARD yield. These changes in sensitivity are related to decreased variability in precipitation data over multi-year periods as described previ-ously based on the coefficient of variation. The model is less sensitive to assumptions regarding water treatment rate and water treatment plant operational efficiency. These parameters are nega-tively correlated with the estimated site water balance because increases in either water treatment rate or operational efficiency lead to decreases in the site water balance.

4.1 Potential for uncontrolled ARD release

The predicted volume of stored ARD developed through Monte Carlo simulation can be used to estimate the probability that the ARD storage capacity will be exceeded in future years, which could lead to an uncontrolled ARD release. The maximum ARD storage capacity at the site is 961 million liters. Based on the Monte Carlo simulation, the probability that the total volume of stored ARD will be less than 961 million liters at the end of the 2008, 2010, 2012, 2014 water years is 99-percent. These estimates indicate that it is unlikely that an uncontrolled release of ARD would occur as a result of insufficient ARD storage and treatment capacity. It should be recognized that the probabilities are based on estimated volumes present at the end of the water year (September 30), and that the volume of stored ARD on the site varies throughout the year. The highest volume of stored ARD on the site is generally present in the spring as a result of snow melt and heavy precipitation in the months of April, May and June. Therefore, the probability that the total volume of stored ARD on the site would be less than 961 million liters in the spring would be somewhat lower than the probability analyses indicate, because it considers stored ARD volumes at the end of the water year (i.e., September 30).

5 CONCLUSIONS

The application of stochastic methods to water balance evaluations provides an important tool to reduce the risk of uncontrolled releases of mine water into the environment. This process facili-tates improved site management decisions, because uncertainties in numerous site variables are propagated through the water balance evaluation and the results are presented in terms of the prob-ability that a future event may or may not occur. Site management decisions can then be based on a risk level that is deemed acceptable based on site-specific factors such as the sensitivity of the

Table 4. Sensitivity analysis of assumptions required for the estimate.

Source of uncertainty in estimate

Relative sensitivity of simulation to various sources of uncertainty

2008 2010 2012 2014

Uncertainty regarding future precipitation 88.9% 81.3% 75.0% 67.9%

Uncertainty regarding actual ARD yield 32.1% 42.1% 49.3% 53.8%

Uncertainty regarding WTP operational efficiency

−20.6% −25.2% −29.1% −32.9%

Uncertainty regarding water treatment rate −24.6% −31.8% −33.7% −38.3%


264

environment to a potential release, chemical hazards present in mine water, and costs of achieving lower risk levels.

REFERENCES

Driscoll, D.G., Hamade, G.R. & Kenner, S.J. 2000, Summary of precipitation data compiled for the Black Hills Hydrology Study, Western South Dakota: U.S. Geological Survey Open-File Report 00-329, 151 p.

Younger, P.L., Banwart, S.A. & Hedin, R.S. 2002, Mine Water, Hydrology, Pollution, and Remediation, Klu-wer Academic Publishers, Dordrecht, The Netherlands, pp. 218–226.


265


Using water balance tools for site design, operation and expansion management

A. TrautweinKnight Piésold and Co., Denver, CO, USA

ABSTRACT: As gold production continues to be highly profitable, businesses must make critical decisions how to manage their sites. The recovery and safe storage of water and process solutions is often at the forefront of their concerns. Even though many businesses rely on experi-ence and operating plans to manage daily water supply, there are few tools and solutions available to perform realistic forecasting and simulate new operating scenarios for the near and long term. Modifying spreadsheets and hiring consultants to evaluate numerous combinations is becoming less practical. Meanwhile, a new class of software is emerging. With simple user interfaces and sophisticated, “behind-the-scenes” precipitation models, heap leach models and tailings storage facility models, users can quickly forecast conditions based on design and actual information. Even missing information like actual precipitation and pond depths are being filled “intelligently” through advanced programming. Users needed considerable programming knowledge, training and time to perform these tasks not long ago. Now, users can easily and intuitively modify input information and understand the output in minutes.

With increased interest in producing more gold, imagine being an operator who can simply demonstrate to managers why their production targets compromise safety. Similarly, imagine being a manager who can independently use the same tools to investigate long-term possibilities and avoid safety and environmental hazards. There are many opportunities to increase perform-ance in today’s market. Fortunately, the technology is now available to make better operating and business decisions.

1 INTRODUCTION

A water balance provides a detailed analysis of site conditions using climatological and opera-tional information. It calculates inflows and outflows to measure changes in pond volume. The results are often used to establish pond storage, water treatment and makeup requirements, as well as, appropriate pipeline sizing.

Many water balances have been prepared using spreadsheet applications. Spreadsheets, how-ever, have many limitations. Depending on the complexity of the system and life of the mine, it is not unusual to exceed the number of fields and records supported in a single spreadsheet and the formula length supported in a single cell. While some migrate toward databases to overcome these limitations, there are other useful solutions worth their consideration.

2 STATE OF PRACTICE (CONVENTIONAL TAILINGS TECHNOLOGY)

One of the most important water balance tool requirements is the ability to manage a large amount of information. Since most water balances rely on numerous daily and monthly climatological and operational information, methods to enter, import and edit information need to be easy.

In addition to easily managing information, it must be easy to “run” the water balance. Once the analysis is complete, the output needs to understandable at-a-glance, too.


266

3 PROCESS SCHEMATIC

Older water balance tools often required users to understand programming languages to modify the process circuit. Today, however, users welcome the opportunity to click on a process schematic to change the process logic.

4 COMPLEXITY

As many may agree from their experience, water balances that seem too technically and visually complicated are seldom used. A solution offering a simple user interface and realistic models requiring minimal input information is the “holy grail” of the water balance field.

5 CLIMATOLOGICAL INFORMATION

Many sites lack a long-term climatological record. Since climatological information is an essential part of a water balance solution, estimates for 24-hour storms, daily precipitation, monthly pre-cipitation, maximum air temperature and minimum air temperature need to be developed. While water balance tools use a number of different methods to develop site-specific climatological information, Knight Piésold (WaterMinder™) uses the National Weather Services (NWS) Pre-cipitation-Frequency Atlas of the United States (NOAA Atlas 14, 2004), Oregon State University’s PRISM program (2004) and the United States Department of Agriculture Erosion/Productivity Impact Calculator (EPIC), as necessary.

6 FUNCTIONALITY

While some use water balances solely for design purposes, there is an exciting opportunity for operators to use water balance tools on a daily basis to forecast conditions months into the future using normal and extreme precipitation. By updating actual daily precipitation and pond depths, operators can more accurately predict critical periods due to climatological and operational com-binations with time to take corrective measures.

If a long climatological record is available or can be created, water balance tools are also use-ful for analyzing hundreds of combinations based on historical information to produce long-term forecasts through life of mine. Managers can use information (updated by operators) to make key business decisions with minimal effort and training.

7 PERFORMANCE

Less than 20 years ago, it was not unusual for a computer to take hours to perform a water bal-ance analysis and many more to prepare the output. Now with some of the available water balance solutions being described and modern computer systems, a short-term analysis takes seconds and a long-term analysis (with a long climatological record) takes a couple of minutes. Finally, it is practical to use water balance tools for daily forecasting.

8 OUTPUT

Output from water balance tools is typically presented in detailed tables and graphs. Newer solutions, however, present a summarized “dashboard” with the ability to access detailed


267

information on-demand. Color coding and other visual cues help users understand the output at-a-glance.

9 PERFORMANCE

Less than 20 years ago, it was not unusual for a computer to take hours to perform a water bal-ance analysis and many more to prepare the output. Now with some of the available water balance solutions being described and modern computer systems, a short-term analysis takes seconds and a long-term analysis (with a long climatological record) takes a couple of minutes. Finally, it is practical to use water balance tools for daily forecasting.

10 MAINTENANCE

While the precipitation, heap leach and tailings storage facility models seldom change, the proc-ess schematic and process logic occasionally do. Users must either make their own modifications using the available programming framework or (more commonly) pay for expert help. Fortu-nately, this usually only happens during site expansion. Don’t forget about deployment issues, though. While stand-alone and web-based solutions offer similar results, stand-alone solutions often require more troubleshooting on individual workstations and more on-site support.

11 VALUE

In the mining industry, avoiding environmental and safety hazards is a primary concern. Any solu-tion with the capabilities described above will help. Considering the consequences of a disaster, a water balance forecasting system always seems to be a good investment.

12 EXTENSIBILITY

New water balance solutions don’t end with avoiding environmental and safety hazards. Look for functionality that:

• allows the user to evaluate various rinse scenarios to determine impacts on the process circuit.• allows the user to evaluate various scenarios to determine the best way to close a facility.• allows the user to assign recovery curves to each heap lift.• allows the user to optimize the leaching schedule with regards to gold recovery.• allows the user to evaluate different expansion scenarios to determine from a gold recovery

perspective what option is preferred.• allows the user to determine the gold inventory of the heap leach pad.In the mining industry,

avoiding environmental and safety hazards is a primary concern. Any solution with the capa-bilities described above will help. Considering the consequences of a disaster, a water balance forecasting system always seems to be a good investment.

13 ABOUT THE AUTHOR

Adam Trautwein is a computer programmer with Knight Piésold and Co. He has worked alongside water balance experts to develop water balance spreadsheets, stand-alone water balance solutions and web-based water balance solutions (WaterMinder™) since the early 1990’s. Usability, devel-oping sophisticated computer models and adding value are his main interests.


268

REFERENCES

Brandyk, T. & Wessling, J.G. 1987. “Soil Moisture Flow in Drainage-Subirrigation System,” Journal of Irriga-tion and Drainage Engineering, ASCE, Vol. 113, No. 1, February.

Daly, C., Gibson, W., Doggett, M., Smith, J. & Taylor, G. 2004. Up-To-Date Monthly Climate Maps for the Conterminous United States, Oregon State University, Corvallis, Oregon 14th AMS Conference on Applied Climatology, American Meteorological Society, Seattle, Washington, January 13–16, 2004.

Frost, K.R. & Schwalen, H.C. 1955. Sprinkler evaporation losses, Agricultural Engineering 36: 526–528.Gumbel, E.J. 1954. A Statistical Theory of Extreme Values and Some Practical Applications, National

Bureau of Standards Applied Mathematics Series 33, U.S. Government Printing Office, Washington, D.C., February 12, pp. 51

Gumbel, E.J. 1953. Probability Tables for the Analysis of Extreme-Value Data, National Bureau of Standards Applied Mathematics Series 22, U.S. Government Printing Office, Washington, D.C., July 6, pp. 32

Hann, C.T., 1982, Statistical Methods in Hydrology, The Iowa State University Press.Hargreaves, G.H. & Samani, Z.A. 1982. “Estimating Potential Evapotranspiration,” Technical Note, Journal

of Irrigation and Drainage Engineering, ASCE, Vol. 108, No. 3, pp. 225–239.Kite, G.W. 1988. Frequency and Risk Analysis in Hydrology, Fort Collins, Water Resources Publications.National Weather Service, 2004, NOAA, Precipitation-Frequency Atlas of the United States, NOAA Atlas 14,

Vol. 1, Ver. 3. Nevada, G.M. Bonnin, D. Todd, B. Lin, T. Parzybok, M. Yetka, and D. Riley.Richardson, C.W., 1981. “Stochastic Simulation of Daily Precipitation, Temperature, and Solar Radiation,”

Water Resources Research, Vol. 17, No. 1.Ritchie, J.T., 1972. Model for predicting evaporation from a row crop with incomplete cover, Water Resour.

Res. 8: 1204–1213.United States Department of Agriculture, 1990. EPIC-Erosion/Productivity Impact Calculator.Viessman, W. Jr, Knapp, J.W., Lewis, G.L., Harbaugh, T.E. 1977. Introduction to Hydrology, Harper & Row,

Publisher.World Meteorological Organization (WMO). 1973. A Manual for Estimation of Probable Maximum Precipi-

tation, Operational Hydrology Report No. 1, WMO—No. 332, Geneva, Switzerland, pp. 190


269

Desiccation and Rheology in cyclic surface deposition of gold paste tailings

P. Simms, B. Fisseha, J. Henriquez & R. BryanDepartment of Civil and Environmental Engineering, Carleton University, Ottawa, Ontario, Canada

ABSTRACT: The deposition of paste tailings on surface certainly offers many potential advan-tages to conventional deposition, such as reduced reliance of dams, increased water recycling within an operation, reduction in volume of tailings, and lower rates of seepage. The disadvantages include increased risk of acid generation, the challenges of managing the evolving deposition geometry, and cost. This paper presents research on both deposition rheology and evaporation rates of gold paste tailings in a cyclic deposition scheme. Modeling of desiccation of multilayer tailings deposits is presented. Modeling of non-Newtonian flow of successive deposits of tailings in a flume is also presented. In light of the results, the authors discuss a framework whereby tail-ings mangers can reduce dewatering costs by relying on self-weight consolidation and evaporation to maximize stack deposition angles.

1 INTRODUCTION

Disposal of thickened or paste tailings is an attractive option for surface disposal, as it eliminates or reduces some of the risks associated with slurried tailings deposition, most importantly reduc-ing reliance on dams and the associated risk of catastrophic failure associated with conventional impoundments (ICOLD, 2001). Thickened or paste tailings offers other comparative advantages, such as increased water recycling within the mining operation, and reduced groundwater seep-age out of the tailings impoundment. Recent advances in technology have permitted economic dewatering and pumping of thicker tailings to the point where they behave as a non-Newtonian fluid, and form gently sloping stacks during deposition, thus reducing reliance on confinement by dams. Tailings thickened to the extent that no segregation of particle size occurs during transport and relatively small amount of settling occurs post-deposition are often called “paste” (Cincilla et al. 1997). Paste deposition has been employed at a full-scale at the Bulyanhulu Gold Mine in Tanzania (Simms et al. 2007, Suttleworth et al. 2005, Theriault et al. 2003). It has been shown at Bulyanhulu, that if the deposition is cycled between a number of points in the impoundment and the geometry of the flow is properly controlled (Shuttleworth et al. 2005), the tailings will densify and gain significant strength through desiccation, allowing the development of stable stacks with up to a 6 degree slope.

Evaporation is an important parameter to the deposition of thickened or paste tailings. On one hand, evaporation is desirable as it promotes desiccation and strength gain, but on the other hand excess evaporation can lead to desaturation of the tailings and the consequent ingress of oxygen. If the tailings are sufficiently sulphidic, this can lead to acid generation. Therefore, it is desir-able to manage the deposition so as to target an optimal rate of drying. This optimal point can be illustrated by comparisons between void ratio, water content, degree of saturation, and associated increases in strength and oxygen diffusivity, as shown in Figure 1.

It is desirable to obtain a water content in the tailings that gives a sufficiently high degree of densification and strength gain, but still high enough to minimize the risk of acid generation. An optimal range of water content is possible, as the buffering capacity of the tailings can be to some extent relied upon to maintain a high pH even if some oxidation occurs.



270

Is it possible to predict the evolution of the tailings along the shrinkage curve? We will present in the body of the paper results from laboratory simulation of multilayer deposition and modeling using the unsaturated flow code SoilCover, showing the capacity of such kind of models to predict the behavior of desiccating paste tailings.

Clearly, the desiccation of a fresh layer of tailings very much depends on its thickness. There-fore, another important challenge is to be able to control the thickness of paste tailings during deposition, and indeed to control the overall geometry of the impoundment. The experience at Bulyanhulu has been that at the beginning of deposition, the tailings formed conical deposits stretched out in the direction of the slope of the underlying topography. However, when the various deposits began to overlap, the flow of the tailings became irregular and difficult to predict. This forced the mine operators to come up with innovative measures to retain control of the geometry on the run (Shuttleworth et al. 2005). Though the efforts of the Bulyanhuhu operators were suc-cessful in containing the flows within the desired footprint, clearly such a reactive approach to deposition is not ideal. For a more proactive approach, and improved understanding of the over-land flow behavior of paste tailings is desired. A number of researchers included the authors (Fitton et al. 2008, Pirouz & Williams 2007, Henriquez & Simms 2008) are undertaking studies to better understand the evolution of paste or thickened tailings geometry. Some results from the author’s group are presented in this paper.

2 MATERIALS AND METHODS

2.1 Materials

The materials are two gold tailings from two different mines. The geotechnical properties of the tailings are summarized in Table 1.

Figure 1. Shrinkage curve of paste tailings and associated geotechnical and geoenvironmental parameters (modified from Simms et al. 2007).

0.7

0.75

0.8

0.85

0.9

0.95

1

5 10 15 20 25 30 35 40Gravimetric water content (%)

0

0.2

0.4

0.6

0.8

1

Rela

tive s

tren

gth

or

rela

tiv

e o

xy

ge

n

dif

fusio

n c

oe

ffic

ien

t

Relative strength Relative oxygen diffusion

Void ratio

Point of desaturation


271

Both tailings, while shipped at the pumping water content (∼40 % mass of water / mass of solids, hereafter designated GWC), undergo significant consolidation during transport, arriving at the laboratory with a significant water cover, and the underlying tailings consolidated to between 23–25% GWC. The pore-water of both tailings was analyzed, and found to be dominated by iron, sulphate, magnesium, and calcium ions. More details on the properties of the tailings can be found in Fisseha (2008).

2.2 Experimental methods

2.2.1 Drying experimentsDrying tests were performed on sequentially deposited layers of initial thicknesses of 10 cm each with a plan area of 1.7 by 1.7 m. The tests were conducted in a steel box mounted on load cells, allowing evaporation to be directly measured. Evaporation was controlled using fans placed around the experiment. Relative humidity and air temperature were monitored 0.5 m above the surface of the tailings. Tests were performed when the receptacles were only filled with water, to measure the potential evaporation (PE) as a function of relative humidity and air temperature, assuming a constant wind speed. The only radiation impacting the surface of the tailings was ambient lighting. It was found that when the fans were turned off, the amount of evaporation was very low, less than 0.1 mm/day in each case, compared to a PE of 7–8 mm/day when the fans were turned on. The relative low solar radiation negates the influence of changing albedo, due to drying and salt precipitation, which can have a significant impact on drying in paste tailings and tailings in general (Simms et al. 2007, Fujiyasu & Fahey 2000).

While drying, the tailings were monitored for volume change, crack propagation, matric suc-tion (using tensiometers and heat dissipation sensors), and surface gravimetric water content. In the large test, surface samples were also obtained to measure the electric conductivity of the pore-water, as a means to track ion transport within the tailings. Pore-water conductivity was used to calculate osmotic suction using the USDA equation (Fredlund & Rahardjo 1993).

2.2.2 Flow experimentsFlume tests were performed on successively placed layers. A high speed camera, Model IN250 from High Speed Imaging was used to record the tests. Tailings were initially deposited at the pumping water content, 40% GWC. For multilayer tests, each layer was left for about 24 hours to allow it to reach the post-bleed GWC of 30%, before the next layer was added. This water content was confirmed by sampling. Pore-pressures were monitored using a tensiometer to ensure no sig-nificant suction developed. At this water content the older layers remain saturated and no cracking develops. Flume tests were performed by pouring the tailings through a funnel at one end of the flume. The funnel had an opening diameter of 2.4 cm. Once prepared at the appropriate water content, the paste was placed in a bucket and mixed with a stirring machine for approximately 10 minutes prior to being deposited in the flume. The flume is made of see-through acrylic and is 30 cm tall, 15 cm wide, and 2.5 m long.

The parameters of interest are the yield stress, the viscosity-shear stress relationship (flow curve), and the bulk density. Yield stress and viscosity were obtained from flow curves measurements at

Table 1. Properties of tested materials.

Properties Tailings A Tailings B

Specific gravity 2.9 3.0

D10

, D50

, D60

(microns) 2, 35, 55 1, 100, 120

Cu (D60

/D10

) 27.5 120

Wl, W

p, W

sh (%) 20, 19, 20 20,18,18

k sat

(m/sec) 2 × 10−7 2 × 10−7


272

different water contents using a Brookfield RS rheometer using a cone and plate fixture. More detail can be found in Henriquez & Simms (2008).

2.3 Numerical simulations

2.3.1 Simulation of drying experiments using the SoilCover modelThe well-know SoilCover model was used to model the multilayer drying experiments. The model is employed essentially in the same way as described in Simms et al. (2007), with the difference that now multilayer deposition must be considered. Therefore, one would expect hysteresis of the water retention curve to play some role, but in the analysis hysteresis was initially ignored, with the hope that despite this assumption reasonable results would still be obtained. The water-retention curve was measured using a pressure plate, and the volume change was measured to accurately calculate the degree of saturation versus suction curve. This degree of saturation curve, not the raw water retention curve, is used to estimate the unsaturated hydraulic conductivity func-tion, as described in Simms et al. (2007).

2.3.2 Simulation of the flow experiments using lubrication theoryThe flume experiments were compared to transient and steady-state solutions obtained from lubri-cation theory. Lubrication theory simplifies the Navier-Stokes equations by the assumptions of relatively slow flow (viscous forces dominate over inertial forces) and small depth to lateral extent ratios. These assumptions permit steady-solutions for specific geometries to be obtained for non-Newtonian fluids, for example, for two dimensional flow over a flat surface:

h h

gx xy2

02

0

2− = −

τρ

( ) (1)

Where h is depth, x is the runout distance, h0 is the depth at a specific runout distance x

0, τ

y is the

yield stress, ρ is the density, and g is the acceleration due to gravity. Equations for depth integrated flow for modeling transient behavior can also be be derived (Yuhi & Mei 2004):

∂∂

= − − −h

t

g dh

dxh h H h Hy y

ρμ

1

63 2 2( )( ) (2)

Where μ is the dynamic viscosity, h is the vertical height of flow, hy is the depth above which

plug flow occurs, and H is the bed elevation. Equation 2 is based on the assumption of laminar flow in sheared zones and plug flow in regions where the yield stress is not exceeded. Transient solutions presented in the paper were obtained using a finite volume solution to Equation 2, using a unit spacing of 0.02 m and a time step of 1/10000 of a second. More details can be found in Henriquez & Simms (2008). The tailings were analyzed as a Bingham fluid: in others words a constant viscosity was assumed in the numerical simulation.

3 RESULTS

3.1 Simulation of multilayer drying tests

An example of predicted and measured evaporation rates are shown in Figure 2 for successive deposition of two layers of tailings A, one layer deposited at 0 days, the other layer deposited at 10 days. The tailings were subsequently rewet on Day 15 to simulate the effect of rainfall.

It can be seen in Figure 2 that the predicted evaporation and the measured evaporation agree reasonably well until Day 15. This is despite the presence of significant cracking as evidenced in Figures 3 and 4. One can also see in these Figures the buildup of salts. This can be quantified by


273

0.00

2.00

4.00

6.00

8.00

10.00

12.00

0 5 10 15 20

Time (days)

Eva

pora

tion

(mm

/day

)

PE predicted from Penman eqn.AE predicted by modelMeasured AE

Figure 2. Predicted and measured evaporation in multilayer deposits of paste tailings.

the buildup of salts at the surface, and the consequent increase in Osmotic suction, which is shown in Figure 5. Note that only a limited number of data points could be obtained, as the tailings would eventually become too dry to extract sufficient water the conductivity test. It is believed that the buildup of salts at the surface is responsible for the disparity in the predicted and the measured evaporation rates after Day 15.

The SoilCover code was able to reasonably simulate the variation in matric suction that occurred when a fresh layer was placed on a desiccated layer. The most dramatic illustration of the action of the previously desiccated layer is shown in Figure 6, for a test on tailings B. It can be seen that when the second layer is added on Day 8, the suctions in the old layer drop quite quickly, while

Figure 3. Tailings A after Day 14 , after drying of the second layer for 5 days and just before rewetting.


274

0

100

200

300

400

500

600

700

800

0 5 10 15 20

Time (Days)

Osm

otic

suc

tion

(kP

a)

0

2

4

6

8

10

12

14

16

18

Con

duct

ivity

(m

illiS

)

Osmotic suction

Pore-waterconductivity

Figure 5. Electric conductivity sampled at the surface of tailings A, and calculated osmotic suctions.

the suctions in the fresh layer rise up, the suction in both layers converging, before further drying induces the suctions to rise.

Another important aspect of Figure 6 is the delay seen in the development of suctions, both after initial placement of the first layer, and again slightly after placement of the second layer. This is because both tailings undergo significant settling and/or self consolidation after deposition: in the absence of evaporation, both tailings will consolidate to 30% GWC, down from the deposition water content of 40% GWC. This was demonstrated for the Bulyanhulu tailings in Simms et al. (2007). This was modeled in SoilCover by choosing the m

v value (the parameter that gives the

amount of water released by a change in pore pressure when the pore pressure is positive) so as to simulate the release of the correct amount of water.

3.2 Visualization of flume tests

It was found that the predictions of equilibrium profiles fitted the results quite well. One of the interesting outcomes is that the overall angle of each deposit is a function of the total volume, and

Figure 4. Tailings A on Day 19, 3 days after rewetting.


275

-20

0

20

40

60

80

100

120

140

160

180

200

0 5 10 15 20

Time (Days)

Ma

tric

su

cti

on

(k

Pa)

Tensiometer 12 cm frombottomTensiometer 2 cm frombottomSoil Cover 2 cm from bottom

SoilCover 12 cm from thebottomSoilCover 18 cm from thebottom

Figure 6. Modeled and measured matric suctions during a multilayer drying test on tailings B.

this behavior could be replicated by the lubrication theory equations. This may partially explain the discrepancy noted between laboratory and field scale angles commonly observed in practice. Figure 7 shows an example of three different single layer flows, the only difference being the vol-ume of the deposit. Figure 8 shows an example of the predicted and measured equilibrium profile of a multilayer deposit. Figure 9 shows an example comparison between predicted and measured transient profiles. In general, it was found that the transient model somewhat over-predicted the rate of spreading of the tailings. This may be due to the assumption of Bingham flow and constant

Figure 7. Predicted and measured equilibrium flume profiles of single layer deposits using Equation 1.

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Run-out (m)

Dep

th (

m)

Flume 1

Flume 2

Flume 3

Predicted


276

0

0.02

0.04

0.06

0.08

0.1

0.12

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

Runout (m)

Ele

vati

on

(m

)

"Old" tailings

Measured profile of freshdeposit

Predicted

Figure 8. Example of a predicted equilibrium flume profile of a multilayer deposit (Henriquez & Simms 2008).

Figure 9. Predicted and measured transient profiles for the flow of a fresh layer over a consolidated layer.

0

0.02

0.04

0.06

0.08

0.1

00.10.20.30.40.50.60.70.8

Runout (m)

Ele

vati

on

(m

)

Layer 1

Predicted-at 3 sec

Measured-at 3 sec

Predicted-at 6 sec

Measured-at 6 sec

Layer 2

Predicted - Final Profile

viscosity used in the simulation. More details can be found in Henriquez & Simms (2008). Note all these tests were performed on tailings A.

4 DISCUSSION

The suppression of evaporation by salts has been noticed in the field as well (Simms et al. 2007), and is common to many tailings deposits in arid environments (Fujiyasu & Fahey 2000). In a field trial conducted at Bulyanhulu a 50 cm layer was left undisturbed for two months: It was that evaporation could be predicted using the SoilCover model and average weather data reasonably well for the first 20 days, but after this time evaporation virtually stopped. This is shown to be highly correlated with the buildup of a salt crust on the surface. The buildup of salts in the field is illustrated in Figure 10.


277

The formation of a salt crust may be beneficial or detrimental depending on how the deposition is managed. Formation of a salt crust at the surface too early may keep the tailings relatively soft and unstable; however, formation of salt crust after significant desiccation may prevent excess desaturation and serve to minimize the ingress of oxygen and therefore the generation of acid drainage. Also, before closure of the tailings, the contribution to runoff from the salts may affect water quality downstream. Clearly the role of salts is important, and the initial pore-water makeup of the tailings, ion transport within the tailings during deposition and precipitation at the surface should ideally be considered when planning deposition.

The use of lubrication theory based equations to model the evolving geometry of paste tailings was shown to be effective at the small scale. Tests at a larger scale are required to examine its applicability to the field. Many of the assumptions of lubrication theory may not hold: certainly, at some mines the velocity of the tailings as they exit the pipe may be quite fast and the flow may be turbulent, though the authors believe the flow becomes laminar at some distance from the pipe, and therefore the geometry will be largely controlled by laminar flow. Another assumption that may not hold true is the time independence of rheological properties—it is possible that some tailings may undergo significant settling and dewatering as they flow. This fact may turn out to be advantageous to mine operators, as it may be possible to rely on settling enhanced by desiccation through evaporation and the wicking action of the underlying layer, to increase the yield stress to that required for deposition. Though possible, this kind of deposition would be complex to control,

Figure 10. Evolution of tailings in the field over the summer of 2005 (adapted from Simms et al. 2007).

June 24th (3 days after deposition)

July 27th

July 18th


278

and further research is required to understand the rate of dewatering out of the pipe, and to antici-pate the behaviour of such dewatering tailings.

5 SUMMARY

Some results from research on desiccation and deposition geometry of multilayer deposition of paste tailings were presented. On the drying side, it was found that the evaporation rate could be reasonably predicted using the SoilCover unsaturated flow code, until the evaporation rate began to be suppressed by the buildup of salts on surface. Further studies will attempt to better character-ize the transport of the constituent ions to the surface and their precipitation as salts. In the studies described in the paper, cracks did not substantially affect the rate of evaporation.

The flume test visualization studies showed that lubrication theory based equations described the measured flows quite well. This theory may partially explain the discrepancy between overall deposition angles reported in the field and those in the laboratory, as it clearly shows that this angle is scale dependant. Testing at larger scales and with more complex flow geometries is still required to characterize the usefulness of this approach to planning field deposition.

ACKNOWLEDGEMENTS

Support for this research by grants from Barrick Gold, Golder Associates, Ontario Centre of Excel-lence, and NSERC is gratefully acknowledged, as is the assistance of Bulyanhulu site personnel.

REFERENCES

Cincilla, W.A., Landriault, D.A. & Verburg, R. 1997. Application of paste technology to surface disposal of mineral wastes. Proceedings of the Fourth International Conference on Tailings and Mine Waste ’97. pp. 343–356.

Fitton, T.G., Bhattacharya, S.N. & Chryss, A.G. 2008. Three-dimensional modeling of tailings beach shape. Computer-aided Civil and Infrastructure engineering, 23: 31–44.

Fisseha, B. 2008. Flow behaviour in multilayer deposits of unsaturated paste tailings. Ottawa: Carleton University.

Fredlund, D.G, & Rahardjo, H. 1993. Soil Mechanics for unsaturated soils. New York: Wiley.Fujiyasu, Y. & Fahey, M. 2000. Experimental study of evaporation from saline tailings. ASCE Journal of

Geotechnical and Geoenvironmental Engineering, 126: 18–27.Henriquez, J., & Simms, P. 2008. Dynamic imaging and modelling of multilayer deposition of gold paste tail-

ings. Minerals Engineering, doi:10.1016/j.mineng.2008.05.010Pirouz, B. & Williams, M.P.A. 2007. Prediction of Non-Segregating Thickened Tailings Beach Slopes—A

New Method. Proceedings of the tenth international seminar on paste and thickened tailings, Perth, Australia, March 13th–15th 2007, pp. 315–327.

Shuttleworth, J.A, Thomson, B.J. & Wates, J.A. 2005. Surface disposal at Bulyanhulu- practical lessons learned. Proceedings of the 6th International Conference on Paste and Thickened tailings, Santiago, Chile, April 20th–22nd.

Simms P., Grabinsky M.W. & Zhan, G. 2007. Modelling evaporation of paste tailings at the Bulyanhulu mine. Canadian Geotechnical Journal, 44(12): 1417–1432.

Theriault, J., Frostiak, J. & Welch, D. 2003. Surface disposal of paste tailings at the Bulyanhulu gold mine, Tanzania. Proceedings of Sudbury 2003, Mining and the Environment, Sudbury, Ontario, Canada, May 26th–28th 2003. Editors G. Spiers, P. Beckett, H. Conroy, pp. 265–269.

Yuhi, M. & Mei, C.C. 2004. Slow spreading of fluid mud over a conical surface. Journal of Fluid Mechanics, 519:337–358.


Remediation and reclamation



281

Past, present and future for treating selenium-impacted water

J. Gusek, K. Conroy & T. RutkowskiGolder Associates Inc., Lakewood, CO, USA

ABSTRACT: Water quality in selected regions of the US is impaired due to selenium, which typically originates from shale formations that were deposited in a marine environment. In many respects, the geochemical behavior of selenium is similar to sulfur; they both share a common group in the periodic table of elements. Early research work in sulfate reducing bioreactors revealed that selenium was reduced biologically in concert with sulfate. The paper discusses vari-ous past, present and future approaches for the treatment of dissolved selenium in either neutral runoff from agricultural lands or neutral to acidic Mining Influenced Water (MIW) from both coal and metal mining. Both active and passive treatment technologies are considered. Selenium bio-chemical reactor design principles are similar to those developed for sulfate reducing bioreactors commonly used for the treatment of high concentration metal-laden acidic MIW.

1 INTRODUCTION

The Kesterson National Wildlife Refuge in California was the focus of attention over 25 years ago when excessive selenium concentrations in the surface water were found to be the primary cause of mortality in fish and birds living there (National Research Council, 1989). The issue associ-ated with mining is not new either. Selenium uptake in plants grown on reclaimed mine lands was reported at the 1988 American Society of Mining and Reclamation (ASMR), US Bureau of Mines and Office of Surface Mining Joint Conference on Mine Drainage held in Pittsburgh. Not many years later, a full session was dedicated to selenium at the 1995 ASMR Conference in Gillette, WY. A scan of combined ASMR, International Conference on Acid Rock Drainage (ICARD) and the West Virginia Surface Mine Drainage Task Force proceedings since 1980 yielded over 50 papers that addressed selenium issues.

The uses and natural occurrences of selenium are varied; they include:

• Paint, pigments, dye formulating• Electronics• Glass manufacturing• Insecticide production• Pulp and paper• Ash piles, FGD blowdown• Coal/oil combustion• Agricultural water• Petroleum processing• Mining operations

2 WHAT’S THE PROBLEM?

The toxicity and health impacts of selenium in the natural and human environment is well estab-lished, although specific levels considered to represent a threat to human health or the environ-ment are not held in universal agreement. Its bioavailability and toxicity is a function of oxidation



282

state. With respect to human health, it is ironic that selenium is an essential trace nutrient found in health food stores, but that it also has a regulatory limit in drinking water. The range of toxic-ity and nutritional deficiency is narrow, however. Unfortunately, selenium is bioaccumulative; its concentration may increase in organisms at successively higher levels in the food chain. The main issues associated with this enigmatic element include:

• It’s an aquatic life hazard • 1983—Kesterson National Wildlife Refuge—California • Birth defects/death of birds, small animals, fish• It’s a regulated drinking water contaminant• Selenium cycle is not well understood • Uncertainty on bioavailability • Even if bio-available—what is a toxic threshold concentration?• Often times low-concentration/high-volume flows make treatment expensive

The primary goal of any treatment process is to meet applicable standards. This is a controver-sial and complex topic and beyond the scope of this paper. The diversity in numeric standards is reflected in the list below.

• Freshwater aquatic life—5 μg/L• Primary Drinking Water Standard (DWS) Maximum Contaminant Limit (MCL)—50 μg/L • U.S. Fish and Wildlife Service has recommended 2 μg/L to protect fish, waterfowl and endan-

gered aquatic species

Suffice it to say that any treatment process that reduces selenium loading to achieve or approach site-specific standards, whether concentrations are measured in the water itself or in fish tissue, is worthy of consideration. A secondary goal is to achieve dissolved selenium reductions as inex-pensively as possible.

3 BASIC SELENIUM GEOCHEMISTRY

The basic geochemistry of selenium needs to be briefly addressed before treatment alternatives can be considered. In contrast to sulfur, which is found in the same group in the periodic table, selenium has four primary oxidation states, including elemental selenium:

• elemental selenium (Se0)• −2 selenide (Se-2), analogous to sulfide (S−2)• +4 selenite (HSeO

3− and SeO

3−)

• +6 selenate (SeO4

–2) analogous to sulfate (SO4

−2)

The focus of this paper is mitigating mining impacts, so it is appropriate to focus on how selenium occurs in mining/geological environments. Ryser, et al. (2005) investigated the selenium-bearing minerals in the Western US Phosphate Resource Area in southeast Idaho and adjacent states (Utah and Wyoming) using micro-spectroscopic techniques. The researchers took great pains in sta-tistically analyzing samples of a selenium-bearing shale interburden sandwiched between two major phosphate ore zones in the Permian-age Phosphoria Formation. The ore itself is virtually selenium-free. The research included test calibrations of various known selenium-bearing mate-rial standards, including two forms of elemental selenium (monoclinic and orthorhombic crystal forms). In their background research, Ryser et al. found that the most commonly reported forms of selenium in shale are selenium-substituted pyrite or precipitated metal selenides. From the test-ing on the Phosphoria Formation shale interburden, they concluded that there were three distinct mineral phases that contain selenium:

• Dzharkenite (FeSe2)—an iron selenide analogous to pyrite

• An organic diselenide• Selenium-substituted pyrite.


283

Ryser et al. further observed that:

In all of these phases Se existed in the negative oxidation states, indicating that the oxidized Se present in weathered shales is a result of oxidative weathering of the reduced mineral Se to elemental Se, selenite, or selenate.

In other words, the iron selenides in the overburden are analogous to pyrite in generating selenium-MIW. The noted absence of elemental selenium in the un-mined/un-weathered shale has implications regarding passive treatment in biochemical reactors which is subsequently discussed.

4 PAST AND PRESENT SELENIUM TREATMENT—USEPA BDATS

The precipitation of selenium from the dissolved state in MIW does not appear to be driven by equilibrium principles; rather, processes are governed by oxidation-reduction (Redox) conditions, biological activity, and sorption (MSE, 2001). The following processes with their associated MIW constraining conditions have been embraced by the USEPA as Best Demonstrated Available Tech-nologies (BDAT) for selenium removal (MSE, 2001):

• Ferric coagulation/filtration • Typically for pH <7 • Relies on a co-precipitation effect • Effective removal requires reduction of selenate to selenite • Performance can be compromised if arsenic is present• Lime softening• Reverse osmosis • Non-preferential process—all dissolved species present are removed • Pretreatment due to other typical mine water issues (i.e., scaling, fouling) may be required• Electrodialysis• Alumina • Selenite is adsorbed at pH range of 3–8 • Silica can interfere at pH >4 • Selenate adsorption is poor• Ion exchange • Oxidized divalent selenate is needed • Competing ion effects can hinder effectiveness • Some specialty resins were tested• Ferrihydrite precipitation with concurrent adsorption of selenium on the ferrihydrite surface • For adsorption—ferricion (Fe+3) needs to be present • Most effective removal occurs at pH 4–6 • Somewhat effective up to pH 8 • Phosphate, silicate, arsenic, carbonate can interfere

From 1999 to 2001, the USEPA and the US Department of Energy (DOE) co-sponsored research conducted by the Mine Waste Technology Program (MSE, 2001). This effort focused on demonstrating several selenium treatment/removal methods. Three technologies were tested in the field:

• Ferrihydrite Adsorption (baseline)• Catalyzed Cementation• Biological Reduction

A fourth technology was tested at a bench scale, Enzymatic Reduction.The program objective was to practically treat to a selenium concentration of less than 50 μg/L

(the Primary Drinking Water Standard). The research was conducted on MIW from the Kennecott


284

Utah Copper Corporation’s Garfield Wetlands-Kessler Springs site, which had the following typical chemistry:

• <50 to >10,000 μg/L Se (up to 10 mg/L)• 95%+ selenate• TDS 1,000–5,000 mg/L

The following results were reported.Ferrihydrite—Ferrihydrite did not work on a consistent basis despite using various iron types,

concentrations, and Fe:Se ratios. The process could achieve the program objective but at prohibi-tive reagent consumption and with questions on the TCLP stability of the resulting sludge.

Catalyzed cementation—This was a process originally developed for arsenic, selenium, thal-lium removal; it removes metals by cementation on the surface of iron particles. It was believed to work on both selenite and selenate. Using proprietary catalysts, bench test work yielded favorable results but it did not work on a consistent basis during field trials.

Biological Reduction (BSeR™)—This process used anaerobic solids bed reactors in which sele-nium was believed to be reduced to elemental selenium by biofilms and proprietary microorgan-isms; molasses was used as carbon source. The process consistently met the program objective; over 70% of samples had concentrations less than the detection limit (2 μg/L). The comparative economics (Table 1) of the three processes suggested that biological methods hold great promise.

5 OTHER PROCESSES—NANOFILTRATION

Nanofiltration technology was evaluated by the US Geological Survey in 1996 for selenium removal from agricultural drainage. The test work showed that the technology achieved better selenate removal than selenite, which was not surprising since the process was designed for divalent rather than monovalent ions. Regardless, the system yielded 95+% removal at selenium concentrations of less than 1,000 μg/L; membrane scaling would be an issue in high sulfate MIW sources.

6 SUMMARY PROBLEMS WITH PAST & PRESENT PROCESSES

Collectively, the active treatment processes considered to date are typically non-selective for sele-nium; their implementation generates large amounts of secondary waste using a diverse array of reagents. While the processes are typically appropriate for the bulk removal of selenium in the presence of other metals, they have difficultly consistently producing clean effluents with less than 10 μg/L selenium. Of the processes considered, the biological processes appeared to provide the greatest promise in achieving effluent goals at a reasonable cost. A focused discussion of biological processes follows.

Table 1. Economic comparison of selenium removal technologies (from MSE, 2001).

BDAT Cementation BSeR™

Capital $1.0M $1.1M $0.6MO&M $2.1M $1.2M $0.14MNPV $17M $9.5M $1.1M$/cubic meter $3.67 $2.16 $0.35$/1,000 gal $13.90 $8.17 $1.32$/kg selenium $1,836 $1,079 $174

Based on 300 gpm plant, 2 mg/L selenium influent 2001 dollars; depreciation, leases, salvage and taxes were not considered For comparison, the 2005 price quoted for selenium was about $110/kg.


285

7 BIOLOGICAL REDUCTION—GENERAL

The first ASMR paper addressing biological treatment for selenium was presented in 1994 (Wildeman, et al.). Selenium was listed as one of the constituents of concern in a neutral MIW tailings underdrain solution and an acidic MIW from a waste rock dump at a gold mine in Nevada. Earlier work by Gerhart and Oswald (1990) utilized “microalgae-bacterial” treatment of agricultural water in California.

A system in the Panoche Drainage District in the San Joaquin Valley is indicative of the poten-tial success of biological systems (Fisher, 2004). Water chemistry in this locale ranges from 74 to 1,400 μg/L due to selenium rich soils; selenate is the primary species present. Numerous biore-mediation studies in this area examined Algal-Bacterial Selenium Removal (ASBR) and Anoxic ponds which reportedly reduced selenate to selenite to elemental selenium and settled the result-ing suspended precipitates. These systems typically exhibited removal efficiencies of about 80%.

Additional California-based research (Fisher, 2004) evaluated the behavior of constructed wetlands in which nine plant species were tested. The results showed 63% to 71% removal efficien-cies for water containing about 20 μg/L selenium. The effluent contained from 3 to 6 μg/L Se. While the technology worked, the hydraulic retention times (HRTs) were on the order of several days.

Thus, the passive treatment concept for selenium is not new. The process is attractive because microbes degrade/transform the contaminant for a number of reasons, including:

• Selenium is used as a bacterial energy source• Bacterial selenium removal is a detoxification mechanism• Selenium behaves or resembles another ion (sulfate)• Combinations of the above

Anaerobic biochemical reactors (BCRs) are especially attractive because of their ability to produce elemental selenium which is biologically unavailable at typical removal efficiencies of greater than 90%. There is some concern that nitrate and sulfate presence may interfere in these types of systems due to bacterial competition. It is interesting that while selenium speciation in bioreactors suggests the formation of elemental selenium, natural rock formations (at least in the context of phosphate deposits) do not appear to exhibit this species naturally. The implications of this observation are uncertain; perhaps the thermodynamics or kinetics of selenium sequestration in natural anaerobic environments favors organic complexation and co-precipitation with ubiqui-tous iron over the elemental form. This may be a subject suitable for future research.

This pioneering work resulted in more-selective selenium microbes being isolated; this was accompanied by advances in fixed film/biofilm media and a better understanding of operating conditions. The net result was that HRTs have been reduced from days to hours for active systems. Coincident advances were also made in passive technology.

The ABMet® Process is offered by the GE Water and Process Technologies group as a variant on the BSeR™ Process evaluated by MSE for the EPA/DOE. It has been used on several flue gas desulfurization projects at commercial scales with selenium levels ranging from 3,000 to 5,000 μg/L in the presence of up to 20,000 mg/L chloride. Removal efficiencies on the order of 98% to 99% removal are projected coincident with effluent concentrations as low as 10 μg/L (Sonstegard 2007).

8 HISTORICAL EXPERIENCE IN SELENIUM PASSIVE TREATMENT

8.1 Definition of passive treatment

As discussed above, there appear to be several promising technologies for biologically treating selenium-bearing MIW. To properly focus the discussion, the following definition of passive treat-ment was proposed by Gusek (2002):

Passive treatment is a process of sequentially removing metals and/or acidity in a natural-looking, man-made bio-system that capitalizes on ecological and geochemical reactions. The process requires no power and no chemicals after construction and lasts for decades with minimal human help.


286

It is a sequential process because no single treatment cell type works in every situation or with every MIW geochemistry. It is an ecological/geochemical process because most of the reactions (with the exception of limestone dissolution) that occur in passive treatment systems are bio-logically assisted. Lastly, it is a removal process because the system must involve the filtration or immobilization of the metal precipitates that are formed. Otherwise, they would be flushed out of the system, and the degree of water quality improvement would be compromised.

A truly passive system should also function for many years, without a major retrofit to replenish construction materials, and be able to function without using electrical power. Benning and Ott (1997) described a volunteer passive system outside of an abandoned lead-zinc mine in Ireland that has apparently been functioning unattended for over 120 years. Ideally, a passive treatment system should be designed to last for at least several decades.

8.2 Early passive treatment studies of selenium-bearing mining influenced water

Wildeman, Filipek, and Gusek (1994) presented the results of “proof-of-principle” studies of passive treatment of neutral/alkaline cyanide mill tailing and acidic MIW sources at an undis-closed gold mine in Nevada at the 1994 ICARD Conference in Pittsburgh. This study was prima-rily focused on the selection of the preferred removal methods (aerobic or anaerobic/BCR) for addressing the parameters of concern for each MIW source; both sources contained selenium. The tailings derived MIW and the acidic MIW from a waste rock dump contained about 200 μg/L and 1,000 μg/L total (unfiltered) selenium, respectively. The results of the test demonstrated in principle that selenium concentrations could be decreased to less than the 10μg/L detection limit with both aerobic and anaerobic geochemical environments.

The proof-of-principle laboratory scale tests were the basis of building two pilot-scale passive treatment units:

• An algae-filled, multi-celled free water surface pond system with cascades for the tailing MIW, and

• A down-flow biochemical reactor and aerobic polishing wetland (in series) which received the acidic MIW from a pre-existing intermediate holding pond.

Performance data from these two systems has not been previously published. Table 2 below summarizes their respective performance with respect to selenium removal.

8.2.1 Brewer Mine (1993–1995)The Brewer Mine is a closed gold mine in South Carolina with heap leach pad runoff and pit lake derived acidic MIW. The results of an 18-month study of a four-liter per minute capacity pilot scale biochemical reactor were described in Gusek, 2000. The primary parameters of concern were depressed pH, and elevated aluminum, iron, copper, and zinc. Several sampling events that included a longer list of analytes yielded data on of the pilot cell’s performance with respect to selenium removal. Typical MIW selenium concentration were about 1,500 μg/L; BCR effluent selenium concentrations in the final sampling event were about 50 μg/L, a removal efficiency of about 96.7%. The influent sulfate concentration in this final sampling event was 5,200 mg/L; the field pH was 2.0.

Table 2. Passive treatment results—Nevada Gold Mine tailing and waste rock MIW.

Flow rate Influent Effluent PercentPilot system (gpm) pH Se (μg/L) Se (μg/L) removal

Tailings MIW (Aerobic) 10 7.5 40 16 60%Waste rock MIW (BCR) 6 2.7 22 <5 ≥78%


287

8.3 Recent passive treatment studies

Pahler, Walker, Rutkowski, and Gusek (2007) described a bench scale BCR study conducted at a gravel pit adjacent to the Colorado River in western Grand Junction, Colorado. The selenium concentrations in a mine pit dewatering trench ranged from 31 to 93 μg/L during the year prior to their study (Kerr, 2006); during the study, the influent concentration was about 20 μg/L, but it did spike to 70 μg/L during the five-month test. Background concentrations of sulfate ranged between 1,000 to 2,000 mg/L.

Four bench BCRs were constructed using 208-liter (55 gallon) polyethylene drums in 2006. Each was filled with varying amounts of homogenously-mixed sawdust, hay, wood chips, agricul-tural limestone, zero valent iron (ZVI) powder, and manure (Table 3). In the hope of inoculating the BCRs with bacteria acclimated to selenium, the manure was derived from cattle grazing in pasture areas known to have in selenium-rich soils. ZVI was included in varying amounts in order to determine if chemical reduction of selenium could enhance the expected biological reduction of selenium. Reactor 1, lacking ZVI, served as a baseline. Reactors 2, 3, and 4 had increasingly larger weight percentages of ZVI.

Additional details on the conduct of the experiment are provided in the Pahler, et al. paper. The four reactors exhibited selenium removal efficiencies of up to 98%; effluent typically exhibited less than 2 μg/L selenium. In this situation, the ZVI incorporated cells offered no advantage over the baseline BCR which was ZVI-free.

The positive results of the bench scale BCR tests were the basis for designing a pilot scale BCR whose construction is planned for May, 2008. The pilot BCR will be sized to treat 4 L/min (one gpm).

Table 3. BCR media compositions.

Proportion of each component by weight

Component Reactor 1 Reactor 2 Reactor 3 Reactor 4

Sawdust 30% 30% 20% 2.5%Hay 10% 10% 10% 0%Wood chips 30% 30% 20% 2.5%Agricultural limestone 20% 5% 5% 5%Zero valent iron 0% 15% 35% 85%Cow manure 10% 10% 10% 5%

Figure 1. Brewer pilot BCR.


288

At full scale, the treatment cost associated with a BCR of similar design is estimated to be $0.07 per m3 of water treated ($0.26 per 1,000 gallons) or $115 per kilogram of selenium removed. At the 2005 selenium price quoted of $115 per kg, and assuming a nominal cost for recovering a portion of the selenium resource, the treatment economics for this technology appear to be more favorable than active treatment

In the author’s experience, the design of a selenium reducing BCR is similar to that required for a sulfate reducing BCR. Both are typically sized based on metal loading to a point as long as a minimum hydraulic detention time (HRT) is satisfied. The selenium BCR organic substrate may be enhanced with a sacrificial metal if the untreated MIW is lacking in metal (e.g., iron) concen-tration that would encourage the precipitation of iron selenide (see Ryser et al. 2005).

9 RECENT EXPERIENCE IN ACTIVE SELENIUM TREATMENT

A Western US metal mine waste rock leachate required treatment for reduction of TDS and selenium prior to discharge to surface water. The MIW at this site is comparatively dilute with respect to selenium; the influent concentration is typically about 30 μg/L while the effluent goal is 10 μg/L. The MIW has neutral pH with calcium, magnesium, and sulfate comprising the TDS with concentrations from approximately 5,000 mg/L to 8,000 mg/L. The other significant challenge of the project was the space constraints for the treatment system and the flow rate which ranged from 70 to 700 gpm with surge events in excess of 2,000 gpm. The treatment system, based on the results of seven months of pilot testing, includes a series of reverse osmosis (RO) units to provide initial volume reduction. The selenium reduction is achieved by biological treatment of the RO brine. Since the feed to the bio-treatment system is RO brine, the actual selenium concentration in the bioreactor feed is approximately 70 μg/L in a 16,000 mg/L TDS background. The bio-treated brine is then recombined with permeate from the RO to the TDS discharge limit and the combined stream is discharged to a nearby creek. Excess bio-treated brine is discharged to the sewer under a Pretreatment Discharge Permit. A process flow diagram of the system is provided in Figure 2.

The selenium bio-treatment process was developed on the bench-scale and then pilot scale including developing the culture which provided the selenium reduction in a high TDS environ-ment. The culture was developed from onsite sediments and sediment from the Great Salt Lake. The bio-treatment system, fueled by molasses supported by phosphate and urea nutrient amend-ments, is currently operating on a feed with approximately 16,000 mg/L TDS. Sulfide generation has occasionally been an issue in operation.

Figure 2. Process flow diagram active selenium bio-treatment at a western US metal mine.


289

As of early 2008, the full scale system (Figure 3) has been operating for about two years and compliant water at less than 10 μg/L is being produced. The footprint for the entire treatment facility is about one acre.

The treatment cost is estimated to be about $1.50 per m3 ($5.71 per thousand gallons) of MIW; this is slightly more than, but consistent with, the economics for the BSeRTM process cited in Table 1. However, because of the dilute MIW at this site (30 μg/L Se), the unit cost of selenium removal is esti-mated to be about $50,000 per kilogram. About 12 kg of selenium is removed annually at this site.

10 CONCLUSIONS

The treatment of selenium either actively or passively with removal efficiencies in the high ninety percent level has been demonstrated on a variety of MIW types, ranging from dilute, neutral pH to concentrated and acidic. The economics of passive treatment appear favorable if land area is avail-able at a reasonable cost. Resource recovery may be able to offset some of the costs of treatment but only if the original MIW has a reasonably elevated selenium concentration; the “unit mass” cost of treating selenium-dilute MIW is not conducive to realizing resource recovery “credits.”

REFERENCES

Beining, B.A. & Otte, M.L. 1997. Retention of Metals and Longevity of a Wetland Receiving Mine Leachate, in Proceedings of 1997 National Meeting of the American Society for Surface Mining and Reclamation, Austin, Texas, May 10–16.

Fisher, Lela. 2004. The Effect of Molasses Concentration on Bacterial Treatment of Selenium in Agriculture Waste Water in the San Joaquin Valley, U. California Berkley website: http://socrates.berkeley.edu/∼es196/projects/2004final/Fischer.pdf.

Figure 3. Full scale biological treatment system.


290

Gerhart & Oswald, W.J. 1990. Final Report Microalgae-Bacterial treatment for Selenium Removal for the San Joaquin Drainage Waters. Bureau of Reclamation Report. pp. 232.

Gusek, J.J. 2000. Reality Check: Passive Treatment of Mine Drainage and Emerging Technology or Proven Methodology? Presented at SME Annual Meeting, Salt Lake City, Utah, February 28.

Kerr, Brent (United Technologies, Inc.). Personal communication, April 20, 2006.National Research Council. 1999. Irrigation-Induced Water Quality Problems. National Academy Press,

1989.Pahler, Jessie, Walker, R., Rutkowski, T. & Gusek, J. 2007. Passive Removal of Selenium from Gravel Pit

Seepage Using Selenium Reducing Bioreactors, presented at the 2007 Annual Meeting of the American Society for Mining and Reclamation, Gillette, WY.

Ryser, Amy L., Strawna, D.G., Marcus, M.A., Johnson-Maynard, J.L., Gunter, M.E. & Möller, G. 2005. Micro-Spectroscopic Investigation of Selenium-Bearing Minerals from the Western US Phosphate Resource Area. Geochemical Transactions ,Volume 6, Number 1, March 2005.

Sonstegard, J., Harwood, J. & Pickett, T. 2007. Full Scale Implementation of GE ABMet® Biological Tech-nology for the Removal of Selenium from FGD Wastewaters. Proceedings of 2007 International Water Conference.

Wildeman, T.R., Filipek, L. & Gusek, J. 1994. Proof-of-Principle Studies for Passive Treatment of Acid Rock Drainage and Mill Tailing Solutions from a Gold Operation in Nevada. In Proceedings of the International Land Reclamation and Mine Drainage Conference and Third International Conference on Acid Rock Drainage, Pittsburgh, PA, April 24–29, 1994. US Bureau of Mines Special Publication SP 06B-94.


291

Reclamation and closure cost planning and estimation and the mining life cycle

L.E. Boxill & T.E. MartinAMEC Earth and Environmental, Burnaby, British Columbia, Canada

ABSTRACT: Reclamation planning and closure cost estimation are initially evaluated during the mine permitting and application process and typically revisited in a more comprehensive way a few years prior to the anticipated onset of mine closure. It is more often than not the case that the mine plan and the physical and geochemical requirements for closure and reclamation will change, sometimes significantly, from those initially envisioned at project inception. Periodic updates of the mine reclamation and closure plan are often required to assess the financial liabili-ties faced by mining companies for closure. The recent completion of closure cost estimates for some Canadian mines highlighted common issues that typically result in the realization of closure estimates that can significantly exceed both posted closure bonds and the liabilities allocated by mining companies to implement their published reclamation and closure plans.

1 INTRODUCTION

During the application and permitting phase of project planning for many new mines, reclama-tion and closure scoping and costing is consistently less rigorous than is the case for project construction and operation. Prior to project start up, mining proponents are usually keen to dem-onstrate sustainable project feasibility and profitability to project investors, while satisfying the requirements set forth by the applicable regulatory agencies responsible for application and permit approval. This often results in the development of an initial reclamation and closure plan that is: ambitious and ambiguous in scope; under-estimates land disturbance; espouses plans of progres-sive reclamation on the basis of mine plans that can undergo significant change; and projects reclamation and closure costs that are frequently understated.

Lessons learned during completion of final closure cost estimates for several Canadian mines in recent years suggests that failure to carefully assess reclamation and closure liabilities during each phase of the mine life, including conceptual planning and permit application phases, often results in significant underestimation of closure liabilities. As detailed assessment of closure and recla-mation liabilities is typically delayed until a few years prior to the anticipated closure of a mine, the identification of costs that are considerably higher than posted reclamation bonds or internally allocated funds for closure liabilities at this phase in the mine life has the potential to leave mining companies in the undesirable position of: 1) having to notify investors of significantly increased liabilities; and 2) potentially having to designate some or all of the profits derived during the remaining mine life to offset anticipated closure costs. Increased awareness of the detail required to develop sound scopes and estimates of closure costs is needed to avoid such situations.

1.1 The mining life cycle and current reclamation costing practice

The duration and evolution of a mine’s life is typically impacted by several factors. As seen in Figure 1 factors include: results of initial and periodic resource and reserve assessments; compliance with environmental regulations; current and projected trends in the commodities markets; escalation of construction and operation costs; and in some cases waste reclamation and closure considerations.



292

The primary purpose of developing a closure cost estimate is to enable allocation of funds that will hopefully allow for adequate and satisfactory reclamation and final closure of a mine site upon the termination of resource extraction activities. As one of the permit conditions, govern-ment regulatory agencies assign the liability for mine closure and reclamation costs to mining companies by requiring the posting and periodic assessment of a financial security, also known as a bond. In the event of a default by the mining company the bond will be used by the government to contract with an independent third party to complete the necessary reclamation and closure work. It is generally understood by both government regulators and mining companies that bond amounts will be periodically adjusted as follows: 1) to account for changes in the mining plan; 2) to account for costs of reclaiming disturbed areas to satisfy identified end land use objectives; and 3) to account for any other factors that may increase or decrease the cost of reclamation.

In Canada, several provincial regulatory agencies require mining companies to implement a program of environmental protection, and the reclamation of disturbed land and water courses. To support an application for mine approval, submission of a conceptual final reclamation and clo-sure plan which includes assumptions as to both the short and long-term disturbance at the site and an estimate of the costs associated with implementation of the reclamation plan is often required.

Reclamation and closure plans (hereafter RCPs) are typically developed to address key con-cerns generally identified in environmental impact assessments completed during the application and initial permitting phases of the project. However, due to the range of possible impacts indi-cated in the initial project stages, limited bases exist to develop detailed cost estimates. Regular updates of the reclamation and closure plan and the associated cost estimate result in the RCP and its associated costs becoming increasingly refined, in addition to providing a sound basis for implementation of the closure works. However, it is more often than not the case that satis-factory detail in either document is unavailable. While major changes associated with resource extraction are included, these periodic updates typically involve limited revision of the earliest versions of the reclamation and closure plan and the associated estimate. Furthermore, recla-mation and closure costs should be as comprehensive as possible and account for both existing liabilities and those resulting from changes in the mine plan or geochemical factors.

1.2 Typical components of regulatory costing guidelines

Consistent with the regulatory closure guidelines established in several jurisdictions within Canada, reclamation and closure plans and their associated costing generally feature two sections as fol-lows: 1) Costs per hectare to reclaim disturbed areas; and 2) Long term costs typically incurred well

Figure 1. Mining life cycle.


293

beyond the termination of site mining activities that may include the monitoring and treatment of acid rock drainage (ARD), metal leaching, or other conditions impacting environmental quality.

Costs associated with the reclamation of disturbed areas typically include: dump resloping; stabilization and layback of the pit highwall if it exists; the reclamation of pits, ponds and site access roads; the final grading and contouring of the site; overall site revegetation; the excavation and placement of reclamation media; demolition of site infrastructure; abandonment and closure of surface openings where they exist; and capital costs for water treatment facilities if they are needed. Long-term costs typically include operation costs for water treatment facilities, post-closure monitoring and any continuing maintenance costs.

1.3 Shortcomings of current typical cost estimation practice

With the expenditure of significant energies at project startup to demonstrate project feasibility while responding to regulatory concerns typically flagged during the environmental assessment phase, mining companies frequently find themselves making environmental commitments that, while satisfying regulatory agencies and facilitating project approvals, are practically untenable and carry financial implications that would be quite severe were regulators to insist on implementation. Avoidance of this tendency often requires more candid discussions with regulators during the per-mitting phase, which in turn requires a greater degree of regulatory attention to these matters than is commonly applied. Periodic updates of both closure plans and associated costing often become exercises in which: ambitious closure prescriptions remain unchallenged; the scope of outlined closure plan remains limited to the most obvious considerations; and details required to provide sound cost bases are lacking. As a result financial liabilities remain underestimated.

In some instances the importance of reclamation categories provided in regulatory guidelines has been discounted because they were not considered applicable during early stages of mine operation. While liabilities like long-term treatment of acid rock drainage (ARD) conditions or possible metal leaching may appear unlikely during the early phases of mine life, annual envi-ronmental reviews and periodic reviews of the RCP provide opportunities for re-evaluation of potential liabilities that have not previously been considered. Current cost estimation practice also typically underestimates the impacts of changes in the mine plan on selection of stockpile loca-tions for both waste materials and reclamation media.

2 COST DEVELOPMENT PROCEDURE

2.1 Common misunderstandings of cost development procedure

Mining companies commonly retain several misperceptions about the requirements for developing realistic estimates of closure liabilities. The first observation made is that cost estimates appear to quite frequently be developed without accounting for the scope of the RCP. Therefore, it is often the case that in the course of developing the cost estimate, glaring omissions and limitations in the reclamation and closure plan are identified. This finding often requires substantial revision of the reclamation and closure plan to not only make it consistent with the detailed estimate developed, but for the purposes of facilitating internal and external reviews and audits.

In British Columbia, government regulators recognize the existence of two possible cost bases on which closure cost estimates can be developed: 1) cost estimates calculated assuming reclama-tion activities will be completed by a third party contractor; and 2) costs assuming that reclama-tion activities will be completed by the mining company at rates that are typically lower than those realized by third party contractors. The British Columbia Advisory Council on Mining (BCACM) has recommended that a decision as to which cost estimate base is used be dependent upon not only an assessment of the amount of reclamation liability that exists at a particular mine site, but also on the level of risk for default associated with individual mining company. However, in both cases, the developed reclamation and closure plan must be comprehensive. Junior mining


294

companies and those with single mine holdings are generally considered to be at high risk and full security to cover reclamation liability is required. It should also be noted that there appears to be a preference on the part of third party auditors for use of independent third party cost estimates to assess financial liabilities.

Currently development of comprehensive third party final closure costs trends towards the use of third party contracting costs to assess closure liabilities for both government bond revision and ensuring the internal allocation of adequate funds. Internal fleet rates often are not loaded with operating, fuelling, underutilization factors, depreciation costs, and maintenance charges that are included in the development of third party rates. Moreover, the mine fleet is comprised of select pieces of existing mining equipment, namely dozers, front end loaders, graders, and track type excavators, only some of which may be suitable for the reclamation civil works, whereas a 3rd party contractor might utilize a different equipment spread. Pit shovels and ore haul trucks require large working and travel areas and are typically ill-suited to complete reclamation activities many of which are completed along narrow access corridors.

Cost estimate development requires the identification of a clear work scope and as such the presence of vague language expressing general closure concepts that often exists in RCPs is unhelpful. Additionally, mining companies often wish to include potential salvage values (espe-cially for mill equipment) as a means of reducing financial liabilities. Salvage values are expressly excluded from closure cost estimates since the estimate is an indication of existing liability and not potential revenue. However, since salvage values can be significant, mining companies are encouraged to have a specialized consultant complete and independent assessment for purposes of assessing existing assets.

Costing efforts also tend to be focused on the removal and dismantling of obvious signs of land disturbance while funding for indirect costs such as monitoring, testing, reporting and treatment tends to be limited even when these costs tend to extend well beyond the termination of resource extraction activities at the mine site.

2.2 Evolution of the mine plan and closure cost estimation

The scope of the RCP is directly tied to the activities completed during execution of the mine plan. Integration of reclamation costs with the planning and scheduling of overall mining costs is desirable. It therefore follows that from the time of initial development of the mine plan and its subsequent changes, the implications of mine plan execution should be addressed in the RCP. It is periodically the case that short term benefits in terms of alteration to the mine plan can result in negative net returns due to significantly increased closure liabilities. If the RCP is updated in conjunction with considered changes to the mine plan, such situations of “short term gain for long term pain” can be avoided.

As a mining project is developed careful consideration should be given to not only the achieve-ment of short term goals (e.g. construction of a given access road) but to how long-term execution of the mine plan either facilitates ultimate reclamation and closure activities (e.g. creation of designated and quantifiable overburden stockpiles in locations that can be preserved throughout the mine life), or impedes execution of the RCP (e.g. the inadvertent burial or inundation of reclamation media thereby limiting quantities of reclamation media available to close the mine site). The mine plan should therefore be developed and executed in a way that does not contravene the stated objectives indicated in the published RCP. In other words, the RCP and the mining plan should be in dynamic, responsive, and consistent relationship throughout all phases of the mining project. This principle is most effectively demonstrated by the implementation of progressive reclamation activities.

2.3 The role of progressive reclamation

The intent of progressive reclamation is to limit overall final closure liability by completing reclamation of previously disturbed areas that are not anticipated to be redisturbed during the remaining mine life. Progressive reclamation is also a means of demonstrating to mine regulators


295

the active limitation of the disturbed areas which must be accounted for in annual environmental reports and as a means of minimizing the closure bond. In many cases, regulators offer financial incentives in the form of rebates for completed reclamation. Activities typically include: the grad-ing and revegetation of borrow areas; regrading and revegetation of inactive areas on waste rock dumps and tailings impoundments; and the regrading and revegetation of other disturbed areas at the mine site that have been abandoned. However, the extent to which progressive reclamation can occur is typically impacted by not only the mine plan but by the resource exploration plan. In other words, it is quite rare to find large areas of previously disturbed land, or even initial mine explora-tion roads that can be reclaimed prior to the termination of active mining at the site.

While the extent of progressive mine reclamation may be limited, careful monitoring of areas reclaimed under a well developed program can provide information that could be potentially very useful for assessing the effectiveness of existing reclamation and closure prescriptions. Informa-tion potentially gleaned from such programs includes: vegetation suitability and growth rates; the need for fertilizers and associated performance; the effectiveness of erosion control measures; and the performance of revegetated covers.

3 COMPREHENSIVE CLOSURE COST DEVELOPMENT

Regulatory agencies typically acknowledge that reclamation costs are dependent on site-specific conditions and as a result standardized unit costs for reclamation items are not provided. Equip-ment costs can typically be developed using trade guidelines like the B.C. Road builders annual equipment rates provided in the Blue Book or the RS Means which provides annual updates of equipment and fleet rates for projects located in the United States. Costing for the decommission-ing of sewage and water treatment plants and other structures will also likely require costing input from specialty contractors.

Equipment and labor costs associated with the reclamation of disturbed areas typically account for the largest portion of mine reclamation provided long-term water treatment is not required. Although equipment selection impacts the reclamation cost estimate, many estimates are based on the use of equipment that: a) is not well suited to complete the identified reclamation work; and b) is not typical of the equipment fleet that would be used by a 3rd party contractor retained to complete the identified work. While pieces of large mining equipment may work well to trans-port ore and waste rock from the extraction site to various staging and processing locations, this equipment is often too large to complete many reclamation tasks efficiently. Even the re-location of waste rock within the limits of an abandoned open pit below the eventual flooding level may be best achieved using a large track-mounted excavator and suitably sized rigid or articulated truck. Equipment fleet selection should be directly linked to equipment suitability for the size and scope of the task being priced. The capacity and number of haul trucks required to complete a particu-lar earth moving activity is directly impacted by both the size of the loading excavator and the distance between loading and dumping sites. Additionally, equipment rates should at a minimum include labor, fuelling, and maintenance costs. An inefficiency factor should also be included in fleet rate development to account for the unavoidable occurrence of equipment and operator downtime. Typical hourly inefficiency ranges between five and fifteen minutes per shift hour.

Finally, the role of the selected discount rate in determining the adequacy of posted financial security should not be underestimated. To present closure costs in current dollars it is necessary to make some assumptions regarding when costs will be incurred in time. While there will likely be deviations from initially assumed timelines and those realized during closure, variations typically increase the schedule timeframe and thus make the discounted figures in the closure plan conserva-tive. When developing a schedule of closure costs the following assumptions are commonly made:

1. Final closure is assumed to occur instantaneously today where t = 0;2. Final design and refinement of closure measures is assumed to occur during the first year after

closure;


296

3. Construction of all closure measures is assumed to occur during the second and third years, with total construction costs spread evenly over those two years;

4. There is a minimum five year monitoring program after the construction period to evaluate the effectiveness of the closure measures; and

5. A definition of “long-term” as 100 years is adopted to facilitate development and discounting of long-term monitoring and treatment costs.

While mining companies are typically in favor of using discount rates equal to and greater than five percent, in British Columbia and Ontario, a discount rate of three percent is generally accept-able to regulators.

4 LESSONS LEARNED FROM COMPLETED ESTIMATES

In the last few years several third party comprehensive closure cost estimates for mines located in various Canadian jurisdictions were completed. In all instances it was observed that several key items were excluded from the internal estimates developed by the mining companies, resulting in significant increases in the assessed financial liabilities associated with closure and reclamation. In one instance final estimated closure costs resulted in quadrupling of the undiscounted cost esti-mate. In another instance, discounted closure costs resulted in a 50 percent increase in reclamation financial liability. Exclusions from internal estimates included:

1. Funding to develop the engineering designs to implement the reclamation and closure plan; 2. Detailed designs for the decommissioning of runoff diversion systems; 3. Indirect costs associated with the completion of reclamation activities; 4. Sound bases to assess the need for and hence develop a cost estimate for water treatment

facilities that would likely be required; 5. Funding to determine how identified metal leaching concerns could be adequately addressed

and monies to design adequate cover systems or mitigation strategies; 6. Consideration of the potential impact of long-term costs for water quality and dam safety

monitoring; 7. Adequate funding for the dismantling and reclamation of mine site infrastructure; 8. Monies for the restoration of overall site hydrologic patterns; 9. Disposal plans for miscellaneous scrap metal, non-biodegradable waste (especially for tires)

typically found in laydown areas; and10. Long-term maintenance costs that may be required to facilitate site access for periodic inspec-

tions post mine closure.

While the scope of reclamation and closure plans and the associated cost estimates are site spe-cific, Figure 2 and Figure 3 show general components of overall closure costs for scenarios includ-ing and excluding conventional water treatment systems. It is important to note that even in cases without water treatment needs, long-term monitoring costs can represent a significant percentage of overall closure costs. The magnitudes of components of closure costing indicated on the figures will vary depending on the specific scope of closure work required for each mine site.

The following sections describe the major factors that should be accounted for during the devel-opment of comprehensive reclamation and closure estimates.

4.1 Engineering design

Engineering design costs include expenditures to complete: long-term assessments of ARD potential; open pit stability; wildlife and vegetation impacts; developing a regrading plan for waste management facilities; overall site grading for drainage; decommissioning of runoff diversion systems; and development of short and long term geotechnical and water quality monitoring pro-grams and remote monitoring systems. Detailed design and development of construction drawings


297

Initial Engineering, 2%

Initial Construction, 30%

Long Term Treatment, Maintenance and Monitoring, 40%

Site Abandonment, 2%

Contingency, 15%

Construction Management & QC, 1%

Closure measure Assessment and Reporting, 1%

Short-term Monitoring, 10%

Figure 2. Example of cost breakdown for overall closure costs with water treatment.

Initial Engineering and Construction

20%

Short-term Monitoring2%

Long Term Maintenance and Monitoring

51%

Site Abandonment10%

Contingency17%

Figure 3. Example of cost breakdown for overall closure costs without water treatment.

and specifications to implement closure measures is completed by engineering consultants and is typically assumed to occur in Year 1. Costs also include indirects such as on-site construction man-agement, quality control, cost control, inspection and as-built reporting which typically coincide with the period in which short term closure measures are implemented.


298

4.2 Indirect cost development

During development of third party independent contractor costs, indirect costs must be accounted for. These costs can be quite significant and are often associated with completion of each reclama-tion and closure activity. These costs include:

1. Fleet mobilization;2. Establishment and dismantling of field offices and associated field office operation costs;3. Creation of fuel storage, a refueling station, and laydown areas;4. General maintenance and repair of equipment;5. Work crew travel to and from site or camp operation costs;6. Completion of independent topographic surveys for QA/QC verification;7. General costs associated with project and construction management;8. Project insurances and bonds; and9. Contractor profit.

Based on the remoteness of the site, the scope of work to be completed, and estimated project duration, indirect costs can range between thirty and sixty percent of the amount calculated for direct costs for the completion of civil works.

4.3 Waste rock dumps

The height and existing steep slopes of some waste rock dumps sometimes render impractical regrading and revegetation of all dumps slopes. However, since reclamation of waste rock dumps features prominently in many reclamation and closure plans, careful planning is required through-out mine operations to facilitate execution of the proposed closure plan. Reclamation activities will likely include: full or partial surface resloping; placement of rooting media; and surface revegetation. The mine plan should provide for dump development that limits the amount of res-loping (in effect, material rehandle) that may be necessary and incorporate benches to reduce slope lengths. Periodic resloping of lower slopes may also limit reclamation efforts.

Waste rock management plans must incorporate water quality and control aspects including: rock characterization, dump volume, location, and chemical characterization of water having contact with these materials. The location and quantities of geochemically problematic rock is documented in annual environmental reports and various strategies are available for handling such rock. It should be noted that while disposal of potentially acid generating (PAG) often receives much attention, the leaching of neutral metals like Selenium, has become increasingly important. Moreover, due to potential issues associated with metal leaching and the complexity of dump flow regimes, mine companies should be wary of placing these materials with NPAG materials due to the significant costs associated with installing engineered composite cover systems over entire dumps. Care should therefore be taken to isolate materials that have any suspected metal leaching potential so that if use of an engineered cover is required the extent of the cover can be limited to the area of concern.

4.4 Tailings management facilities

The condition of the downstream slope, crest, beach and abutments of a tailings impoundment are key factors affecting not only the scope of the reclamation prescription required for reclamation and closure, but for selection of a suitable equipment fleet. Surfaces to be revegetated must first have shallow enough slopes to facilitate placement and long-term retention of reclamation media. It is also important for tailings impoundments to be designed, constructed and operated in a man-ner consistent with closure objectives especially as it relates to slope configuration, freeboard allowances, and beach creation. Available quantities of reclamation media are distributed among the various reclamation sites to achieve the coverage depths identified in the RCP. Furthermore, depending on substrate conditions, reclamation media cover depths generally range from 0.15 m to 0.6 m thick based on a review of closure prescriptions for several mining projects across Canada completed in the last twenty years.


299

Based on the tailings dam’s hazard classification, inspection and dam safety monitoring criteria set forth in the Canadian Dam Safety Guidelines must also be accounted for. Costs also need to account for: decommissioning of existing diversion systems; development of long-term abutment and groin drainage; and installation of closure spillways and erosion protection.

4.5 Long-term tailings impoundment monitoring, repair, and reporting requirements

Costs associated with completing dam safety inspections and periodic maintenance need to be accounted for. Depending on the remoteness of the mine site, consideration should be given to including costs to install a remote monitoring system that would include the use of closed circuit television (CCTV) and programmed alerts indicating the occurrence of conditions like spillway blockages that could potentially have adverse impacts on dam safety. Funds should also be allocated to complete repairs, some of which could be major, that are identified during inspec-tions. The maintenance of long-term post closure dam insurance as a means of assisting with the continued management of liability associated with the tailings dam post closure might be considered. The cost of completing periodic inspections, comprehensive dam safety reviews con-sistent with the hazard classifications provided in the Canadian Dam Safety Guidelines (2007), and the preparation of the associated reports should also be accounted for.

4.6 Long-term water treatment, monitoring and reporting requirements

The potential for long-term monitoring and treatment costs should be carefully evaluated through-out the mine life. Environmental, geochemical, hydro-geological, and engineering assessments required to determine the potential for and likely scope of long-term dam safety, water treatment, monitoring, and maintenance needs should be completed since they enable a more comprehensive assessment of closure and reclamation liabilities.

From a water quality perspective, detailed analyses should be completed to determine the long-term potential for ARD and/or metal leaching impacts on environmental quality post closure, particularly if rock slides have exposed rock with high acid generating or metal leaching potential. The presence of metal leaching liability could also trigger the need for development of an engi-neered composite cover system which is generally quite costly. In some instances water quality effects are not observed until well into closure but must still be accounted for.

Due to the significant capital and operation costs associated with a water treatment plant, com-pletion of assessments to determine conclusively: a) if a water treatment plant will be required; b) what type of treatment will be required; and c) what treatment duration will be required is paramount. In instances where it remains unclear as to when construction of a water treatment plant might occur, the effect of the selected discount rate and anticipated construction year can be assessed using basic sensitivity analysis. Figure 4 provides an example of how capital costs for a relatively small conventional water treatment facility vary with time and shows the impact of the selected discount rate. Figure 5 shows the variance in operation costs with start up time for a small conventional treatment plant. Long-term treatment typically assumes operation of a conventional water treatment plant for a period of at least 100 years after plant commissioning. Due to the long operation period, the selected discount rate and operation start time significantly impact the dis-counted operation costs.

Costs also account for the completion of water quality reporting as required by the respec-tive provincial and federal regulators. Inspection and reporting costs should be developed based on assessments of the staffing levels and estimates of time required to complete the required activities.

4.7 Restoration of site drainage

While the physical restoration of drainage courses requires funding allocation, costs associated with completion of hydrological analysis to review overall site drainage and management of


300

catchment areas should not be overlooked. Funds should also be allocated for decommissioning impoundments and other diversion structures.

4.8 Long-term site maintenance

These costs should be accounted for in instances where site access beyond mine closure is anticipated. This need typically arises when long-term operation of a water treatment facility and physical monitoring and inspections are required. Costs typically include periodic operation of a grader and snow plow to maintain the required site access. It should also be noted that costs should

Figure 4. Variance of WTP capital cost with construction time frame.

Figure 5. Variance of WTP operation costs with start year for operation.

$0

$500,000

$1,000,000

$1,500,000

$2,000,000

$2,500,000

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Construction Year

Dis

co

un

ted

Co

st

3% Discount Rate 5% Discount Rate

$0

$2,000,000

$4,000,000

$6,000,000

$8,000,000

$10,000,000

$12,000,000

$14,000,000

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80

Start Year for WTP Operation

Dis

co

un

ted

Co

st

3% Discount Rate 5% Discount Rate


301

account for any required width reduction of existing haul roads that may be incorporated into the long-term site access plan.

4.9 Reclamation of potentially contaminated sites

These costs account for characterization of soils in the vicinity of the existing refuelling and equip-ment laydown areas and development of suitable soil remediation strategies for contaminated soils identified. Costs are developed with the assistance of a senior environmental engineer primarily responsible for the design and implementation of soil remediation technologies. While clay liners are generally incorporated within the footprint of tank farms and refueling sites, assessment of limits of the contaminant plume should be completed to facilitate costing of mitigation measures.

Publicly available costing for soil remediation technologies provided by the United States Envi-ronmental Protection Agency and United States Army Corps of Engineers may also provide some guidance during cost development.

4.10 Infrastructure decommissioning

Infrastructure decommissioning costs typically account for removal of structures providing offices, equipment maintenance and repair, warehouses, mill and process facilities, steel bents supporting conveyor systems, accommodation and dining facilities, and sewerage treatment facilities. Costs of decommissioning existing electrical service lines and their associated substations are typically handled as a separate exercise since long-term energy requirements are typically considered if the need for a water treatment facility is identified. Costs to remove steel frame structures are signifi-cant and should be developed in consultation with an experienced contractor specializing in the dismantling of the involved building types. Dismantling costs generally assume the removal and relocation of building contents to an offsite salvage depot or scrap yard. Overburden stockpiles to reclaim the footprints of the plant site structures should also be identified so that associated earthwork costs can be calculated. Site reclamation costs may also include removal of PAG mate-rial that had been used to create building foundations.

4.11 Lay-down areas and disposal of incidental waste

The disposal of waste products like used tires, scrap metal, storage tanks, and other non-biodegradable wastes on the mine site should also be accounted for. While these costs may be small in comparison to other reclamation tasks, the reclamation of these areas should be accounted for.

It should also be noted that salvage values from the sale of scrap metals are not accounted for in closure costing. Arrangements should be made during mine operation to minimize quantities of miscellaneous waste that exist on site. Implementation of metal and tire recycling programs during mine operations will not only reduce onsite waste product quantities but provide useful information about costs and potential revenues realized through the offsite relocation of these waste products. The possibility of in-pit or underground disposal for some or all incidental waste products should also be explored with the regulatory agencies.

4.12 Other cost considerations

Other tasks that are commonly omitted or underestimated in closure cost estimates include: rec-lamation and closure of dump ponds and sediment control structures containing tailings; and development of re-vegetation costs for forested and non-forested areas. Key points associated with these tasks are as follows:

1. Reclamation of dump ponds requires that all contaminated waste contained be relocated to an acceptable disposal site usually identified within the limits of a tailings impoundment or beneath ultimate flooding limits of an open pit. The cost of transporting saturated tailings is


302

generally well in excess of transport costs for dry material and need to be accounted for accord-ingly. Costs for this work package account for the dewatering of all sediment control structures including tailings dump ponds and impoundments. This work may be completed using either a series of pumps or a combination of haul trucks and an excavator.

2. Revegetation costs should account for a mortality rate since sustained site revegetation is the intended objective of this task. To achieve long-term establishment of planted vegetation, a secondary revegetation campaign is typically required within two years of the initial program. Information about both success and mortality rates for the various species included in trial vegetation programs completed throughout the mine life is also helpful in determining more comprehensive revegetation costs.

Finally, if site remoteness and access are challenges during mine start up, similar and poten-tially greater challenges could exist during mine closure especially as it relates to: the disposal of petrochemical or other contaminated waste products; the completion of periodic monitoring; or directly impact the potential salvage values realized for site materials and equipment. Funds should also be allocated to prepare a report describing the effectiveness of the implemented closure measures that may be required by regulators within five years of mine closure.

5 SUMMARY

Initial RCPs should be developed with a clear understanding as to how promised concessions and reclamation measures both impact the proposed mine plan and the potential costs associ-ated with implementation of identified closure prescriptions. RCPs should therefore: have a firm basis in practicality; be developed with careful consideration of the initial mine plan; allow for revisions and expansions resulting from likely modifications to the mine plan; and comprehen-sively respond to changed geochemical and environmental conditions that likely occur during mine operating life.

The preparation of cost estimates that do not fully account for existing reclamation and closure liabilities do not ultimately serve the interests of mining companies, their investors, or regulatory agencies. Time taken throughout the mine life to carefully consider closure liabilities is therefore time well invested and has the potential of: realizing significant cost savings due to its direct impact on material handling practice (i.e. minimizing the occurrence of material re-handling); the strategic positioning of reclamation media and other stockpiles; and the implementation of effec-tive progressive reclamation plans.

The comprehensive assessment of existing and potential liabilities and their associated costs throughout the mine life will not only result in the development of more realistic bonds or fund-ing allocations but will likely result in innovative methods that could potentially translate into increased closure cost savings. It is further believed that if reclamation and closure cost estimation could be transformed from a reactive to a proactive and progressive exercise, many of the surprises associated with the realization of inadequate allocation of funding to meet closure liabilities just prior to the onset of mine closure can be at worst minimized and at best avoided.

REFERENCES

BC Road Builders and Heavy Construction Association. 2006. 2006–2007 Equipment Rental Rate Guide The Blue Book. Burnaby: Westholme Graphics, Inc., June.

Canadian Dam Association. 2007. Canadian Dam Safety Guidelines.Ministry of Energy and Mines. 2004. Mine Reclamation Costing and Spreadsheet Manual, Victoria.Ministry of Energy and Mines. 2003. Health, Safety and Reclamation Code for Mines in British Columbia,

Victoria.


Policies, procedures, and public safety



305


Uranium tailings facility design and permitting in the modern regulatory environment

K. Morrison, J. Elliott & J. JohnsonGolder Associates Inc., Lakewood, CO, USA

B. MonokEnergy Fuels Resources Corp., Lakewood, CO, USA

ABSTRACT: Strategically located within the Uravan Mineral Belt District, the Piñon Ridge Project is the first new uranium mill being proposed for construction in the USA in over 25 years. In 1992, dra-matic changes to the regulatory environment for uranium ore processing occurred making regulatory compliance an increased challenge for a new facility. Such changes to the regulations include enforce-ment of a prescriptive liner for surface impoundments which includes primary and secondary liner systems separated by a leak detection and collection layer, with the lower composite liner system com-prised of a geomembrane underlain by three feet of low permeability clay. The project includes design and permitting of three 30.5-acre tailings cells, as well as the process facilities, evaporation ponds, and ore pads. This paper will focus on the regulatory requirements pertaining to the tailings cell design.

1 INTRODUCTION

Energy Fuels Resources Corporation (EFRC) is in the process of designing and pursuing licensing of a new conventional uranium mill strategically located within the Uravan Mineral Belt District. The proposed mill, termed the Piñon Ridge Project, is located near Naturita, Colorado, and is the first uranium mill being proposed in the USA in the past 25 years. With the resurgence of the ura-nium industry, local interest in the project is high as this area has exhibited slow economic growth since the collapse of the uranium industry nearly 30 years ago.

With climate change policies in play on Capitol Hill, the government’s interest in ‘clean’ power sources is in full swing. Nuclear energy is considered clean energy as it does not contribute to greenhouse gas emissions (Weakly 2008). The USA is currently the largest generator of commer-cial nuclear power, with nuclear energy contributing to approximately 20 percent of the Nation’s electricity (Weakly 2008). In 2002, the Nation’s nuclear power plants required about 18,000 metric tons of Uranium (tU) with the USA producing only 1000 tU, importing the remainder from pre-dominantly Canada and Australia (Finch 2003). The Piñon Ridge Project, in conjunction with the development of new or re-opening of closed uranium mines in the Uravan Mineral Belt is antici-pated to decrease some of this dependence on foreign uranium sources.

As part of the Piñon Ridge Project, uranium tailings are proposed to be disposed in three 30.5 acre tailings cells with a combined capacity to store seven million tons (Mt) of tailings. The project has a 20-year design life at a design mill throughput of 1000 tons per day (tpd) of tailings. Project construction is anticipated to commence in 2010. Figure 1 is a photo of the proposed mill site, located in the Paradox Valley.

1.1 Regulatory background

Under the Uranium Mill Tailings Radiation Control Act (UMTRCA) of 1978, as amended, the Environmental Protection Agency (EPA) has the responsibility of establishing standards for


306

exposure of the public to radioactive materials originating from mill tailings for active uranium extraction facilities licensed by the US Nuclear Regulatory Commission (NRC) or its Agreement State. NRC regulations are issued under the United States Code of Federal Regulations (CFR) Title 10, Chapter 1.

Located in the Agreement State of Colorado, the Piñon Ridge Project is pursuing permitting through the Colorado Department of Public Health and Environment (CDPHE). As an Agreement State, the state has the responsibility for licensing the possession and use of radioactive materi-als in Colorado under the Radiation Control Act (Title 25, Article 11), and Colorado’s Rules and Regulations Pertaining to Radiation Control (6 CCR 1007-1). As of January 2006, thirty-three states have entered into agreements with the NRC, under which regulatory authority has been delegated to the state over most radioactive materials used in non-federal facilities, pending that the state program is compatible with NRC requirements.

1.2 Key regulatory requirements

Regulations pertaining to the design and operation of uranium tailings disposal cells became increasingly more stringent in 1992, including the requirement for tailings cells to have a double liner separated by a leak detection and collection layer, with the lower composite liner system comprised of a geomembrane underlain by three feet of low permeability clay. Further, the revised regulations promote fully below-grade tailings disposal. Because the uranium industry is now awakening from a long recession, these regulations are now coming to the forefront, needing to be implemented for the first time since their inception over 15 years ago for new uranium tailings disposal projects.

Figure 1. Photo of the proposed Piñon Ridge uranium mill site, looking north across the valley.


307

The Colorado state regulations indicate that flexibility is provided in the criteria which allows for optimal tailings disposal on a site-specific basis. Several of the guidelines for uranium tailings disposal include (6 CCR 1007-1, Part 18, Appendix A):

• Consideration of disposal of tailings below grade, either in mines or pits, which is stated as the ‘prime option’ for disposal of tailings (Criterion 3);

• Providing good wind protection to the tailings via topographic features (Criterion 4B); and• Employing the groundwater protection standards set forth in 40 CFR Part 192, Subparts D

and E (Criterion 5).

2 GENERAL TAILINGS CELL DESIGN CONCEPTS

The Piñon Ridge Mill is designed to operate at 1000 tpd with an expected life of 20 years. Each of the three proposed tailings cells (Cells A through C) have been designed (i) to provide capac-ity for 6.7 years, (ii) with liner footprint areas of 30.5 acres, and (iii) with a minimum capacity

Figure 2. Tailings cell layout indicating location of section A-A'.


308

to accommodate storage of 2.45 million tons (Mt) of tailings with three feet of freeboard. The tailings cell layouts are illustrated in plan view in Figure 2. Applicable criteria of 6 CCR 1007-1, Part 18 (Appendix A) have been considered in the tailings cell investigation and design work.

The tailings cells were designed for construction predominantly in the existing subgrade, with an excess cut of approximately 230,000 to 300,000 cubic yards dedicated to future closure cover construction. The tailings cells were developed by designing a perimeter embankment with a width of 15 feet to facilitate berms and one-way light truck traffic. The tailings cells have internal side slopes of 3H:1V, and a minimum base grade of one percent. The limits of the tailings cells are lined with a double layer liner system with an intervening leak detection system to contain process solutions, enhance solution collection, and protect the groundwater regime. Intermediate benches have been incorporated in the design to provide additional anchorage of the underliner component of the secondary composite liner system, as well as buttressing of the liner to limit wind uplift.

As a precautionary measure, the first tailings cell (Cell A) slated for construction has been designed as a split cell to facilitate separate collection of process solutions for redundancy during facility start-up if unforeseen problems with the liner system develop, allowing half of the cell to be decommissioned and repaired while continuing mill operations (Fig. 3). Design of future tail-ings cells (Cells B and C) includes a split-cell option with anticipation of a single-cell design.

3 TAILINGS CELL LINER SYSTEM DESIGN

The groundwater protection standards (40 CFR 264.221) which the uranium tailings disposal cells must meet effectively define the liner system requirements. The prescriptive liner system consists of a double layer liner with the primary and secondary liners separated by a leak detection system (from top to bottom):

• Upper primary geomembrane liner;• Leak detection layer, consisting of drainage gravel or geosynthetic material (i.e., geonet) which

meets prescribed minimum permeability or transmissivity values;• Lower secondary geomembrane liner; and• A minimum of three feet of low permeability clayey soil underliner with a maximum perme-

ability of 10−7 centimeters per second (cm/sec).

3.1 Liner system summary

At Piñon Ridge, investigative drilling to depths of up to 600 feet below the ground surface did not encounter groundwater under the planned location of the tailings cells. Additionally, a number of aquitards were identified during the geotechnical field program, further limiting any potential impacts to the groundwater regime during the active life of the mill. Despite this site specific char-acteristic, the tailings cells were nevertheless designed with the standards applicable to hazardous waste treatment, storage and disposal facilities in accordance with the above prescriptive liner system for groundwater protection (from top to bottom) (Fig. 4):

• 60 mil high density polyethylene (HDPE) upper (primary) geomembrane;• Leak Detection System (LDS) consisting of HDPE geonet on the base of the tailings cells, and

a drainage geocomposite on the side slopes;

Figure 3. Cross-section A-A' through proposed tailings Cell A.


309

• 60 mil HDPE lower (secondary) geomembrane;• Geosynthetic clay liner (GCL) as the underliner component of the secondary composite liner

system; and• Prepared subgrade.

When closure includes leaving the liner system in place in perpetuity, which is applicable to the Piñon Ridge Project tailings cells, the geosynthetic components are required to be designed of materials which prevent wastes from migrating into the liner during the active life of the facility. When closure of the facility includes removal of the liner system or decontamination of all waste residues, wastes are permitted to migrate into the liner during the active life of the facility.

3.2 Upper primary geomembrane liner

Since the upper primary geomembrane will be exposed to atmospheric conditions as well as a variety of chemicals associated with uranium and thorium processing, use of a high density polyethylene (HDPE) liner is proposed. Although standard (black) HDPE has been proven to be highly resistant to ultraviolet radiation over long periods of time, white HDPE is proposed to fur-ther increase this weathering resistance by reflecting more solar radiation while at the same time reducing the range of expansion/contraction resulting from temperature fluctuations and reducing desiccation effects to subgrade soil materials. While the bottom of the cells will be exposed for a matter of weeks or months during initial tailings deposition, the upper portion of the primary geomembrane liner is anticipated to be exposed for nearly seven years.

Single-sided texturing (textured side down) on the upper primary geomembrane is considered to increase frictional resistance at the contact with the LDS layer. Textured rubsheets will be extru-sion welded where required by mill operations to facilitate tailings deposition and access during operations.

Additionally, use of conductive liner is proposed for the upper geomembrane to aid in quality assurance testing. After the liner system is installed, spark testing will be conducted on the con-ductive liner as a cost-effective and precise way to detect defects in the liner.

3.3 Leak detection system layer

An important feature of the tailings cell liner system is the Leak Detection System (LDS) layer, designed per 40 CFR 264.221 (by reference from 10 CFR 40 and 6 CCR 1007-1, Part 18). If a leak occurs in the upper primary geomembrane, the LDS is designed to minimize the hydraulic

Figure 4. Tailings cell liner system detail.


310

heads on the lower geomembrane liner by utilization of HDPE geonet in the base of the tailings cells and a drainage geocomposite on the side slopes. Geonet was not considered suitable for use on the long side slopes of the tailings cells due to its anticipated low interface shear strength when placed in contact with geomembrane. Instead, a drainage geocomposite, comprised of a geonet laminated on both sides to a nonwoven geotextile filtration media, is proposed to increase fric-tional resistance with the overlying and underlying textured geomembrane liners.

In the event that leakage occurs through the upper geomembrane liner, it will be collected in the LDS layer and routed via gravity flow to a LDS sump. The LDS sump was designed to accommodate eight hours of maximum flow in the LDS layer, assuming one liner defect per acre for good installation (Giroud & Bonaparte 1989), an effective porosity of 30 percent in the sump, and applying a factor of safety of 1.5. Collected solution will be recovered via an automated submersible pump installed in one of two HDPE riser pipes, and pumped back to the tailings surface.

3.4 Lower secondary composite liner system

The lower secondary composite liner system underlies the LDS layer to maximize the amount of solution recovered in the LDS and act as a final flow barrier, protecting the subgrade. This com-posite liner system consists of a 60 mil HDPE double sided textured geomembrane overlying a geosynthetic clay liner (GCL). HDPE was selected due to its natural resistance to the chemicals in the solution, and the double sided texturing is used to increase the frictional resistance with the overlying and underlying geosynthetic layers.

Due to lack of locally-available clay sources in the vicinity of the Piñon Ridge Project, alter-native underliner materials which meet or exceed the prescribed underliner (i.e. three feet of 10−7 cm/s clayey material) were considered. Alternatives evaluated included bentonite amend-ment of on-site silty and sandy soils to achieve a low permeability underliner, and use of GCL as the underliner material. The use of GCL was ultimately recommended for this site as soils amended with up to three percent bentonite tested to be nearly one order of magnitude more permeable than the prescriptive clay liner requirement. Calculations according to the method proposed by Giroud et al. (1997) demonstrate that the secondary liner system containing a GCL performs better than the secondary liner system containing the prescriptive clay liner in terms of limiting fluid flow. At the Piñon Ridge Project site, use of a GCL underliner appears to the most cost-effective and constructible solution to meet (or exceed) the regulatory requirements.

4 UNDERDRAIN SYSTEM DESIGN

Per Criterion 5E(3) of 6 CCR 1007-1, Part 18, Appendix A, the tailings cells have been designed to facilitate dewatering of the tailings (i.e. lower the phreatic surface and reduce the driving head for seepage) via an underdrain system installed at the base of the impoundment. The tailings are expected to consist of silty sand to sandy silt materials, which are considered amenable to dewater-ing, particularly if some segregation by particle size results from deposition as dilute slurry.

The tailings underdrain system is comprised of collection pipes at the base of the tailings cell which convey solution to underdrain sumps. The underdrain collection pipes are proposed to con-sist of perforated corrugated HDPE pipes placed in trenches, which are backfilled with imported granular drainage materials. The underdrain sump will be constructed above the LDS sump to provide head for pumping of collected seepage. The underdrain sump is proposed to be backfilled with coarse underdrain fill overlain by fine underdrain fill to ensure filter compatibility with the overlying tailings. Two underdrain riser pipes are proposed within each sump to add redundancy to the system. The lower ends of the risers are slotted in the sump area to provide solution access. Solution is recovered via an automated submersible pump installed in the riser and returned to the mill circuit.


311

5 TAILINGS DEPOSITION MODELING

Tailings deposition for tailings Cell A was modeled using the proprietary GoldTail software pack-age. The purpose of the tailings deposition modeling is to provide mill operations personnel with a method for tailings discharge which enhances design of the tailings cells by providing protection to the constructed underdrain system from potential slimes clogging, as well as providing initial buttressing to the geomembrane liner system. Tailings deposition was modeled within Cell A in the following five simplified phases:

• Phase 1—Deposition commences within sub-cell A1 (or A2) in the vicinity of the underdrain sump to provide approximately 10 feet of tailings deposition over the sump area. This phase of deposition provides coarse-grained underflow tailings over the underdrain sump to enhance the effectiveness of the tailings underdrain system;

• Phase 2—Continued deposition within the remainder of the first sub-cell to push the pond toward the sump area;

• Phase 3—This phase was modeled with deposition commencing within the other sub-cell in the vicinity of the underdrain sump, again providing approximately 10 feet of coarse-grained underflow tailings over the underdrain sump area. During actual operations, it is recommended

Figure 5a. Tailings cell pre-deposition. Figure 5b. Tailings cell phase 1 deposition.

Figure 5c. Tailings cell phase 2 deposition.

Figure 5e. Tailings cell phase 4 deposition.

Figure 5d. Tailings cell phase 3 deposition.

Figure 5f. Tailings cell phase 5 deposition.


312

to reverse the order of the modeled phases 2 and 3 in order to buttress the geomembrane liner system within both sub-cells at the on-set of operations, prior to completely filling the first sub-cell;

• Phase 4—Continued deposition within the remainder of the second sub-cell to push the pond toward the sump area; and

• Phase 5—Once both sub-cells are filled, tailings deposition will proceed along the perimeter of the entire tailings cell in stages (as dictated by tailings operations), until the tailings cell is full (with three feet of freeboard provided at the perimeter of the cell).

Tailings deposition modeling was effective for determining a tailings discharge method, includ-ing approximate discharge locations and duration of discharge for each deposition phase. The modeled deposition phases are illustrated in Figures 5a–5f.

6 CLOSURE CONSIDERATIONS

The tailings cells for the Piñon Ridge Project have been designed to consider closure and to inte-grate the design for compatibility with the following concepts:

• Minimize the need for long-term active site care and maintenance during the post-closure period;

• Perimeter berms developed with side slopes of 5H:1V (per Criterion 4C of 6 CCR 1007-1, Part 18, Appendix A), which may be flattened to 10H:1V during construction of the final closure cover;

• Placement of an interim cover over each tailings cell as deposition is complete within the tailings cell to limit exposure to radiation until final cover placement at the end of milling operations;

• Dewatering of the tailings as feasible prior to placement of closure cover materials;• Provide additional capacity within the tailings cells to accommodate future closure considera-

tions, such as disposal of the liner systems removed from the evaporation ponds and ore pads, etc., during site closure activities; and

• Construction of a final closure cover which meets the requirements of Criterion 4D (6 CCR 1007-1, Part 18, Appendix A) with regard to erosion protection, as well as limiting radon flux to acceptable levels (per Criterion 6, 6 CCR 1007-1, Part 18, Appendix A).

7 CONCLUSIONS

Sixteen-year old regulations governing the design of new uranium mill tailings facilities are cur-rently being applied for the first time at a new facility for the Piñon Ridge Project, located in west-ern Colorado. As an Agreement State, Colorado has the authority to license the possession and use of radioactive materials in Colorado under the Radiation Control Act (Title 25, Article 11), and Colorado’s Rules and Regulations Pertaining to Radiation Control (6 CCR 1007-1).

In general, the regulations provide minimum requirements, many of which are somewhat flexible in order to allow for optimization on a site-specific basis. However, thorough justi-fication is required for any deviation from the prescriptive design standards. With mounting public concern for the environment, it is now more important than ever to design and con-struct potentially hazardous facilities in a manner which minimizes present and future adverse impacts on the environment. Abiding by the proposed regulations is the first step in curbing these impacts.

Although the project team has been in constant communication with the responsible regulatory agency, the Colorado Department of Public Health and Environment (CDPHE), formal approval of the proposed tailings cell design has not yet been issued.


313

REFERENCES

6 CCR 1007-1, Part 18—“State Board of Health Licensing Requirements for Uranium and Thorium Process-ing”, specifically Appendix A (Criteria relating to the operation of mills and the disposition of the tailings or wastes from these operations).

10 CFR Part 40—“Domestic Licensing of Source Material”, Appendix A to Part 40 (Criteria Relating to the Operation of Uranium Mills and the Disposition of Tailings or Wastes Produced by the Extraction or Con-centration of Source Material from Ores Processed Primarily for their Source Material Content).

40 CFR Part 192—“Health and Environmental Protection Standards for Uranium and Thorium Mill Tail-ings”, Subpart D (Standards for management of uranium byproduct materials pursuant to section 84 of the Atomic Energy Act of 1954, as amended).

40 CFR Part 264—“Standards for Owners and Operators of Hazardous Waste Treatment, Storage, and Dis-posal Facilities”, Subpart K (Surface Impoundments).

Finch, W.I. 2003. Uranium—Fuel for nuclear energy 2002. U.S. Geological Survey Bulletin 2179-A, 18 p.Giroud, J.P. & Bonaparte, R. 1989. Leakage through liners constructed with geomembranes—Part I.

Geomembrane Liners. Geotextiles and Geomembranes 8: 27–67.Giroud, J.P., Badu-Tweneboah, K. & Soderman, K.L. 1997. Comparison of leachate flow through compacted

clay liners and geosynthetic clay liners in landfill liner systems. Geosynthetics International 4(3–4): 391–431.

Weakly, L.A. 2008. Renewables: an inconvenient energy reality. Mining Engineering 60(4): 18–28.



315


Tailings structure closure for economic development in Ghana

M.B. ThorpeGolden Star Resources, Littleton, CO, USA

F. NyameGolden Star (Bogoso/Prestea) Limited, Bogoso, Ghana

B.A. AddoAbosso Goldfields Limited, Damang, Ghana

ABSTRACT: Economic development in Africa is often associated with mine development and options for continued economic success are required once a mine closure (or facility closure) decision is made. The closure of the tailings disposal facilities offers the greatest potential benefit to local economic development in the Western Region of Ghana. Through the involvement of our Golden Star Oil Palm Plantation (GSOPP), we will build on the successes achieved in other areas of Ghana, more specifically at the Damang mine, to develop an oil palm plantation on the decom-missioned tailings disposal facility. GSOPP will provide the expertise and management required for the success of the project and through a smallholder program will result in employment for about 16 family units (supporting about 200 people in total) while the mine is still operating. By building on the base skills of the local people, we will be able to turn a potential liability into a community economic asset and contribute to the long-term economic stability to a region of the world that suffers from high levels of poverty.

1 INTRODUCTION

Economic diversity within the Western Region of Ghana is very limited. The major source of income is subsistence farming. There is some cash crop farming of cocoa and oil palm and exten-sive oil palm plantations are present in some areas. Economic diversity has mainly originated because of mining and mine development that, in some cases, dates back to the 1890’s. Patterns of economic and population development are associated with ports and mining areas with very few larger settlements in other areas. There has been a population move into the area as farming offers better opportunities in the Western region, which is mostly tropical rainforest, than in the other regions of Ghana.

Mining in the Western Region centers on Tarkwa, which is a large town about 100 km north of the coastal city of Takoradi that has five operating mines with 40 km of its centre, including the two Golden Star Resources operations at Bogoso and Wassa. The economy in Tarkwa is primarily driven by mining and its associated support activity and by government as it is the district capital. Further to the north and west is the Golden Star Bogoso/Prestea operation. Our operation has an extensive strike length (80 km) and covers ground that has been part of mining operations since the 1890’s. The town of Prestea is currently stagnating due to the shutdown of the underground operation that formerly employed over 1500 people directly.

2 GOLDEN STAR (BOGOSO/PRESTEA) LIMITED

The Bogoso/Prestea mine consists of a series of open pits with two processing plants and an under-ground mine located in Prestea that is currently under care and maintenance. Production during


316

2007 was mostly from the Pampe, Buesichem, and Chujah pits with some additional oxide ore from some smaller satellite pits.

The Bogoso/Prestea mining fleet consists of 31 Caterpillar 777 (100 tonne) haul trucks assisted by a variety of heavy equipment including 5 Liebherr 994 excavators, 5 graders and 11 dozers. Mining is carried out using a normal drill, blast, load and haul method with ore transported to the run of mine (ROM) pad where it is fed into the comminution section of the processing plant.

The Pampe and Buesichem open pits are 26 km and 8 km, respectively, from the Bogoso pro-cessing facility. Therefore, contractors with road haulage fleets are used to transport ore from these pits to the Bogoso processing facility, mostly along our private haul roads. The Bogoso/Prestea Mining Department works closely with the haulage contractors to ensure that all safety and environmental policies are implemented.

With the commissioning of the sulfide processing plant, the focus at Bogoso/Prestea has been on optimization. The oxide processing facility was modified to allow it to receive either oxide or sulfide ores, on a campaign basis. The oxide product from the mills are fed directly to the carbon-in-leach (CIL) tanks and sulfide material from the mills is sent to the flotation section to produce a concentrate for the BIOX® section of the sulfide processing plant.

In 2007, 2006 and 2005, the royalty rate for Bogoso/Prestea was 3% of revenues, therefore, royalties of $2.6 million, $1.9 million, and $1.8 million, respectively, were paid to the Government of Ghana.

2.1 Community engagement and support

Our commitment to fund the Golden Star Community Development Fund is $1 per ounce of gold sold and 0.5% of pre-tax profits. This is just one of the ways we share the success of the Company

Figure 1. Location of the Bogoso/Prestea operations in the Western Region of Ghana.


317

with our local communities (Golden Star Resources 2008). Our Development Foundation supports projects that are recommended by community committees. Each of our local stakeholder commu-nities is able to select infrastructure and other community projects that are then reviewed by the community committee. This committee then selects the projects that will be recommended for funding by the development foundation. This section of our corporate social responsibility mostly addresses infrastructure needs and we are currently building two schools and a police station. Other projects that have been recommended for funding include a nurse’s station and a scholarship scheme. The community mine consultative committees meet each quarter and are able to request funding for projects that reflect the changing needs of the community. Additional support for community water is provided directly by the operation and in the past, Bogoso/Prestea has built schools, libraries, toilet facilities and numerous community water supplies.

In addition to our Development Foundation, we made a substantial investment in Golden Star Oil Palm Plantations (GSOPP), which has the potential to support 3,000 family groups when fully developed. This is the cornerstone of our economic diversification and builds on the existing level of knowledge within the local communities (Golden Star Resources 2008). The economic realities of life in the small towns and villages in the Western region of Ghana are such that many children attend school only a limited extent and then have to work on farms or assisting with the family life. This is particularly true with girls, who are often pulled out of school to look after younger siblings.

2.2 Rehabilitation

Rehabilitation at GSR is an integral part of our ongoing operations with open pits and waste dumps being rehabilitated when they are no longer required for operations. The Environmental Departments at the mines operate nurseries to provide trees for rehabilitation and grass for slope stabilization.

Rehabilitation work at GSR mines includes capping hard rock waste dumps with oxide materi-als to provide subsoil, adding available topsoil as a growth medium, planting with legume species (trees and forbs) to build nutrients and then ongoing maintenance work. This process results in a developing an ecosystem that will develop its own soils capable of supporting a more complex and diverse ecosystem.

The use of Acacia species provides slope stabilization, good soil development with the high production of leaves, sequestering of carbon in the wood and the opportunity to build biomass, which is the key to ecosystem development in the tropics. The trees can be used for firewood locally but provide little vertical biodiversity until soil conditions improve so that other pioneer forest species can become established in the developing ecosystem.

The resulting ecosystem remains in a state of flux and will eventually change to a more mature tropical system with time. Key drivers include the availability of seed (proximity of other forest patches; connectivity with these patches of forest to allow animals to migrate into the new ecosys-tem; and human intervention, either through direct intervention and tree planting or through forest change by developing farmland.

3 ECONOMIC DEVELOPMENT

In Ghana, there have been, and continue to be a host of economic diversity initiatives including raising grasscutters (a large rodent) for food, growing giant land snails, soap making, pastry mak-ing, poultry farming etc. Additionally, some operations are providing training in trade-type skills but are not generally considering the need for the specific trade skill, or where people will be able to work.

Economic development efforts with a view to the future was formulated in 1964 with the 7-year development plan (Vordzorgbe & Caiquo 2001) and have continued with various initia-tives through to the Vision-2020, which was started in 1991 with a view to providing development


318

through 2020. Additional development work is being carried out as part of the Millennium Development Goals. This 5-year $547 M program is focused on private sector-led agribusiness development with the aim of increasing production and productivity of cash and food staple crops (Millennium Development Authority 2006).

Many of the development programs in Ghana are sustained by the program itself. For example, the grasscutter program promotes the development and expansion of grasscutter ranching and farmers are selling breeding pairs of animals at a price above the market price for the meat. Once the build-out of the production is completed and no additional breeding grasscutters are required, the animals have to be sold for meat with greatly reduced profits. Examples of this are found in many countries, a key example of which is llama production in North America, which is currently experiencing the crash (Ingram & Krowka 2002).

3.1 Golden Star economic development efforts

As part of our corporate social responsibility, Golden Star initiated a series of alternative liveli-hood programs and vocational training, which was developed in partnership with local councilors, local community entrepreneurs, and trainers. Committees selected 48 people to participate in the various training programs. Participants were able to select among batik fabric printing and tie dying, soap and cosmetic cream making, and bread and pastry baking. Graduates formed collec-tive associations that, with the help of the District’s Cooperative Officer, became officially regis-tered cooperatives. After a year, 37 graduates (nearly 80%) are self-employed, operating thriving businesses, supporting their families, saving money, and planning for their futures. Other outreach and development programs and the associated expenditures are presented in Table 1.

Currently, in addition to the GSOPP and the Development Foundation, we have four projects running: Oil Palm out-grower scheme; Poultry Farming; Beposo Fish Farm and Demonstration Farm and Training Centre.

4 TAILINGS DISPOSAL FACILITY CLOSURE

At the Goldfields Damang mine, there was a need to close a tailings disposal facility and initial community consultations indicated that there was a lot of concern over the long-term liability associated with the facility. Additionally, the local traditional leaders did not want to accept the land back due to the potential risk to their people.

4.1 Selection of end land use

Field trials were started as the facility was closed to determine the options for closure. At this point, it was decided that the final closure plan for the facility must contribute, at least in part, to economic development within the existing traditional area.

Table 1. Sustainable alternative livelihood programs developed by Golden Star Resources.

Program # People Costs (US$) From–To

Poultry-Broilers 24 25,569 2005–06Poultry-Layers 18 24,320 2006–07Oil Palm Out-Growers 319 414,431 2001–07Vocational Training

Pilot Project 48 12,198 2005–07Sericulture 25 5,500 2003–06Pit Conversion to

Fish Pond (trial) 0 5,100 2003–06

Total 434 487,118 2001–07


319

Initial field trials of a variety of species planted into small holes filled with topsoil proved to be successful. Key drivers to success were:

− Availability of symbiotic bacteria for legumes.− Surface temperature.− Time of planting.− Tolerance of sunlight/heat.

With the initial success of the field trials, an evaluation of the possible economic end use for the tailings disposal facility was completed. This included a review of the cash crops in the area, an evaluation of the types of soil required and the long-term viability of the crop based on local agri-cultural skills. Crops considered included cocoa, plantain, coconut, and oil palm, all of which have existing markets capable of expansion to use any additional production. The environmental condi-tions within the tailings disposal facility included a high clay/silt content (due to the grind size) and a relatively high water table as a small pond was maintained in the centre of the facility for sediment control before water passed over the spillway. These environmental conditions resulted in the selection of oil palm as the preferred crop to contribute to the economic development in the area and so an oil palm plantation was selected as the end land use for the tailings disposal facility. Due to downstream environmental concerns and the need for some diversified habitat, the closure scenario mixed oil palm with other general species and a wetland for sediment control and wildlife habitat.

4.2 Initial plantings

Work on the surface of the tailings facility included stabilization with grass and legumes to reduce wind and water erosion, planting of various tree species for habitat diversity and to reduce the phreatic level in the tailings disposal facility adjacent to the embankments and establishment of rushes and reeds adjacent to the pond for increased habitat diversity. Additionally, the spillway was planted for stabilization with a variety of grasses and ground-covering legumes. A trial area of cocoa was initially established in the tailings. However, the plants did not survive, probably due to the high level of sunlight not only directly onto the plants but also reflected from the white of

Photo 1. Aerial view of the Damang tailings disposal facility showing the oil palm plantation.


320

the tailings surface. The initial establishment of the vegetation at the site was then inspected by the local traditional leaders and other key stakeholders. However, there was skepticism as to the possible success of the project.

The oil palm seedlings were obtained from one of the local commercial oil palm plantation nurseries and were planted into holes (0.5 m diameter × 0.5 m deep) filled with topsoil as there are no nutrients in the tailings. After establishing the first several hundred oil palms plants, effort was focused on reducing the albedo of tailings surface so inter-planting of the oil palms with Pueraria (as species of bean) was carried out, as shown in Photo 2.

4.3 Ongoing maintenance and development

The lack of nutrients and organic matter in the tailings mean that ongoing maintenance of the plantation is more critical than under normal oil palm plantation models. The development of the plants was assisted by fertilizer applications throughout the year. Up to 100 kg/ha were applied to the area beneath the fronds as this is typically where the roots are located. The development of the oil palm was so good that a small amount of fruit was produced after 2 years with most of the trees bearing fruit after 3 years. Harvesting of the fresh fruit bunches was carried out after four years. After 6 years, maintenance is required on the plantation but fruit production is now sufficient to pay for all the inputs.

Ongoing reviews of the plantation are carried out by the Environmental Department and har-vests from the period July 2007 to May 2008 are now totaling 70.7 tonnes for an area of 25 ha (Photo 3). The harvest is not the full complement of the plantation’s capacity as harvesting and maintenance are not regular/strict and it is accessible to local communities. Proceeds from the sale of these fresh fruit bunches are used in funding some community assistance and projects. Field

Photo 2. Initial planting at the Damang tailings disposal facility.


321

(open) days are continuing with the main discussions among the traditional/opinion leaders being how to improve on methods used in rehabilitating the land for farming and other purposes. The Damang mine continues to operate its tailings oil palm plantation as an ongoing demonstration farm. The Environmental Protection Agency (EPA) has not signed it off to be handed over to the government, who will then decide who takes over the rehabilitated area. The success of the model developed at the Damang mine will be applied at the Bogoso/Prestea operation but using a dif-ferent approach for the long-term integration into the local economy. This will be achieved by integrating the tailings disposal facilities into the GSOPP plantation.

5 BOGOSO/PRESTEA ECONOMIC DEVELOPMENT MODEL

The Bogoso/Prestea mine tailings disposal facility differs from that at the Damang mine as com-pensation was paid for the crops in the area of the tailings disposal facility before it was built and for some of the area, a payment was made to the traditional authority for the purchase of the land. Therefore, some sections of the tailing disposal facility are owned directly by Bogoso/Prestea. This provides some variation in options for the closure of the disposal facility and its integration into the GSOPP program.

5.1 Golden Star oil palm plantation limited

Agriculture practices in the Western Region of Ghana are based largely on slash and burn with some longer-term cash crops such as cocoa and oil palm. Concerns have been raised about the effect of large oil palm plantations on biodiversity. However, where Golden Star is working, most

Photo 3. Oil palm plantation and some of the fresh fruit bunches at the Damang tailings disposal facility.


322

Photo 4. Local workers weeding the Golden Star Oil Palm Plantation at Bogoso.

Photo 5. Layout of the tailings disposal facilities at Bogoso/Prestea showing the oil palm plantation.

of the vegetation cover is small farms, abandoned farms, or secondary forest. Therefore, the oil palm plantations planned by Golden Star are not removing primary forest.

The current model for GSOPP is that the local traditional authority is approached to provide access to lands for the development of the oil palm plantation. These lands are normally a con-tinuous block that allows the plantation to be developed taking into consideration the protection of watercourses and high value forest. Planting is not carried out within 25 m of watercourses and the natural (or remnant) riparian vegetation is left intact. Similarly, high value areas of


323

forest are secured while the plantation develops around them. Plantation development occurs as follows:

− Agreement signed with the traditional authority.− Area cleared of brush and existing vegetation.− Infrastructure marked out (e.g. roads).− Palm plantation layouts pegged.− Palm seedlings planted.

Following the initial plantation development, there is a period of between 3 and 4 years of care and maintenance, which includes weed and other pest control, fertilizer application and general plantation development. The plantation is then marked out into 4 ha blocks and these are allocated to family groups. The family groups are responsible for maintaining the oil palm plantation and will receive benefits from the sale of the fruit. From the sales, 5% is directed to the traditional authority and 30% goes to GSOPP for administration and support.

5.2 Tailings disposal facility closure

The Bogoso/Prestea tailings disposal facilities consist of three distinct areas: decommissioned #I facility; TSF II for CN tailings; and TSF III for sulphide flotation and neutralized tailings. The areas of the tailings disposal facilities total 230 ha plus 69 ha in the decommissioned TSF I.

The tailings disposal facilities fit well into the overall GSOPP expansion program as they are immediately adjacent to one of the larger (276 ha) plantations. The closure plan will be phased to address operational needs with the currently decommissioned TSF I being the first mine clo-sure area to be integrated into the oil palm plantation. This facility is still being used as part of the operational water control system. However, water levels are controlled such that planting along the margins of the facility will be completed during the rainy season of 2008.

The rehabilitation of the tailings disposal facilities will be staged based on the mine life and the requirement for the disposal of tailings. The initial work will establish about 64 ha of oil palm plantation over a 2-year period from 2008 to 2009. This will be maintained throughout the first 4 years by GSOPP using local labour to care for the plants and carry out the required maintenance. Due to the nature of the tailings, the rehabilitation will require additional fertilizer applications above the anticipated 100–200 kg/ha for a developing oil palm plantation following the initial low inputs in years 1–3 (Fairhurst & Mutert 1999).

The build out and integration of the decommissioned TSF I offers Golden Star (Bogoso/Prestea) Limited the opportunity to prove that the predicted closure scenario will achieve the success obtained by the Damang Mine in the closure of their facility.

6 CONCLUSIONS

The economic development opportunities and efforts in the Western Region of Ghana need to be focused and adapted to the limitations of the area—poor infrastructure, limited education within the population, and the need to build on existing capacities within the resident population and the existing knowledge base. The initial aim of contributing to economic development in the area was based on the need for a market for the products and skills generated. As there is an pervasive farming culture around both the Damang and Bogoso/Prestea operations, developing additional agro-based economic activity was deemed to have the greatest potential for success. Key factors in the selection of oil palm were the existence of small plantations locally, the proximity of large plantations that required additional fresh fruit bunches, and a willingness within the local popula-tion and the broader stakeholder community to embrace such closure initiatives.

Challenges exist for the ongoing establishment of an oil palm plantation on the Bogoso/Prestea tailings disposal facilities. However, using the knowledge based gained as a result of the Damang mine closure and sound agro-forestry practices brought to the project from GSOPP increase


324

the potential for success. By integrating a closure plan with an existing economic development initiative, Golden Star Resources is able to contribute economically to the local economy while reducing its own liability for the retirement of its assets.

REFERENCES

Fairhurst, T.H. & Mutert, E. 1999. Introduction to Oil Palm Production. Better Crops International Vol. 13, No. 1.

Golden Star Resources. 2008. Annual Sustainability Report.Ingram, G. & Krowka, J. 2002. The straight scoop on breeding llamas.Millennium Development Authority. 2006. Ghana Compact Summary.Vordzorgbe, S.D. & Caiquo, B. 2001. OECD/DAC Dialogues with Developing Countries on National Strate-

gies for Sustainable Development, Report on status review in Ghana, Draft Working Document.


325


Lessons learned from tabletop reviews of emergency action plans for high hazard dams in West Virginia, USA

J.D. QuarantaDepartment of Civil and Environmental Engineering, West Virginia University, Morgantown, WV, USA

H.M. Childers & P. MylesWheeling Jesuit University, Wheeling, WV, USA

ABSTRACT: This paper presents two case histories on the application of the US Federal Emer-gency Management Agency (FEMA) tabletop exercise process on Emergency Action Plans (EAP) as applied to coal waste impoundments in West Virginia. Coal waste dams are important com-ponents in the processing of mined coal and are used for storing the coarse and fine coal slurry generated as a result of material washing and preparation. The FEMA established the follow-ing three classification levels for dams as: Low, Significant, and High hazard potential. Dams assigned the high hazard potential classification are those where failure will probably cause loss of human life.

The Coal Impoundment Project performed tabletop exercises in collaboration with two coal operators; county emergency management agencies; responders including local law enforcement and fire departments; and the WVDEP and MSHA regulatory agencies. The exercises were devel-oped and facilitated using two different formats: One format involved a role-playing approach where participants received hand-delivered situation update messages. The second format was structured by scenarios and discussion questions within a project developed Situation Manual. The incident scenario for both exercises was based on a “sunny day” event failure. Both exercises were designed for participants to go through the four tiers of awareness in the mine specific Emergency Action Plan (normal conditions, adverse conditions, standby alert, and evacuation conditions). Results of the exercises are presented and discussed.

1 INTRODUCTION

West Virginia, with its mountainous terrain and historical coal mining production, has a legacy of pre-law (Surface Mining Control and Reclamation Act of 1977) and post-law coal waste impound-ment sites. These sites were historically used as impounding structures for coal slurry, process black water, and coarse refuse disposal. These structures are regulated by the US Department of Labor—Mine Safety and Health Administration (MSHA) and by the West Virginia Department of Environmental Protection (DEP). In West Virginia there are over 135 Coal Waste Impound-ments having a Hazardous rating; indicating that failure of these structures could cause significant human and property loss.

Following the breakthrough and release of coal slurry from the Martin County Coal Corporation impoundment near Inez, Kentucky on October 11, 2000 the United States Congress requested the National Research Council (NRC) to examine ways to reduce these types of accidents. The NRC completed their study which identified numerous areas of concern and the committee presented recommendations for improving the design, operation, and safety of coal waste impoundments (NRC, 2002). In 2003 the Coal Impoundment Project began as a program of the National Technol-ogy Transfer Center (NTTC) at Wheeling Jesuit University (WJU) (Quaranta, et al. 2004).


326

The need for emergency planning and response review of critical facilities due to natural disasters is important to maintaining public safety and limiting property and environmental dam-age. In 2005, the hurricanes of Rita and Katrina caused death and destruction to the United States on a massive scale. Numerous emergency response breakdowns occurred during hurricane Katrina and resulted in an increase in human suffering. The lack of community knowledge and involve-ment in the Emergency Action Plan (EAP) process was identified in part though a risk assessment survey (McSpirit, 2005), through public meetings throughout West Virginia, and in light of the emergency response failures of the federal and local emergency responders in the Gulf States, September 2005.

The goal of this project was to support the safety of communities by developing and performing tabletop reviews of select coal waste impoundments in West Virginia. The project involved bring-ing together various federal and state agencies as well as select community emergency manage-ment agencies to participate in this pilot program. The FEMA has published a general guidance manual with information on how to develop and execute a tabletop exercise. This project used the FEMA training courses as a guide and then formed a working group to further develop real-life in-cident scenarios to tailor a program which could be offered to West Virginia coal companies as a useful and practical guide for performing tabletop reviews at their sites. This initial pilot program focused on developing and performing a tabletop reviews with two coal companies having coal waste impoundments with inundation maps affecting Ohio, Marshall, Boone, and Lincoln counties of West Virginia.

The NTTC, Center for Educational Technologies (CET), and West Virginia University collabo-rated with two coal companies to hold tabletop review exercises of their current EAP. These exer-cises involved the participation of various community emergency response agencies, and included both the West Virginia Department of Environmental Protection and the US Mine Safety and Health Administration.

2 BACKGROUND ON EMERGENCY ACTION PLAN DOCUMENTS

The West Virginia Department of Environmental Protection, Division of Water and Waste Management—Dam Safety Section requires that dam owners submit for approval and implemen-tation a Monitoring and Emergency Action Plan in accordance with the West Virginia Dam Con-trol and Safety Act (WVDEP, 2006). The DEP organized the EAP documents into the following four sections:

Part 1: Monitoring Plan and Inspection ScheduleA. Normal ConditionsB. Adverse ConditionsC. Standby AlertD. Evacuation Conditions.

Part 2: Emergency Action and Evacuation PlanA. Notification of AgenciesB. Evacuation Notification of Downstream PersonsC. Evacuation Map.

Part 3: Post-Evacuation Notification ProcedureA. No failure of damB. Failure of dam.

Part 4: Administrative and Record KeepingA. Qualified Persons for Monitoring and Inspecting the ImpoundmentB. Signature and Distribution ListC. Inspection Record.


327

In practice the first three parts of the plan would function sequentially. Advancing actions in Part 1 lead to activities performed in Parts 2 and 3.

There is presently no requirement by the WV DEP or any county government requiring a mock disaster drill or tabletop exercise of an EAP. However, EAP documents are required to be reviewed and approved annually. The WV DEP also may conduct on-site reviews of the EAP procedures with mine owners on case-by-case situations. The WV DEP requires that the approved EAP be distrib-uted by the mine owner to offices of emergency services, county sheriffs, and state regulators.

In December 2007, MSHA released a draft version of the updated Coal Refuse Disposal Engineering Manual recognizing the importance of emergency action plans and has dedicated Chapter 14—Emergency Action Plans to this topic (MSHA, 2007). Currently MSHA does not require EAP documents for approval of permits (MSHA, 2008).

3 OBJECTIVES AND GOALS

This project’s objectives were to: provide an exercise to bring together participants who would respond in a real emergency at a coal waste impoundment in West Virginia; familiarize the stake-holders with the anticipated response process; and demonstrate the effectiveness of the stakehold-ers to interact in the solution of the problem scenario addressing the emergency response and communication stages of an Emergency Action Plan. The exercises would be evaluated and feed-back used to assess the effectiveness and the current format of the WVDEP-EAP document.

4 METHOD

The tabletop exercise process was developed and performed using two different formats. Both formats were prepared based on the Federal Emergency Management Agency training course for conducting tabletop exercises (http://training.fema.gov/EMIWeb/IS/is139lst.asp). Planning for these exercises was initially organized to include workgroups of developers, reviewers, evalua-tors, facilitators, and participants. The developers met first to prepare an incident scenario based on a FEMA “sunny day” event failure. A team of reviewers and evaluators prepared measures of monitoring the exercise to score evaluations for discussion. The participants were invited as listed within the coal company EAP. Members of the public community at large were not invited.

There were two different tabletop exercises performed on the project, the first occurred in August 2006 and the second followed in June 2008. Each of these exercises is discussed in this paper.

4.1 Exercise 1 August 2006

The first exercise followed a format which incorporated a hypothetical emergency at a coal waste impoundment. The participants were segregated into different rooms and brought into the exercise using hand written messages as the EAP advanced with notifications depending on the partici-pant affiliation and anticipated sequence of notification. This version of the exercise involved a role-playing approach with un-rehearsed response actions from the participant(s). A facilitator led the exercise and paced the communication events thru the EAP stages (normal conditions, adverse conditions, standby alert, and evacuation conditions). Exercise evaluators used criteria to rank the exercise for subsequent post meeting discussions and evaluations.

The exercise began with the facilitator reading of a short narrative which set the stage for the hypothetical emergency. The facilitator next stimulated discussion by introducing various event scenarios. These scenarios described detailed or major events which were addressed either to individual participants or to participating departments or agencies. Each statement was prepared in advance and the facilitator delivered the information as participants determined their response actions. The pre-scripted messages were written with the intent to trigger the expected actions


328

and to involve all participants in the exercise. The facilitator adjusted the timing of information for more complex situations depending upon how the participants responded. Specifically, the exercise simulated a series of initiating events which could lead up to a critical point and then events at the impoundment stabilized. The scenario did not advance to a failure of the coal waste impoundment (CIP, 2006).

Discussion generated by the scenario focused on roles (how the participants would respond in a real emergency), plans, coordination, the effect of decisions on other organizations, and similar concerns. Participating organizations were able to reference emergency preparedness plans as well as the coal impoundment EAP. Maps, charts, and packets of materials were added for reference to enhance the realism of the exercise. The exercise ended when the scenario described the initiating seepage event had stopped due to corrective actions by the mine operator. No evacuation release was required to be issued. The Evaluator’s and Participant’s comments were recorded during the debriefing and were included in the after action report (CIP, 2006).

4.2 Exercise 1 evaluation

Developers found the evaluation team to be the most difficult team to find volunteers. Comments included that for future exercises, recruitment for evaluators should happen during the initial plan-ning stage. Travel stipends or other compensation might be necessary to field a team of experi-enced evaluators. Ideally, the evaluation would include both a performance-based evaluation and a standards-based evaluation.

The exercise evaluation plan included having experienced professionals in safety and emer-gency management observe the exercise to access whether or not the exercise activities met the pre-defined objectives. Four evaluators were planned—two from the International Union of Oper-ating Engineers National Hazmat Training Program and two from the Belmont County Emergency Management Agency. Developers asked the two evaluators from the National Hazmat Training Program to focus on the first two objectives and asked the emergency management agency to focus on the second two objectives. On the day of the exercise, the two evaluators from Belmont County were unavailable because of an actual emergency. The objectives and the evaluator assess-ments of whether the exercise met the objectives for Exercise 1 were:

Objective 1—During standby alert conditions, verify that the person responsible places the coal impoundment under constant surveillance and notifies the proper agencies according to the coal impoundment’s emergency action plan.

Evaluator notes: “This objective, in my opinion, was met. From my observation of the company and WVDEP participants, they appeared to know specifically what response was needed based on the problem at hand. I think if we were to observe them in a real crisis, their process would be sec-ond nature. For the purpose of this exercise, it would have been nice to have company representa-tives, a technical, with a management person. I know that it is tough to get people to participate, but if the Consol guy has oversight of the impoundment then the right person was there. From a process standpoint, he seemed to know what needed to be done. Also, I didn’t have time to really look the company plan during this activity, but what they did appeared to be reasonable. I have to assume that they pretty well knew what was in there, because they dealt with it well. I focused on the company and WVDEP command center, rather than the emergency responders” (CIP, 2006).

Objective 2—To determine whether or not there is a consensus among stakeholder participants as to the conditions for standby alert.

Evaluator notes: “This objective was also met. From a process standpoint, the command center worked well to reason out the problem, react to changes in the conditions, and then to proceed. It gives me hope that it appears that there is a good working relationship among these folks in real life, that they would handle problems reasonably. Everybody in this group knew their jobs, and they seem to work well in that environment” (CIP, 2006).

Objective 3—During evacuation conditions, verify that the person responsible notifies the proper offices during an evacuation notice according to the coal impoundment’s emergency warn-ing plan and the county’s all-hazards emergency action plan.


329

The evaluator for this objective was unavailable due to an actual emergency.Objective 4—To determine whether or not there is a consensus among stakeholder participants

as to the conditions for an evacuation notice.The evaluator for this objective was unavailable due to an actual emergency.Examples of additional evaluator comments from the final report follow (CIP, 2006):

• “More notification advance is needed in order to have the response plans for all the key players

• I like the movement of all participants outside the discussion area until needed. It could be bor-ing for the ones that don’t participate early in the exercise but that are what they bought on to do. We need to be careful that we don’t short change the deliberation time to accommodate waiting participants.

• I think the idea of some common timeline on the screen would be good for everybody to see. Also the computerized view you had for this exercise was great!

• The exercise went smoothly, but I think that we need to review and make sure that sufficient time is being allotted for analysis and reaction for each of the changes in conditions”.

4.3 Discussion of Exercise 1 after-action report

In the after-action report (CIP, 2006), a review of the evaluator and participant comments was per-formed and the results tended to indicate that this form of blending mock notifications distracted the exercise away from the core content of testing the function of the EAP. What apparently tended to occur was that the exercise emphasis shifted into control by the emergency responder’s and away from discussions of the Emergency Action Plan process. The exercise was intended as a half-day event; however, the meeting discussions forced the meeting into a full day event.

The project developers identified that many of the emergency response personnel were not knowledgeable about the EAP or what a coal waste impoundment was, the facility complexities, physical size and shape, and the significant impact a dam failure could produce. After the first initial minutes of the exercise, the EAP was not referenced by the first responders to determine what the process was and who / how communication was being established and performed.

The developers identified the need to produce a handbook for use by the coal industry, regu-lators, and emergency responders when testing their EAP. The handbook was intended to focus conversation onto the EAP, specifically the Part 1 activities occurring (normal conditions, adverse conditions, standby alert, and evacuation conditions). The exercise needed to be limited to one-half a day, and the participants needed a face-to-face layout where a conversation driven table-top review could be performed. A site visit to the impoundment would have been worthwhile to expose participants to the magnitude of the potential problems.

To accomplish this effort, two manuals were prepared: a Situation manual for use by par-ticipants and a Facilitator/ Evaluator manual for use by the exercise evaluators and facilitators. This new form of the tabletop exercise focused only on the EAP monitoring plan main sections, specifically: Part 1—Monitoring Plan and Inspection Schedule, Part 2—Emergency Action and Evacuation Plan, Part 3—Post-Evaluation Notification, and Part 4—Administrative and Record Keeping.

5 SITUATION/EVALUATOR MANUAL

Figure 1 is a snapshot of a page form the manual. The exercise layout starts with presentation of the time/date and initial condition at the impoundment; the location and relevant facility information are presented to give the participant background on the site. Discussion questions then follow the description narrative. The questions establish the baseline for discussion such as what the impoundment inspection practice is under normal conditions. During the course of the exercise the questions become specific to the stage of the pending emergency event such as


330

Figure 1. Situation manual example.

when the conditions change to Adverse Conditions, then to Standby Alert, and finally Evacuation Conditions. Figure 2 is a snapshot of an Evaluator Checklist form used to assess and measure aspects of the objectives.

5.1 Exercise 2 June 2008

The second tabletop exercise was performed in June 2008 and followed a different format from the first August 2006 exercise. This second exercise limited participants to: the coal company,


331

WV DEP, MSHA (District 4 and Pittsburgh Technical Center), and select emergency response agencies (Boone County Emergency Services and the county Sheriff’s office), the participants totaled twenty-four.

This exercise was hosted at the mine company’s training center and began with opening com-ments from the mine owner and review of safety training needed for site training and impoundment access. Following a field tour of the impoundment facility and discussion of the site operations the

Figure 2. Evaluator checklist.


332

participants returned to the training center to begin the exercise. The exercise followed the format outlined in Figure 3 and had the following four key scenario events:

1. 10 a.m., July 15, 2008—Suspicious SeepageOn this beautiful sunny day, during the placement of coarse refuse, the dozer operator observes

seepage around the outside of the decant pipe. The dozer operator radios the shift foreman that there appears to be seepage developing around the outside surface of the decant pipe. The shift foreman subsequently notifies the mine superintendent. They meet at the decant pipe and decide to start monitoring the flow of the seepage.

2. 8 a.m., July 16, 2008—Black MorningThe flow of the seepage has increased slightly, and the water has turned dark with fine coal

particulate and erosion developing around the decant pipe. Suspended solids have increased, and the decant ponds are now black with suspended coal fines. Decant pumps have now started to transfer black water into clarifier and treatment ponds. The company decides to contact its consulting engineer, and in meetings with the engineer it develops a new monitoring plan.

3. 4 p.m., July 16, 2008—From Black to WorseA whirlpool appears to be developing within the pond. Flow rate and suspended solids increase

drastically at the seepage area. A visible sinkhole on the downstream face develops. Everything gets worse in all of the data messages—flow increases, suspended solids increases, sinkhole is observed. A link on the track on the company’s dozer being used to move coarse refuse breaks, and the dozer is inoperable. The company does not have another D9 or D10 dozer immediately available.

The black water has broken through the settling/treatment ponds and has started to flow into the stream. People in the community are noticing the black water, and traffic traveling to the impoundment increases. People call 911 asking questions. The media are contacted and begin to show up on site.

4. 9 p.m., July 16, 2008—Disaster AvertedSlurry flow slows. There has been a large reduction in the seepage flow rate. Injury and

damage to personal property have been avoided. However, slurry and black water have been released into Sunny Day Branch.

5.2 Exercise 2 evaluation

The objectives and the evaluator assessments of whether the exercise met the objectives for Exer-cise 2 are the following (CIP, 2008):

Objective 1—Introduce participants to the complexities of the coal refuse impoundment and review the serious potential hazards associated with a High Hazard Potential impoundment.

Evaluator notes: “The presentation and site visit was good for the participants to get an appre-ciation of the size of the impoundment and the high hazard potential classification was explained. However, the exercise did not thoroughly explain specific site hazards of the coal and water mate-rial impounded”.

Objective 2—Examine the Emergency Action Plan and establish a working knowledge of the four tiers of awareness (normal conditions, adverse conditions, standby alert, and evacuation con-ditions) in an emergency action plan implementation process.

Evaluator notes: “Participants had the latest approved EAP. None of the participants indicated that any of the contact information was incorrect or out of date. Discussions included a review of the EAP sections and highlighted expected actions for normal conditions, adverse conditions, standby alert, and evacuation conditions. Participants were familiar with the notification proce-dures, evacuation procedures, and post evacuation procedures”.

Objective 3—Discuss impoundment monitoring programs and plans under normal conditions.Evaluator notes: “This point was well covered. The person responsible was fully aware of the

checkpoints and scheduled frequency for monitoring and checking the impoundment during nor-mal conditions”.


333

Objective 4—Evaluate mechanisms to determine whether the impoundment is stable or developing into a hazardous condition.

Evaluator notes: “Based on the discussion, these points were complied with. The person respon-sible increased the frequency of monitoring, monitored additional parameters, and compared data to previous conditions and critical thresholds as expected”.

Objective 5—Verify that during standby alert conditions, the person responsible places the coal impoundment under constant surveillance and notifies the proper agencies according to the coal impoundment’s emergency action plan.

Evaluator notes: “The Company notified all agencies as expected. Excellent discussion and food for thought unfolded on how soon to contact the local emergency responders. A highlight was when the superintendent of Patriot Coal asked the Sheriff and emergency manager when they would want to be notified. This demonstrated one of the main points of the whole exercise—the exchange of perspectives”.

Objective 6—Discuss the decision-making process used to determine whether a downstream evacuation is warranted.

Evaluator notes: “There was good discussion about the command post; incident commander; transportation of evacuees; evacuation centers; and notification of utilities. However the exercise didn’t get into the detail level of police security and roadblocks. When participants determined that an evacuation was not warranted, they communicated that to all agencies”.

Objective 7—Examine the effectiveness of internal communication, understanding, and response execution within the coal company at the exercise level.

Evaluator notes: “Company personnel emphasized the need for them to communicate a problem—when the conditions warrant—to higher level of management. When asked what he would do if he saw a problem, one of the “qualified persons” said he’d go get the prep plant superintendent. It was also good that the company indicated early on that they would notify their engineering consultant”.

Objective 8—Examine the effectiveness of communication, understanding, and response exe-cution between and among the coal company and all responding agencies.

Evaluator notes: “I don’t recall much discussion about how the agencies will communicate with one another (i.e. Radios, Internet, emergency warning systems). However, there was some discus-sion about the use of direct personal phone numbers or cell phone numbers rather than the com-pany calling a main number for a Sheriff Office or Office of Emergency Services. Explanations of the incident and possible impacts were clearly communicated to the emergency responders. Inundation maps, evacuation times, and routes were discussed. Discussions addressed resources needed for evacuation and evacuation centers”.

Examples of additional evaluator comments follow:

• “The facilitator opened the exercise by explaining that one of her main jobs as facilitator was “to get everyone here.” My thought was that just getting all of the involved parties in the same room accomplished a great deal, because it created the opportunity for the exchange of information, perspectives, ideas and concerns.

• I thought the exercise was extremely beneficial for all of the parties. While Patriot Coal person-nel and representatives from MSHA and WVDEP are used to working with one another, this was not the case with respect to the Boone County Sheriff’s Office or the Boone County Emer-gency Management Agency.

• The exercise demonstrated the importance of building relationships with local emergency responders prior to a real incident occurring. The Sheriff and Emergency Manager had never before met the Prep Plant Superintendent or the agency representatives.

• The exercise allowed the Sheriff and Emergency Manager to provide insights from their per-spectives during a dam-related emergency. The Emergency Manager, for example, explained that he would like a heads-up as early as possible during the incident so that he could locate personnel—even if it were just on a standby basis.


334

• The Sheriff expressed the same opinion, indicating that, depending on the circumstances, he would need to contact off-duty deputies and possibly enlist deputies from surrounding counties for assistance. This would be especially true if the situation was progressing to the point where an evacuation was ordered.

• The Sheriff expressed the opinion that too often they get their information from the public (e.g., a person living downstream notices a black water discharge). The Sheriff said that a rumor that the dam is leaking “can be the same as a dam break” for its effect on the public.

• The Plant Superintendent asked the Sheriff and Emergency Manager when they would want to be notified. This resulted in a good discussion. The consensus seemed to be that as soon as persons on site recognize that there is a real potential for the situation to deteriorate to the point where the dam may actually fail, the local emergency responders should be notified and advised of the situation.

• The Emergency Manager made an interesting point that he would be concerned with the judg-ment and integrity of the people who contacted him. He would be concerned that they may downplay the seriousness of the situation, so he would want to get someone from his office to visit the site. The issue of the credibility of the persons contacting the local emergency respond-ers is another benefit of the local responders being familiar with key coal company and agency personnel”.

5.3 Discussion of Exercise 2 after-action report

This after-action report identified several key findings, comments and suggestions for improvement.

• “The coal operator gained a new perspective as to how much advance notice the Sheriff’s office needs in order to conduct a successful evacuation and now realizes that the Sheriff and Office of Emergency Services needs to have a “heads up” prior to a possible evacuation situation.

• WV DEP gained an appreciation for the National Incident Management System (NIMS) and will explore the possibility of altering the EAP format so that the EAP’s may be NIMS compli-ant in the future.

• One participant suggested that a similar exercise be done county-by-county so that the Sheriff and Emergency Response Manager for each county would gain the benefit of the exercise.

• Another participant suggested that the scenarios used in the exercise should not be ideal, because in a real situation, “what can go wrong, will go wrong.” They suggested not having the scenario be a “sunny day” when a problem is more likely during a rainy period or when it’s extremely cold”.

This exercise maintained close adherence to discussing the process of the EAP. The exercise was paced to complete on schedule as a partial day event (CIP, 2008).

6 CONCLUSIONS

The feedback results from the participants indicated that both of the Table Top Exercises benefited all organizations. Based on the participant and reviewer’s comments along with achieving indi-vidual task objectives we believe the goals of the project were reached.

Each exercise format identified strengths and weaknesses when implementing the individual company Emergency Action Plan. The Situation Manual (June 2008) exercise appeared to be more conducive to working with the specific stages and details of the EAP. This exercise format also led to more direct interactions at the front-line of the response between the mine owner, emergency management agencies, state, and federal regulatory agency staff.

A drawback of the Situation Manual (June 2008) approach was that the current format was not flexible to permit participants to respond and react to changing events as the exercise developed;


335

specifically the scenarios assumed responses at critical events and did not allow for the participant’s input prior to the action. This did effect the interaction of the participants.

The second exercise identified communication methods and requirements needed for the EAP which the first exercise did not identify. These include that the emergency management and law enforcement agencies require earlier engagement than initially planned for in the EAP and that communication pathways need to be direct to the heads of the agencies in order to minimize communication leaks. Similarly, the EAP format used does not currently conform to the National Incident Management System (NIMS). The NIMS process establishes the structure for incident command and was refined after the Katrina disaster.

The beneficial outcome of identifying NIMS compliance is that this process can be evaluated by the state engineers for incorporation into future EAP format requirements. Further refinement of the Situation Manual format will be performed prior to releasing the document for industry and regulatory use.

ACKNOWLEDGEMENTS

This project was made possible through funding from the National Technology Transfer Center (NTTC) at Wheeling Jesuit University through a grant from the US Mine Safety and Health Admin-istration. The authors wish to thank Mr. J. Davitt McAteer, Program Director and Mr. Joseph Pav-lovich (Exercise 1 facilitator). The authors wish to thank the MSHA Pittsburgh Safety and Health Technology Center—Mine Waste and Geotechnical Engineering Division and the WV DEP for their involvement and support of this work. The authors wish to thank the engineers and staff members from the Consol Energy Corporation and the Eastern Associated Coal Corporation—Patriot Coal Division who participated in this exercise and supported the program.

REFERENCES

McSpirit, Stephanie 2005. Coal Impoundment Risk Assessment: A Survey of Mingo and Wyoming County, West Virginia Households July 2005. Wheeling, WV: National Technology Transfer Center, Wheeling Jesuit University. Retrieved December 6, 2006, from http://www.anthropology.eku.edu/martincounty/Webpage/CoalImpoundmentReport/CIP_TOC2.htm.

MSHA. 2007. Engineering and Design Manual: Coal Refuse Disposal Facilities Advance Draft For Industry Review And Comment. U.S. Department of Labor, 937 pp.

MSHA. 2008. [Title 30 Code of Federal Regulations]. 30 CFR §77.216 Water, sediment, or slurry impound-ments and impounding structures; general. U.S. Department of Labor.

NRC [National Research Council]. 2002 Coal Waste Impoundments: Risks, Responses, and Alternatives. Washington, DC: National Academy Press, 230 pp.

Quaranta, J.D., Gutta, B., Stout, B., McAteer, D. & Ziemkiewicz, P. “Improving the Safety of Coal Slurry Impoundments in West Virginia,” Tailings 2004, Vail CO.

WVDEP. 2006. [West Virginia Department of Environmental Protection] 47CSR34, 15.7, TITLE 47, Legislative Rules, Division of Water and Waste Management, Dam Safety Rules.

CIP [Coal Impoundment Project]. 2006. “Coal Impoundment Tabletop Exercise—August 16, 2006 Final Report.”, Wheeling Jesuit University, Wheeling, WV.

CIP [Coal Impoundment Project]. 2008. “Coal Impoundment Tabletop Exercise—June 2008 Final Report.”, Wheeling Jesuit University, Wheeling, WV.



337


Working for responsible management of tailings facilities

Elizabeth GardinerThe Mining Association of Canada, Ottawa, Ontario, Canada

David GladwinTwo Ducks Resources Inc., Ottawa, Ontario, Canada

The Mining Association of Canada Tailings Working Group

ABSTRACT: The Mining Association of Canada (MAC) has worked over the past 12 years to improve the management of tailings facilities by its member companies and the broader mining industry. This effort began out of a realization that tailings present a major business risk to the mining industry, that the risk needed to be better managed, and that a series of major tailings fail-ures around the world in the 1990’s, while individually related to specific technical issues, were more fundamentally indicative of need for improved care and management practices by tailings dam and facility owners and operators. MAC established and continues to sponsor the Tailings Working Group, which works within the industry to promote safe and environmentally responsible tailings management practices.The most visible thrust of this initiative has been the development and publication of a three-volume set of guides to improve tailings management:

• “A Guide to the Management of Tailings Facilities”, initially published in 1998, updated edition published in 2008;

• “Developing an Operation, Maintenance and Surveillance Manual for Tailings and Water Man-agement Facilities”, published in 2003; and

• “A Guide to Audit and Assessment of Tailings Facility Management”, published in 2008.

The three guides provide a strong message to tailings facility owners, operators and contractors—the key to safe and environmentally responsible management of tailings is consistent applica-tion of sound engineering capability within an effective management framework. The guides also provide a means to achieve the management side of this formula. This presentation will outline the MAC tailings management framework, as put forth in the guides, and provide guidance for its implementation and application as gained from member company experiences.

1 BACKGROUND

In June 1996, The Mining Association of Canada (MAC) Board of Directors established a task force to promote the safe and environmentally responsible management of tailings and mine rock. The task force determined that engineering capability exists and generally is applied throughout the Canadian mining industry in the safe design, construction, operation and closure of tailings facilities. The key to managing tailings is consistent application of that engineering capability within an effective management framework through the full life cycle.

To promote the exchange of information and best practices, the task force arranged two work-shops, one on management of tailings and mine rock (December 1996) and another on tailings risk assessment (May 1997). These workshops and related consultations identified the need for a guide to tailings management.


338

“A Guide to the Management of Tailings Facilities” was initially published in 1998 and updated in 2008. It was developed in a collaborative effort by representatives of the Canadian mining industry, working as the MAC Tailings Working Group, to provide guidance on good practices for the safe and environmentally responsible management of tailings facilities. Its purpose is threefold: to provide information on safe and environmentally responsible management of tailings facilities; to help companies develop tailings management systems that include environmental and safety criteria; and to improve the consistency of application of sound engineering and manage-ment principles to tailings facilities.

The MAC Tailings Working Group has subsequently developed two companion guides:

• Developing an Operation, Maintenance and Surveillance Manual for Tailings and Water Man-agement Facilities, published in 2003; and

• A Guide to Audit and Assessment of Tailings Facility Management, published in 2008.

Together, the three MAC tailings management guides provide a strong and consistent message to tailings facility owners, operators and contractors: the key to safe and environmentally responsible management of tailings is the consistent application of sound engineering capability within an effec-tive management framework and through the full life cycle of a facility. The Guides are not techni-cal manuals; technical guidance may be found in other publications. Nor do the Guides replace professional expertise or regulatory requirements. Mining companies are encouraged to obtain professional and/or expert advice to be sure that each company’s specific needs are addressed.

All three MAC tailings guides are available for download at the MAC website: www.mining.caSince the first edition of the Tailings Management Guide was published in 1998, the tailings

management framework has been applied at mining operations across Canada and around the world. The updated Guide (2008) reflects information and experience gained throughout the course of developing the companion guides and working with tailings management systems around the world.

2 INTRODUCTION TO TAILINGS MANAGEMENT SYSTEMS

Tailings facilities are site-specific complex systems that have unique environmental and physical characteristics. They pose a significant business risk that must be effectively managed for the long term. The mining industry has the technology and resources to safely site, design, construct, oper-ate, decommission and close tailings facilities, but there remains a need to continually improve their management in a consistent, safe and environmentally responsible manner through the full life cycle.

One way to do this is to establish a comprehensive tailings management system, one that inte-grates technical and managerial aspects, and one that individual companies may adapt and imple-ment under often widely ranging conditions. With this approach, the industry can self-regulate, demonstrate due diligence, complement government legislation and regulations, and protect the environment and the public. Perhaps more importantly, such an approach will help companies to integrate environmental and safety considerations in a manner that is consistent with continual improvement in their tailings operations.

The MAC A Guide to the Management of Tailings Facilities provides a basis for the develop-ment of customized tailings management systems that address the specific needs of individual mining companies and local regulatory and community requirements. The Guide presents a tail-ings management framework that offers a foundation for managing tailings in a safe and environ-mentally responsible manner through the full life cycle of a tailings facility from site selection and design, through construction and operation, to eventual decommissioning and closure. As well, it can help companies implement due diligence.

Mining companies and tailings facility owners and operators are encouraged to adapt and extend the principles contained in the tailings management framework to meet their own site, operational and community requirements, incorporating appropriate site-specific performance measures.


339

3 THE TAILINGS MANAGEMENT FRAMEWORK

The tailings management framework, as outlined in the following, is the foundation for a tailings management system. It addresses management requirements with application from initial concep-tual planning through design, construction, operations and eventual closure.

1 Policy and CommitmentEstablish tailings management policies that include commitments to:

• implement the principles outlined in this framework;• locate, design, construct, operate, decommission and close tailings facilities in a manner such

that:

• all structures are stable;• all solids and water are managed within designated areas; and• all aspects of tailings management comply with regulatory requirements and conform with

sound engineering practice, company standards, the MAC TSM Guiding Principles, this tail-ings management framework and commitments to Communities of Interest;

• take responsibility for implementing this framework through the actions of its employees;• consult with Communities of Interest, taking into account their considerations relating to the

tailings facility management; and• establish an ongoing program of review and continual improvement to manage health, safety

and environmental risks associated with tailings facilities.

2 Planning

2.1 Roles and responsibilities Assign overall accountability for tailings management to an executive officer of the company

(CEO or COO), with responsibility for putting in place an appropriate management structure and for providing assurance to the corporation and its Communities of Interest that tailings facilities are managed responsibly.

Assign responsibility and budgetary authority for tailings management. Define the personnel roles, responsibilities and reporting relationships, supported by job

descriptions and organizational charts, to implement the tailings management framework through all stages in the facility life cycle.

2.2 Objectives Plan to manage tailings through the full life cycle in conformance with regulatory require-

ments, company standards, this framework, commitments to Communities of Interest, and sound engineering and environmental practices.

Plan for eventual closure, including:

• protection of public health and safety;• mitigation of negative environmental impacts; and• acceptable post-closure use within a feasible technical and economic framework.

Identify and assess significant environmental, health and safety aspects and their associated risks. Prepare and document tailings facility plans, including descriptions of:

• objectives and performance measures;• permits and approvals;• communication procedures among the team and with management and Communities of

Interest;• site selection and characterization criteria;• safety, environmental and engineering design criteria;• construction, operating, decommissioning and closure procedures;• requirements for documentation, including as-built records;


340

• maintenance, surveillance, inspection, reporting and review requirements; and• knowledge and skills (awareness, training and competence) requirements.

Incorporate Communities of Interest considerations in tailings facility planning.

2.3 Managing for compliance Ensure that:

• applicable legislation, regulations, permits and commitments are identified, documented and understood;

• actions needed to ensure compliance are understood; and• processes and procedures to ensure measurement and compliance have been established,

documented and communicated to all facility employees.

Establish procedures for reporting compliance and non-compliance.

2.4 Managing risk Conduct risk assessment, define acceptable risk in the context of the facility, and identify and

evaluate possible triggers and failure modes. Plan for risk management to:

• minimize the likelihood of adverse safety or environmental impacts; and• detect and respond to potential failures at the facility.

Prepare contingency plans as well as emergency preparedness and response plans.

2.5 Managing change Prepare and document procedures to ensure that the integrity of both the management system

and the approved facility designs and plans is maintained by:

• managing changes in personnel, roles and responsibilities;• managing changes, including temporary changes, made to approved designs and plans; and• responding to changes in regulatory requirements.

2.6 Resources and scheduling For effective and efficient implementation of the tailings management system, including even-

tual decommissioning and closure, identify and secure:

• adequate human and financial resources; and• a schedule.

2.7 Emergency preparedness and response Develop and maintain emergency preparedness and response plans to identify possible acci-

dent or emergency situations, to respond to emergency situations and to prevent and mitigate on- and off-site environmental and safety impacts associated with emergency situations.

Establish procedures for periodic review, testing and distribution of the emergency prepared-ness and response plans within the organization and to potentially affected external parties.

3 Implementing the Plan

3.1 Operational control Assemble a qualified team and assign responsibilities for implementation of the tailings

facility. Select a site, design, construct, operate, decommission and close tailings facilities in

compliance with regulatory requirements and in conformance with the approved plans, appropriate engineering and environmental practices, risk management, the MAC TSM Guiding Principles, commitments to Communities of Interest and this tailings management framework.


341

Identify, evaluate the impact of, and document changes made to approved designs, plans and procedures.

Routinely inspect, monitor, test, record, evaluate and report on key characteristics of the tail-ings facility, including compliance with requirements and commitments.

Implement and periodically test contingency plans and emergency preparedness and response plans.

3.2 Financial control Establish a budget and financial controls, obtain budget approval, and track capital and operat-

ing costs against the budget.

3.3 Documentation Prepare, maintain, periodically review and revise the documents required to design, construct,

operate, decommission and close a tailings facility. Maintain current versions of all documents at designated, readily accessible locations. Promptly remove from use and archive obsolete versions of documents.

3.4 Training, Awareness and Competence Employ qualified personnel. Provide appropriate training to all personnel, including contractors and suppliers, whose

work may significantly affect the tailings facility. Maintain records of all training.

3.5 Communications Implement documented procedures for communications among tailings and related personnel

and with management and Communities of Interest.

4 Checking and Corrective Action

4.1 Checking In addition to routine monitoring and inspections, conduct periodic inspections and reviews of

the tailings facility to:

• evaluate operating and financial performance, compliance with regulatory requirements, and conformance with plans and commitments;

• revisit the facility design, construction, operation and decommissioning and closure plans;• re-evaluate downstream risks (which may change during the life of the facility)• update the risk assessment; and• evaluate need for changes or updates to risk management plans, contingency plans and emer-

gency preparedness and response plans.

Conduct periodic audit and assessment of the entire tailings management system. Identify items requiring corrective action. Document and promptly report to the designated responsible official, observations and rec-

ommendations arising from inspections, reviews, audits and assessments.

4.2 Corrective action Develop and implement action plans to address items that require corrective action as identi-

fied during inspections, reviews, audits or assessments. Document completion of corrective actions.

5 Annual tailings management review for continual improvement

Conduct an annual review of tailings management to:

• evaluate the performance of the tailings management system, considering inspection, audit and assessment reports, changing circumstances, monitoring results, spills and other inci-dents, recommendations, and the commitment to continual improvement;


342

• evaluate the continuing adequacy of, and need for changes to, policies and objectives for, performance of, and financial resources allocated to the tailings management system; and

• address the need for changes to commitments to Communities of Interest.

Report the observations and conclusions of the annual review of tailings management to the accountable executive officer.

4 IMPLEMENTING THE TAILINGS MANAGEMENT FRAMEWORK

The tailings management framework has been designed for application through the full life cycle of a tailings facility. Implementing the tailings management framework requires:

• confirming and/or customizing the relevant management actions as derived from the tailings management framework;

• assigning responsibility and authority for the management actions to individuals within the organization;

• determining relevant site-specific performance measures as indicators of progress on manage-ment actions and objectives, quantified where practicable, to enable tracking of progress;

• identifying a schedule to provide a time frame for completing significant milestones for a man-agement action, which may include specific delivery dates or times, and/or frequency of ongo-ing or periodic activities such as monitoring and reviews, and providing a clear timeline for key actions;

• adding references, including technical, managerial and regulatory information relevant to the management action and to the site.

The framework is intended to be customized to suit the requirements of specific sites, company policies and local regulatory and community requirements. One way to customize and implement the tailings management framework is through checklists. These can provide a basis for develop-ing customized management systems, operating procedures and manuals, exposing gaps within existing procedures, identifying training requirements, communicating with Communities of Interest, obtaining permits, conducting internal audits, and aiding compliance and due diligence, at any stage of the life cycle. Sample tailings management checklists are provided in the Tailings Management Guide.

When fully implemented at a particular site, a management system based on the MAC tailings management framework will encourage continual improvement in the safe and environmentally responsible management of tailings facilities.

5 AFTER IMPLEMENTATION—VERIFYING PERFORMANCE THROUGH AUDIT AND ASSESSMENT

Integral to responsible tailings management, as per the tailings management framework, is the principle of continual improvement. This is implemented both on a routine basis through checking and corrective action for specific activities, and by periodic verification of implementation of the tailings management as a whole.

Such verification of conformance with the tailings management framework is done through either audit of tailings facility management and an additional qualitative assessment of that man-agement. The third MAC guide, A Guide to Audit and Assessment of Tailings Facility Manage-ment, provides protocols for such verification.

“Audit” and “assessment” describe two distinct protocols that, when employed, can provide assurance that tailings facilities are being managed effectively and are in accordance with the tailings management framework. The recommended protocols provide guidance on the approach and methodology of verification to be implemented either as an audit or as a more extended


343

assessment. This guidance provides the information necessary to design a detailed plan for an audit or assessment, customized to site-specific needs.

• Audit is the formal, systematic and documented examination of an organization’s or facility’s conformance with explicit, agreed, prescribed criteria, in this case, the MAC tailings manage-ment framework.

An audit is not based on opinion, nor is it designed to determine the root cause of deficiencies or to evaluate management system effectiveness.

• Assessment goes beyond measuring against stipulated criteria (as in an audit) to incorporate professional judgment in evaluating the effectiveness, implementation, application and mainte-nance of a management system.

An assessment is driven by a concern for the quality of system design and management process implementation. It can identify system deficiencies and determine their root cause to provide a basis for improving the process.

Applying the audit and assessment protocols can:

a. assist facility and corporate management to evaluate the implementation and effectiveness of a tailings management system, to reduce risk, and to drive continuous improvement; and

b. provide assurance to a company’s CEO, Board of Directors and Communities of Interest that tailings facilities are being effectively managed in conformance with the MAC tailings man-agement framework.

The application of systematic verification also provides a process to ensure that tailings facility management is being implemented comprehensively and effectively. This can further benefit an organization by:

• increasing the awareness and understanding of tailings management issues by managers and employees;

• improving the facility management’s ability to achieve tailings management objectives;• providing a basis for and demonstrating due diligence in risk management;• affirming compliance with regulatory requirements;• reducing potential liability; and• demonstrating a commitment to continual improvement.

Any areas of potential risk or weakness in management system design or implementation, as identified during a management system audit or assessment, will warrant follow-up.

6 OPERATION, MAINTENANCE AND SURVEILLANCE MANUALS

In its continuing work, the Tailings Working Group identified that, while mining companies were making significant progress toward implementing and documenting tailings management sys-tems, and integrating them with overall environmental management systems, further effort was warranted in their application at the operational level. In particular, a need was identified for further guidance in preparing manuals that outline procedures for the safe operation, maintenance and surveillance (OMS) of tailings and water management facilities.

This led to the preparation of the second MAC guide Developing an Operation, Maintenance and Surveillance Manual for Tailings and Water Management Facilities, published in 2003. It recommends rationale, organization and contents for an OMS manual, and describes procedures that should be addressed.

The objective of an OMS manual is to define and describe:

• roles and responsibilities of personnel assigned to the facility;• procedures and processes for managing change;• the key components of the facility;


344

• procedures required to operate, monitor the performance of, and maintain a facility to ensure that it functions in accordance with its design, meets regulatory and corporate policy obliga-tions, and links to emergency planning and response; and

• requirements for analysis and documentation of the performance of the facility.

An OMS manual should enable the performance of a facility to be compared to expectations, design criteria and operating intent, particularly in the event of significant incidents.

7 CONCLUSION

This paper has outlined some of the key background to and principles underlying the MAC tail-ings management guides. Much additional material and guidance is available in the publications. Readers are encouraged to download the guides directly from our website: www.mining.ca Please make use of them and distribute them widely—it is in all our best interests that the mining industry practice to the fullest extent, safe and environmentally responsible tailings management.

The mining industry is currently in a period of unprecedented boom in activity. New projects are being developed and opened at a very high rate. This rapid growth presents additional chal-lenges to the industry to secure and maintain seasoned expertise to design and manage tailings. It is hoped that the MAC tailings management guides will indeed help maintain focus on sound tailings management through the implementation of effective tailings management systems.

REFERENCES

The MAC Tailings Working Group, 1998, updated in 2008, A Guide to the Management of Tailings Facilities, Ottawa, The Mining Association of Canada.

The MAC Tailings Working Group, 2003, Developing an Operation, Maintenance and Surveillance Manual for Tailings and Water Management Facilities, Ottawa, The Mining Association of Canada.

The MAC Tailings Working Group, 2008, A Guide to Audit and Assessment of Tailings Facility Management, Ottawa, The Mining Association of Canada.


345


Developments in the safety and security of mining industry dams

J.W. FredlandMine Safety and Health Administration, Directorate of Technical Support, Pittsburgh, PA, USA

ABSTRACT: Dams are an integral component of U.S. mining operations—making dam safety an important topic for the mining industry. The Mine Safety and Health Administration (MSHA) lists over 2300 dams in its dam inventory. MSHA’s Dam Safety Program includes dam inspections, investigations of problems, training in dam-safety, and, for coal mines, approval of engineering plans for dams larger than a certain size and dams which can present a hazard. This paper dis-cusses the following topics relative to dam safety in the mining industry:

• Updating of the “Engineering and Design Manual: Coal Refuse Disposal Facilities;”• Lessons learned from dam incidents;• Emergency action planning for high-hazard potential dams; and• Security of dams.

Through proper engineering, attention to construction, effective inspections, appropriate emer-gency action planning, and alertness to security issues, the mining industry can ensure that dams on mine property are safe and secure.

1 INTRODUCTION

MSHA lists over 2300 dams in its dam inventory. Of this total, over 350 dams are classified as having “high-hazard potential,” meaning that in the event of dam failure, loss of human life is likely to occur. For this reason, and because of the environmental harm that can occur as the result of failure of a tailings dam, the safety of dams is an important topic in the mining industry.

MSHA’s dam safety program includes dam inspections, investigations of problems with dam performance, dam-safety training, and, in the coal industry, approval of engineering plans for dams larger than a certain size and dams that can present a hazard. The standards applying to dams on coal mine property (“Water, sediment, or slurry impoundments and impounding structures”) are contained in 30 CFR 77.216 and 77.217. For metal and nonmetal mining operations, “Retain-ing dams” are addressed in 30 CFR 56/57.20010.

This paper discusses four important topics concerning the safety of mining industry dams. Con-sideration of these topics can help mine operators to improve their dam safety programs.

2 UPDATING OF “ENGINEERING AND DESIGN MANUAL: COAL REFUSE DISPOSAL FACILITIES”

On February 26, 1972, a “coal waste” dam failed near Saunders, WV. The resulting release of water and coal waste devastated the Buffalo Creek Valley for a distance of approximately 16 miles, causing 125 deaths and leaving 4000 homeless. This tragedy had a profound effect on the coal waste disposal practices in the United States. In 1975, more stringent Federal regulations were promulgated which included requirements for coal companies to submit engineering plans for dams for pre-approval; inspect impoundments at least every seven days for signs of instability; and submit an annual certification that construction was in accordance with the approved plan.


346

In the dam inspections that occurred following the Buffalo Creek failure, it became apparent that many coal waste disposal dams had not been properly engineered. That is, the principles of geotechnical engineering had not been applied to ensure slope stability and seepage control; the principles of hydrology and hydraulics had not been applied to ensure that storm runoff would be handled without causing the dams to be overtopped or damaged by erosion; and construction spec-ifications had not been developed or followed to ensure that dams were properly constructed.

To help remedy this situation, the Mining Enforcement and Safety Administration (MESA), the predecessor agency to MSHA, contracted for development of the “Engineering and Design Man-ual: Coal Refuse Disposal Facilities” (Design Manual). Although coal waste disposal dams share many of the same engineering aspects as conventional dams, they also have unique characteristics with respect to their purpose, materials, and methods of construction, which must be considered in design. The Design Manual was published in 1975 with the intent “to serve as a uniform guide to safe refuse disposal practices for those concerned with this necessary function of coal mining and preparation.”

Significant improvements occurred in the design and construction of coal waste disposal dams as a result of industry and government actions following the Buffalo Creek failure. Remedial actions were taken to correct deficiencies at existing sites while new sites were engineered from the ground up. In the 36 years since the Buffalo Creek failure, no fatalities have occurred as the result of the failure of a regulated coal waste disposal impoundment. However, there have been accidents related to coal waste disposal. In 1981, a woman was killed at Ages Creek, KY, by the failure of a coal waste pile which was being constructed using alternating layers of fine refuse and coarse refuse. In 2000, a slurry impoundment broke into underground mine workings in Martin County, KY, releasing over 300 million gallons of water and slurry and contaminating miles of streams in two watersheds. The Martin County accident prompted the U.S. Congress to ask the National Research Council (NRC) to evaluate coal waste disposal practices. The NRC issued a report, “Coal Waste Impoundments: Risks, Responses, and Alternatives,” (NRC 2002) which included a number of recommendations for MSHA.

One of the NRC report’s recommendations was that MSHA “continue to adopt and promote the best available technology and practices with regard to the site evaluation, design, construc-tion, and operation of impoundments.” MSHA recognized that it would be beneficial to update the Design Manual to reflect what had been learned about coal mine waste disposal practices in the years since the original manual was published. MSHA wanted the revised Design Manual to address topics that had not been covered in detail in the original manual, such as evaluating sites for undermining and breakthrough potential, and defensive design measures to avoid break-through incidents. MSHA also wanted the revised manual to provide information on topics where significant technical developments had occurred or where increased emphasis was now consid-ered appropriate. Topics in these categories included the design of upstream-construction dams for seismic loadings, and the benefits of increased emphasis on monitoring, instrumentation and emergency action planning.

MSHA prepared a statement of work and requested proposals in July 2005. A contract was awarded to D’Appolonia Consultants in September 2005. D’Appolonia assembled a team which included representatives from several design firms with experience and expertise in the design of dams and slurry impoundments. The D’Appolonia proposal included provisions to survey other impoundment designers on design practices. The proposal also provided that interested parties have the opportunity to comment once the draft manual was prepared and prior to the manual being finalized. These steps were taken to attempt to ensure that the manual reflected the full spectrum of coal industry practices.

MSHA’s intent with the updated manual is to provide the best available technical information on the design and construction of coal mining dams. At the same time, MSHA recognizes that a specific recommendation in the manual may not need to be followed in a particular situation, provided a sound technical basis is provided for the proposed design or practice. MSHA’s impoundment plan review process continues to be based on evaluating submitted plans to ensure


347

that proposed design or construction provisions have a sound technical basis, in accordance with current, prudent engineering practice.

One of the significant benefits of the updated manual is that it will be available in electronic format. Besides being searchable by keywords, full copies of references will be included on the DVD for those references for which copyright approval can be obtained. These features should make the updated manual a valuable resource for anyone involved in the design, construction or regulation of coal waste disposal facilities. Information on how to obtain a copy of the manual, either on DVD or as a download, will be provided on MSHA’s web site at www.msha.gov.

3 LESSONS LEARNED FROM DAM INCIDENTS

The updating of the Design Manual highlights the need for all types of mining industry dams to be properly engineered. From MSHA’s investigations of dam failures and problems, deficiencies that have been found at dams that were not properly engineered include the following:

• Slopes that are too steep: The slopes of dams must be flatter than the slopes of other types of embankments or piles because dam slopes have to resist both the forces of gravity and the forces created by water impounded against and seeping through the dam. Water or tailings impounding structures must be constructed at flat enough slopes to have an adequate factor of safety against failure with all reasonable loading conditions (gravity, seepage forces, earthquake effects, etc.) considered in the stability analyses. MSHA has investigated some dams built with slopes near the material’s angle of repose. It is not surprising that stability problems occur with these structures.

• Inadequate internal drainage: In a homogeneous dam constructed without internal drainage pro-visions, seepage will eventually exit on the downstream slope. Such seepage will soften the downstream toe area, make the slope less stable by decreasing the effective stresses, and possibly cause internal erosion to occur. All of these effects are detrimental to the stability of the dam. Internal erosion is a process by which particles are carried along with the seeping water allowing a void to form at the seepage discharge point. Over time the void can progressively work back toward the reservoir and cause the dam to fail (Fig. 1). Internal drainage provisions, such as blan-ket drains and finger drains, are used to reduce the saturation level in the downstream portion of a dam and to release seepage water in a controlled manner without internal erosion.

• Inadequate compaction: Compaction of the dam material increases its shear strength and reduces its permeability. If dam material is not adequately and uniformly compacted, then zones of weaker shear strength and higher permeability will result. Such zones can lead to instability through slope movement, internal erosion, or creation of zones of high pore-water pressure. Proper installation of a conduit that passes through a dam, including adequate compaction of backfill and provisions for drainage, is especially important as the area around the conduit can otherwise provide a path for excessive seepage and internal erosion (FEMA 2005).

• Inadequate freeboard and/or spillway discharge capacity: Dams must be capable of handling the runoff from significant storms without overtopping. The appropriate size storm depends on the size and hazard potential classification of the dam. Sufficient freeboard, i.e. the vertical distance between the water level and the crest of the dam, must be maintained so that runoff into the reservoir can either be stored or passed through a spillway without the dam being over-topped. If an earthen dam is overtopped, there is a good chance that the dam will fail. Typically the minimum freeboard at the highest pool level that occurs during the design storm should be at least 3 feet.

• Inadequate dam inspection: Dams should be inspected regularly for signs of instability and to check that the amount of freeboard and the outlets are prepared to handle the runoff from a large storm. These inspections should be performed by personnel with training in at least the basics of dam safety. At coal mining operations, 30 CFR 77.213–3(a)(1) requires that regulated dams


348

be inspected by a qualified person for signs of structural weakness at least every seven days, unless less frequent inspections are approved by the District Manager. Less frequent inspections are normally only approved for dams with low hazard potential.

In addition to routine inspections, the condition of dams should also be inspected during and/or after significant events such as large storms or earthquakes. At many mining operations, person-nel routinely work around or travel past the dams. These personnel—especially if provided with basic training about dams—can be a valuable resource as far as observing the overall condition and function of the dam. Such personnel should be instructed to immediately report any unusual conditions or signs of instability. Particular attention should be given to evidence of cracking, slope instability, excessive seepage, undue settlement, significant erosion, sinkholes, boils, and blockage or improper functioning of spillways and drains. Routine inspections should be sup-plemented with more formal inspections by engineering personnel. High-hazard potential dams, for example, should typically be inspected at least annually by an engineer familiar with dam safety.

The above discussion is consistent with the findings of a report by the International Com-mission on Large Dams (ICOLDS) entitled “Tailings Dams: Risk of Dangerous Occurrences” (ICOLDS 2001). This report lists 221 incidents of tailings dam failures or accidents. The leading causes of these accidents, in order of occurrence, were slope instability, earthquakes, overtopping, poor foundation, and seepage.

For dams at coal mines, MSHA reviews the plans for the design and construction to verify that they are in accordance with current, prudent engineering practice and that the types of struc-tural deficiencies described above do not exist. While MSHA has no plan review and approval requirement for dams at metal and nonmetal mines, the “Retaining dams” standard at 30 CFR 56/57.20010 states that: “If failure of a water or silt retaining dam will create a hazard, it shall be

Figure 1. Failure of a tailings dam due to internal erosion.


349

of substantial construction and inspected at regular intervals.” Sections 56/57.2, “Definitions,” define “substantial construction” as follows: “Substantial construction means construction of such strength, material, and workmanship that the object will withstand all reasonable shock, wear, and usage, to which it will be subjected.” Thus it is important that the operators of metal and nonmetal mines make sure that their dams are adequately engineered and inspected so that the types of problems indicated above are avoided. This point was well stated in the previously mentioned 2001 ICOLD report where the Tailings Dam Committee concluded that “effective reduction of the cost of risk and failure can only be achieved by a commitment from Owners to the adequate and enforced application of available engineering technology to the design, construction and closure of tailings dams and impoundments over the entire period of their operating life.”

4 EMERGENCY ACTION PLANNING

Even though every effort may be made during the design, construction, and monitoring of a dam to ensure that the dam is safe, dam failures do occur. The significant potential energy created by a dam means that it is prudent, even for the best designed and constructed dams, for the dam owner to develop and maintain an Emergency Action Plan (EAP) for dams whose failure would endanger human life.

In the Buffalo Creek dam failure, 77 of the 125 fatalities occurred to persons located 5 miles or more downstream from the failed dam (Wahler 1973). It was estimated that the flood wave took 45 minutes or longer to reach these locations. These lives, and possibly many more, may have been saved had there been pre-planning for the process of monitoring and evaluating potential emergency conditions, and warning and evacuating affected downstream residents—the type of measures that modern EAPs are intended to address.

An EAP is a formal document that identifies potential emergency conditions at a dam and specifies preplanned actions to be followed to minimize loss of life and property damage. EAPs are recommended in the Federal Guidelines for Dam Safety (FEMA 2004a). The guidelines state: “Each Federal agency which owns or is responsible for dams and each public or private owner of a federally regulated dam should evaluate the possible modes of failure of each dam, indicators or precursors of failure for each mode, possible emergency actions appropriate for each mode, and the effects on downstream areas of failure by each mode.” The guidelines go on to indicate that the dam owner should provide notification and evacuation procedures where failure would pose a significant danger to human life and property. Furthermore, the guidelines advise that emergency planning be coordinated with local officials to establish plans for notifying and evacuating local communities should conditions warrant.

“Emergency Action Planning for Dam Owners,” (FEMA 2004b) provides specific information to guide dam owners in developing an EAP. The Natural Resource Conservation Service (NRCS), working with the Association of State Dam Safety Officials, prepared a sample EAP, which includes a “fillable form,” to further assist dam owners in preparing EAPs. More information can be found on the NRCS web page. Go to http://directives.sc.egov.usda.gov and click on “Manuals,” then “Title 180,” then “National O&M Manual,” then on “Subpart F-Emergency Action Plan.” See the links under Section 500.52, “EAP Content and Format.”

An EAP specifies the actions that a dam owner should take to moderate or alleviate a serious problem at their dam (Fig. 2). If conditions at the dam indicate that failure is reasonably likely to occur, an EAP provides procedures and information to assist the dam owner in issuing early warn-ing and notification messages to responsible downstream emergency management authorities. An EAP also contains mapping showing the downstream area that would be flooded in the event of dam failure.

It is important that EAPs be kept updated to reflect any changes concerning emergency respond-ers and their contact information. The importance of building relationships with local emergency responders prior to an incident cannot be overemphasized. Tabletop exercises are a good way to ensure that all of the involved parties are familiar with each other, knowledgeable about the


350

location and features of the dam, informed about the potential inundation area, and aware of their respective roles in an emergency.

While the benefits of an EAP to a dam owner appear obvious, according to the National Inven-tory of Dams, in 2007, only about half of the over 8000 high-hazard potential dams in the U.S. had an EAP. This shortcoming is a major concern of the National Dam Safety Review Board (Board) and the Interagency Committee on Dam Safety (ICODS). As called for in the Dam Safety Act (last reauthorized in 2006 as Public Law 109–460), these are the two main bodies for monitoring and enhancing dam safety in the United States. In January 2008, the Board and ICODS held a joint meeting at which both bodies voted to establish a goal that within the next five years, all high-hazard potential dams in the U.S. will have an EAP.

MSHA’s regulations do not specifically call for EAPs which are consistent with the Fed-eral Guidelines. In the coal standards, when a potentially hazardous condition develops at an impoundment, 30 CFR 77.216–3(b) calls for the mine operator to do the following: take action to eliminate the potentially hazardous condition; notify the MSHA District Manager; notify and prepare to evacuate miners from coal mine property which may be affected by the potentially hazardous condition; and direct a qualified person to monitor all instruments and examine the structure at least once every 8 hours, or more often as required by an authorized representative of the Secretary. Thus the standard only specifically addresses the evacuation of coal miners from mine property. Obviously a dam failure can affect an area much farther downstream than mine property. For metal and nonmetal operations, EAPs are not addressed in the standards on “Retain-ing dams.” In 1994, MSHA issued Program Information Bulletin (PIB) Number P94–18 which encouraged coal mine operators to prepare EAPs consistent with the Federal Guidelines for dam safety. In 2004, metal and nonmetal mine operators where encouraged to prepare such EAPs in PIB Number P04–09.

MSHA’s dam inventory lists 219 dams at coal mines and 144 dams at metal and nonmetal mines that are classified as having high hazard potential, that is, likely to cause loss of life in the event of

Figure 2. Seepage and internal erosion at Teton Dam shortly before the dam failed.


351

failure. Less than half of these dams have EAPs. The dams that have EAPs are typically located in states, such as West Virginia and Pennsylvania, where the state regulations require an EAP.

To be consistent with the Federal Guidelines for Dam Safety and to meet the goals of ICODS and the National Dam Safety Review Board, MSHA will explore ways to obtain EAPs for all high-hazard potential dams at mining operations. MSHA again encourages mine operators to voluntarily comply with this goal.

5 DAM SECURITY AWARENESS

As highlighted by the terrorist attacks of September 11, 2001, the owners of structures, includ-ing dams, which can pose high risks in the event of failure must consider the potential for such facilities to be deliberately attacked by people intent on causing harm. Dams located upstream of populated areas or located where their failure could cause significant societal impact, such as disruption of water supplies, are the greatest security concerns.

The owners of dams which have high or significant hazard potential should identify where vulnerabilities may exist and be aware of actions that can be taken to reduce the risk of attack. Guidance on the security of dams has been developed by the Department of Homeland Security (DHS). The “Dams Sector Security Awareness Guide: A Guide for Owners and Operators,” (DHS 2007) is available on the internet page of the Association of State Dam Safety Officials (www.damsafety.org), or a copy can be obtained by sending a request to the DHS at the following e-mail address: [email protected].

The goals of the security guide are to enhance dam owners’ and operators’ security postures by providing information regarding identification of site vulnerabilities, detecting surveillance and suspicious activities, and reporting such activities to the appropriate authorities. The key points in the guide are summarized here.

5.1 Surveillance of potential targets

Potential aggressors engage in surveillance activities to identify any security vulnerabilities they can exploit. The overall objective of surveillance activity is to determine possible targets, attack modes, and the likelihood of success of an attack against an asset. An aggressor’s specific surveil-lance objectives could be to identify the following features of an asset:

• Presence or absence of security and security cameras;• Identification cards of employees and contractors;• Security screening procedures for employees, visitors, contractors;• Access point locations;• Locations and characteristics of vulnerable structural com ponents; and• Patterns of concentration of people and vehicles.

5.2 Detection of surveillance and suspicious activities

Mine operators should instruct their personnel to be alert indicators that a dam or other asset is possibly under surveillance. Some indicators are:

• Persons using or carrying video/camera/observation equipment;• Persons with maps, photographs, or drawings with highlighted areas or notes regarding the dam;• Persons parking, standing, or loitering in the same area over a multiple-day period with no

apparent reasonable explanation;• Personnel being questioned off-site about practices pertaining to the dam;• Persons not associated with the dam showing an increased general interest in the area

surrounding it;


352

• Persons observing deliveries, especially of explosives or hazardous materials;• A noted pattern or series of false alarms requiring a response by law enforcement or emergency

services (possibly used to check response times);• Theft of contractor identification cards or uniforms or unauthorized persons in possession of

identification cards or uniforms;• Recent damage (e.g., significant holes or cuts) to a perimeter fence or gate, or damage to perim-

eter lighting, closed-circuit televisions, or other security devices;• Unfamiliar contract workers with passable credentials; contract workers attempting to access

unauthorized areas;• A seemingly abandoned vehicle in the area of the dam;• Delivery of equipment or materials that is unexpected, unusual, out of the norm, without expla-

nation, or with suspicious or missing paperwork.

Mine operators should instruct their personnel to be alert to suspicious activities. An example would be the theft of a large amount of ingredients for explosives (e.g., fuel oil, nitrates). With the routine use of explosives at mining operations, the opportunity for explosives to be used in a terrorist-type attack is especially a concern.

5.3 Reporting incidents

The security guide lists the types of surveillance and suspicious activities that should be reported to authorities. Examples are: gathering of inappropriate information; breach of a restricted area; attempted intrusion into a restricted area; suspicious photography; observation taken to an unu-sual degree; sabotage; and weapons discovery. The guide encourages reporting of information concerning suspicious or criminal activity to local law enforcement as well as to the DHS or the Federal Bureau of Investigation. The National Infrastructure Coordinating Center (NICC), within the DHS, can be contacted at 202–282–9201 or by e-mail at [email protected]. The Guide advocates building relationships with agencies that deal with security issues, such as local law enforcement and local emergency responders, before an incident occurs.

5.4 Addressing vulnerabilities

Some of the steps which mine operators can take to address vulnerabilities include: instructing security personnel and other company personnel to be alert to possible surveillance and the activi-ties of strangers; preventing the public from having access to driving vehicles on the crest of a dam by installing locked gates or barriers at each end of the crest or at other convenient locations on the access road leading to the dam; posting appropriate signs where the public is not allowed (e.g. “No Trespassing”); and preparing an Emergency Action Plan which includes the information for contacting appropriate authorities if a security incident occurs and rapid response by law enforce-ment or emergency management officials is necessary.

6 CONCLUSION

Dams are a vital component of U.S. mining operations. Every owner of a dam that would poten-tially cause loss of life in the event of failure should ensure that the dam is properly engineered and constructed, and routinely inspected. The owner should also develop and maintain an Emergency Action Plan. This is a prudent measure that can save lives in the event of a dam failure. Through proper engineering, adequate attention to construction, effective regular inspections, appropriate emergency action planning, and attentiveness to security issues, the mining industry can ensure that dams on mine property are safe and secure.


353

REFERENCES

Federal Emergency Management Agency (FEMA). 2004a. Federal Guidelines for Dam Safety. FEMA 93.Federal Emergency Management Agency (FEMA). 2004b. Federal Guidelines for Dam Safety: Emergency

Action Planning for Dam Owners. FEMA 64.Federal Emergency Management Agency (FEMA). 2005. Conduits through Embankment Dams: Best Prac-

tices for Design, Construction, Problem Identification and Evaluation, Inspection, Maintenance, Renova-tion and Repair. FEMA 484.

International Commission on Large Dams (ICOLD). 2001. Tailings Dams: Risk of Dangerous Occurrences. Bulletin 121.

National Research Council (NRC). 2002. Coal Waste Impoundments: Risks, Responses, and Alternatives. National Academy Press.

U. S. Department of Homeland Security. 2007. Dams Sector Security Awareness Guide: A Guide for Owners and Operators.

Wahler, W.A., and Associates. 1973. Analysis of Coal Refuse Dam Failure: Middle Fork Buffalo Creek, Saun-ders, West Virginia. PB-215 142, U. S. Bureau of Mines.



355

Life cycle assessment and tailings management trade-off studies—concepts

D. van ZylUniversity of British Columbia, Vancouver, B.C., Canada

ABSTRACT: Trade-off studies for the evaluation of tailings management options, alternative designs, etc. are typically based on cost, environmental and site specific considerations. This short paper presents the concepts of using Life Cycle Assessment (LCA) as a basis for trade-off studies to evaluate tailings management options, alternative designs, etc. LCA has been formalized by the International Organization for Standardization in a series of standards, ISO 14040. In an LCA materials flows, environmental impacts, and energy flows are typically used to evaluate a specific option as well as for trade-off studies. A short review of the LCA prin-ciples, framework and approach is followed by summaries of a few mining and project related studies using LCA’s at the country, mine site and project scale. By defining the “functional unit” clearly as tailings management all the LCA standards can be applied in the evaluation of trade-offs at this scale.

1 INTRODUCTION

Planning, designing and implementing a tailings management facility for a mining project is a big responsibility typically undertaken by a team of engineers and scientists. A series of steps can be involved in this task, including: develop tailings management options, develop concep-tual designs for tailings management alternatives, perform site evaluation and selection stud-ies, select a final site and tailings management approach, perform detailed site investigations, develop a final design for permitting and construction and finally construct and operate the facility. These activities do not necessarily all occur in series and may not be formalized in all projects. However, this process involves numerous trade-off studies where the decision crite-ria are typically based on cost, environmental and other site specific considerations. Typical methodologies in these evaluations and trade-off studies may include Multiple Account Analysis (MAA) to develop multi-stakeholder project expectations, Risk Assessment (typically Failure Mode and Effects Analysis (FMEA)) to evaluate the risks associated with one or more options, and site specific cost analyses using Discounted Cash Flow (DCF). In most of these analyses the full life cycle of the project and facility is included, i.e. construction, operations, closure and post-closure (Van Zyl, 2008).

This short paper presents the concepts of using Life Cycle Assessment (LCA) for tailings man-agement trade-off studies. LCA has been formalized by the International Organization for Stand-ardization in a series of standards such as ISO 14040 (Standard 14040, 2006). The methodology has also been applied by a number of researchers to specific mining issues as summarized below. In an LCA materials flows, environmental impacts, and energy flows are typically used to evalu-ate a specific option as well as for trade-off studies. While these measures can provide useful comparative data the practical nature of using LCA methodology for tailings management must be further evaluated.



356

2 LIFE CYCLE ANALYSIS (LCA)

2.1 International Organization for Standardization

ISO 14040 entitled Environmental management—Life cycle assessment—Principles and frame-work and ISO 14044 entitled Environmental management—Life cycle assessment—Requirements and guidelines (Standard 14040, 14044, 2006) are the latest versions of the standard. There are a number of accompanying standards for specific components of the LCA. This section provides some further discussion of the standard.

“LCA addresses the environmental aspects and potential environmental impacts) (e.g. use of resources and environmental consequences of releases) throughout a product’s life cycle from raw material acquisition through production, use, end-of-life treatment, recycling and final disposal (i.e. cradle-to-grave)” (Standard 14040, 2006).

The standard is also applicable to services and in the context of this paper tailings management may be considered a service. A fundamental decision that must be made in performing an LCA is the definition of the “functional unit”. By selecting tailings management as the “functional unit” all the LCA Standards can be applied.

The following statement from the Standard provides further insight in the application of LCA for the evaluation of functional units such as tailings management: “LCA is one of several envi-ronmental management techniques (e.g. risk assessment, environmental performance evaluation, environmental auditing, and environmental impact assessment) and might not be the most appro-priate technique to use in all situations. LCA typically does not address the economic or social aspects of a product, but the life cycle approach and methodologies described in this International Standard may be applied to these other aspects.”

The four phases in an LCA study are described as follows by the Standard:

• The goal and scope definition phase. The scope, including system boundary and level of detail, of an LCA depends on the subject and the intended use of the study. The depth and the breadth of LCA can differ considerably depending on the goal of a particular LCA.

• The inventory analysis phase. The life cycle inventory analysis phase (LCI phase) is the second phase of LCA. It is an inventory of input/output data with regard to the system being studied. It involves the collection of the data necessary to meet the goals of the defined study.

• The impact assessment phase. The life cycle impact assessment phase (LCIA) is the third phase of the LCA. The purpose of LCIA is to provide additional information to help assess a product system’s LCI results so as to better understand their environmental significance.

• The interpretation phase. Life cycle interpretation is the final phase of the LCA procedure, in which the results of an LCI or an LCIA, or both, are summarized and discussed as a basis for conclusions, recommendations and decision-making in accordance with the goal and scope definition.

Much of Standard 14044 is focused on developing complete and consistent methodologies to develop and describe the LCA. This will be explored in a future paper on LCA methodologies for evaluating tailings management.

2.2 Mining and project related LCA

LCA’s have typically focused on the materials flows, environmental impacts and energy flows for a metal (or other product) from cradle-to-grave. A number of such LCA’s have been pub-lished or are available on public web sites (note that some of these are referenced in Duru-can, et al., 2006). In such analyses the international production, manufacturing, use, re-use, recycle, and disposal of these metals are considered, clearly a daunting task that is extremely data intensive.


357

The same methodologies can also be used at a smaller scale. Stewart and Petrie (2006), in a special edition of the Journal of Cleaner Production that focused on LCA, presents an LCI based methodology to evaluate the environmental performance of the minerals processing Sec-tor. The information is based on a combination of expert knowledge and process modeling. They develop environmental profiles for significant commodities of the South African and Australian industries and conclude that “because of the mix of technologies in place in these two regions, these profiles provide a valuable first order understanding of how the sector as a whole might perform.” This paper and much of their research is focused on country wide impacts and is still of large scope.

Durucan, et al. (2006), in the same issue of the Journal of Cleaner Production, sets out to estab-lish a cradle-to-gate approach for LCA, i.e. providing a better LCI and LCA for the mining related part of the metal cycle, because “very little or no emphasis has been placed on the extraction of the mineral ore and the consequent waste handling aspects of the industry in relation to the alloca-tion of environmental burdens. In other words, the mining system has been largely simplified in a single fact sheet.”

The research by Durucan, et al. (2006) establishes mine specific LCI’s (using relational data-bases) and LCA’s for the following activities: mineral extraction, mineral processing, waste handling and remediation, electrical supply, and life cycle impact assessment. They provide an example of the LCA application for a Bauxite mine in Hungary.

Lundie, et al. (2004) describes the application of LCA methodology to “examine the potential environmental impacts of Sydney Water’s total operations in the year 2021.” They continue: “To our knowledge this is the first study to create an LCA model of an integrated water and wastewater system with this degree of complexity.” They further motivate the study as follows: “Whether publicly or privately operated, water and wastewater systems are often directed by government organizations which need to demonstrate that their performance and their planning processes meet the public’s expectations in terms of ecologically sustainable development (ESD). This concern with sustainability creates a role for LCA. LCA was chosen to examine the potential environmental impacts of Sydney Water’s operations in the year 2021. LCA is seen as a tool more holistic, quantitative, comparative, and predictive than the few alternatives available for comparing alternative technical systems.”

In terms of relating their study to ISO 14040 Lundie, et al. (2004) states: “The LCA was intended to show which aspects of the business place the largest burdens on the environment and to compare alternative future scenarios. By allowing this holistic perspective of environmental issues, the LCA offers a means to move beyond ‘end-of-pipe’ thinking in understanding the sus-tainability of water service provision. The investigation was based on a functional unit defined as the provision of water supply and sewerage services in the year 2021.”

In evaluating the outcomes of their project Lundie, et al. (2004) concludes: “Building a model of the scale and complexity as this one can be a resource-intensive process, but the discipline has benefits beyond the performance of the LCA itself and the insights which it gives to strategic plan-ners. The process of constructing the LCI involves information exchanges between planning and operational staff which can enhance communication in a large organization. It has also provided information which has enriched communication with external stakeholders. Performing LCA has enabled Sydney Water to capture environmental effects associated with the consumption of mate-rials, which does not routinely occur in other strategic planning processes.”

These studies demonstrate how LCA approaches can be used at smaller scales than the LCA of a metal, for example copper. It provides insights in the application of the LCA methodology that can be useful in developing the approach further for tailings management trade-off studies.

3 CONCLUSIONS

The LCA Standard provides details on how to perform and document the process so that it is comprehensive and consistent. Previously researchers and engineers have shown that the LCA


358

methodology can be applied equally to metals (internationally), country wide evaluations of process options, and at the mine level including the full range of activities. Applying the same methodology to tailings management systems is not a conceptual stretch.

While the methodology focuses on “products” it also refers to “services”; tailings management is clearly a service. Furthermore, by selecting the “functional unit” as tailings management all the LCA Standards can be applied. The devil is always in the details and comprehensive application of the LCA Standard to trade-off studies in tailings management will undoubtedly result in a number of challenges. It is expected that the biggest challenge will be to obtain all the applicable informa-tion, i.e. the LCI aspects of the study.

Another important consideration is that although the LCA approach typically includes mate-rials flows, environmental impacts and energy flows the standard leaves it open to also include economic and social issues. Including these aspects may become an important part of the overall LCA for tailings management and will be explored in future work.

REFERENCES

Durucan, S., Korre, A. & Munoz-Melendez, G. (2006) Mining life cycle modelling: a cradle-to-gate approach to environmental management in the minerals industry, Journal of Cleaner Production, 14, 1057–1070.

Lundie, S., Peters, G.M. & Beavis, P.C. (2004) Life cycle assessment for sustainable metropolitan water sys-tems planning, Environ. Sci. Technol., 38, 3465–3473.

Standard 14040—Environmental management—Life cycle assessment—Principles and framework, International Organization for Standardization, 2006.

Standard 14044—Environmental management—Life cycle assessment—Requirements and guidelines, Inter-national Organization for Standardization, 2006.

Stewart, M. & Petrie, J. (2006) A process systems approach to life cycle inventories for minerals: South Afri-can and Australian case studies, Journal of Cleaner Production 14, 1042–1056.

Van Zyl, D. (2008) The mine life cycle system and mine waste management, International Conference on Mining and Industrial Waste, South African Inst of Civil Engineers, 13p.


Case histories



361


Tailings closure at BHPBilliton’s San Manuel operation design and closure construction San Manuel, Arizona

D. OrtmanSRK Consulting, Tucson, AR, USA

ABSTRACT: In 2002 BHPBilliton announced closure of the San Manuel copper operation 40 miles north of Tucson, Arizona. Started by Magma Copper in the late 1940’s, San Manuel disturbed 9,000 acres and included an underground mine that produced over 700 million tons of ore, an 1,100 foot deep open pit, sulfide waste rock dumps, heap leach and SX-EW plant, mill, concentrator, smelter, and tailings facilities impounding over 700 million tons of material, covering 3300 acres. BHPBilliton engaged SRK Consulting to provide site characterization, engineering, permitting, and to assist with construction oversight for the closure. These activities began in 2002 and closure construction started in late 2004. The work was completed in January of 2008 ahead of schedule, under budget, with an exemplary safety record, and without a single contractor claim. This paper presents the tailings closure portion of the design and an overview of the tailings closure construction.

1 INTRODUCTION

The San Manuel copper operation is located approximately 50 miles northeast of Tucson, Arizona. The facility, covering approximately 9000 acres, consisted of both underground and open pit mines, heap leach with SX/EX, mill, concentrator, smelter copper processing facilities, and 3300 acres of tailings impoundments. Started in 1948 by Magma Copper, the mill, concentrator, and smelter processed sulfide copper and molybdenum ore until the current owner, BHPBilliton placed the facility on care and maintenance in 1999. During the 50 year life approximately 700 million tons of tailings were deposited in seven impoundments covering 3,300 acres eastward of the Plant Area and adjacent to the San Pedro River (Figure 1). On October 24, 2002 BHPBilliton announced that that San Manuel would be permanently closed, and characterization of the tailings impoundment facilities was initiated for closure design (SRK Consulting, 2005).

2 FACILITY DESCRIPTION

The tailings facilities include the #1/2, #3/4, #5, #6, and #10 Tailings Impoundments covering an area of approximately 3,300 acres (Figure 1). The tailings are located on a terrace above the San Pedro River. The toe of the tailings is at a distance of between 1,000 and 3,000 feet from the floodplain of the San Pedro River. Construction of the first tailings dam (Tailings Impoundment #1) started in 1950 followed by construction of the #2, #3, #4, #5, #6, and #10 impoundments and consolidation of facilities resulting in the current configuration.

All dams are unlined and were constructed using the upstream method. Until 1996, the embankments were raised using the coarse fraction of cycloned tailings. Thereafter, the embankments were raised with spigoted tailings from the beach adjacent to the embankments. The embankment slopes are generally 3H:1V with the exception of portions of the lower embankment slopes which can steepen to as much as 2H:1V. The beach gradients vary between 0.25 percent and 0.5 percent, and drain away from the embankment crests. The impoundments are


362

up to 300 feet high from crest to toe, with slope lengths up to 900 feet. Tailings supernatant water was decanted through a series of drop inlets and routed to the toe of the embankments via pipes installed beneath the impoundments. The reclaimed water was returned to the plant for reuse.

During operations the embankments were instrumented and evaluated for stability by AMEC. Post-operational analyses conducted by AMEC indicate the static Factor of Safety (FOS) ranges between 1.94 and 2.38 and the pseudostatic FOS ranges between 1.37 and 1.70, depending on the dam being analyzed. These values exceed the Arizona requirements of 1.5 (static) and 1.1 (pseudostatic).

Extensive geochemical characterization, including 40-week humidity cell testing, concluded that the tailings would not be a source of acid rock drainage. The geochemical characterization and the associated groundwater fate and transport analysis indicated no exceedance of a primary water quality standard would occur. During operations approximately 4000 gpm of water infiltrated the tailings impoundments. Following closure the recharge to the tailings facilities was estimated be on the order of 100 gpm or less. Based on these values the post-closure recharge to the tailings will be reduced by approximately 98 percent when compared to the operational period.

3 DESIGN OBJECTIVES AND CONCEPTS

The design is based on regulatory requirements, including state and federal regulations (Arizona Aquifer Protection Permit and Best Available Demonstrated Control Technology, Mined Land Reclamation Plan, Clean Air Act, and Clean Water Act), and the results of characterization and analyses performed as part of a systematic design procedure. The design addressed the following objectives with the given design concepts:

Figure 1. San Manuel tailings area.


363

• Minimize run-on and ponding on impoundments by diverting storm water away from the impoundments and draining the impoundment surfaces;

• Avoid disturbance in Jurisdictional Waters of the United States by not enlarging the footprint of the tailings; therefore the embankment slopes were not flattened and were closed at the existing slope angles (3:1 to 2:1);

• Eliminate storm water contact with tailings by placing a soil layer between storm water and tailings and minimize erosion exposure of tailing surfaces by maintaining dispersed storm water flow on tailings covers and placing blanket rock armor on embankment slopes;

• Prevent overtopping of the embankment during major storm events by installing emergency spillways designed to route the peak flow for the Probable Maximum Precipitation (PMP) event away from the embankment crests;

• Control discharge of sediment laden storm water by upgrading existing and constructing new sediment control structures;

• Eliminate possible discharge from water reclaim facilities by grouting all tailings decant structures; and

• Promote vegetation on all soil covered surfaces by revegetating all soil surfaces.

3.1 Storm water diversion

New diversion channels were designed immediately upstream from the impoundments. The channels were to collect runoff from the closed beach and pond areas of the facilities, capture runoff from upslope of the tailings, and convey both around the impoundments to washes adjacent to the closed tailings (Figure 1). Diversion channels were designed with 3H:1V side slopes and a 20-foot bottom width to accommodate scraper construction. Regulations required the diversions to pass the peak flow from the 100-year, 24-hour event, resulting in a maximum flow depth of 7 feet. However, due to the depth of excavation needed to develop a minimal drainage grade and using the channel excavation as the borrow source for the large amount of cover soil needed for the dams the final channels were significantly larger than required by the hydraulic requirements with depths of 10 to 20+ feet over most of their length.

To avoid long-term ponding supernatant ponds were filled and graded to drain and runoff from the beach and filled pond areas was directed to the diversion channels. The beach areas were left at their existing grade.

3.2 Avoid jurisdictional waters of the U.S.

Many drainages immediately below the toe of the tailings embankments are delineated Waters of the U.S. As part of the overall approach to closure, BHPBilliton requested that the design not disturb any of these areas. Hence the embankment slopes could not be flattened and the design had to accommodate closing the slopes at their existing angles of 3H:1V to 2H:1V.

3.3 Tailings cover and erosion protection

Most tailings surfaces were covered with at least 12 inches of soil to prevent storm water from contacting tailings. The beaches were covered with a nominal 12 inch layer of cover soil and ponds received as much as 5 feet of material to ensure positive drainage. Beach and pond cover required 4 million cubic yards of soil borrowed from the diversion channel alignment. An erosion resistant cover system was placed on all embankment slopes. The slope cover consisted of a 2 foot thick layer of Coarse Alluvium (a local alluvial borrow with a minimum of 35% gravel and cobble) placed on the upper third of the slope and a two-layer cover consisting of 1 foot of Random Alluvium (borrow from the excavation of the diversion channels) topped with 1 foot of Rock Armor (crushed material from a nearby quarry) placed on the lower two-thirds of the slope.

Revegetation studies by GeoSystems Analysis, Inc. (GSA) of Tucson, Arizona indicate that 12 inches of cover soil is sufficient growth medium to sustain a vegetative cover. Once established,


364

the vegetation will promote transpiration to reduce infiltration into the tailings and stabilize the surface to reduce the dust generation potential. On cessation of operations BHPBilliton placed a nominal 4 inches of local alluvial soil on the beaches to mitigate dust generation. The final closure added an additional 9 inches to most beach areas; resulting in a total cover thickness of not less than 12 inches. In designated cases other approaches were used for beach cover. At 10 Dam the existing 4 inches of cover was tilled into the tailings to a depth of 4 inches and covered with an additional 9 inches of soil to create the final cover. This method was suggested by GSA to reduce the amount of imported material and provide a long-term comparison of alternative covers. Portions of the 1/2 Dam beach had received approximately 6 inches of soil cover and been revegetated during reclamation trials in 1991. No additional cover was placed on the portions of beach where vegetation was successfully established.

3.4 Erosion protection of the embankment slopes

Southern Arizona is subject to severe thunder storms during the summer. Although the average annual rainfall at the site is on the order of only 14 inches it is not uncommon to receive localized events of over an inch in a few minutes. During the 2 year construction of the tailings closure the site experienced two storm events of approximately 2 inches of rain in 20 minutes. These short-duration high-intensity storms require robust erosion protection be afforded to all embankment slopes.

Early in the design process it was decided that at blanket armor approach to erosion protection on the embankment slopes was preferable to a lateral channel and spillway system. A number of considerations entered into this decision, but the primary drivers were:

• The direction to not disturb Waters of the U.S. and, hence, to not flatten the existing 3H:1V—2H:1V slopes did not allow sufficient room on the slopes to feasibly construct the large cross-slope channel sections necessary to safely convey the design event peak flows; and

• The assessment that maintaining a distributed flow over the large slope areas was preferable to concentrating flow in constructed channels because the resultant impact from the failure of a channel would likely be much more severe than that from a localized erosion rill cutting through the blanket armor.

The blanket armor design focused on using local materials from sources controlled by the project. Two materials from local sources were used for slope erosion protection, they were:

• Coarse Alluvium—An alluvial deposit of silty sand with a minimum of 35% gravel and cobble content was identified immediately adjacent to the upper side of 6 Dam. This material was evaluated as described below and found to have sufficient gravel and cobble content to resist erosion for slope lengths of 300 feet or less.

• Rock Armor—An 8 inch minus crusher-run limestone developed from the existing Black Hills Quarry that was used to produce smelter flux for the operation. A new source area was identi-fied within the permitted quarry area for production of Rock Armor. The area was drilled and blasted, dozed to the crusher location, passed through a jaw crusher set for a maximum 8 inch size, and stockpiled for haulage. The quarry was approximately 7 miles from the project site.

Both theoretical and empirical work was performed to evaluate the use of these materials for cover on the faces of the tailings dams. Engineering and Hydrosystems, Inc. (E&H) of Littleton, Colorado, conducted the theoretical evaluation to estimate the rock size required to protect the surface of a 900 foot long embankment face with a slope between 3H:1V and 2H:1V. The analyses identified that an eight inch minus, crusher-run rock material with a D50 of four inches would be acceptable protection against erosion under 100 year, 24 hour storm precipitation. E&H also evaluated the use of the Coarse Alluvium for use on shorter slope lengths and concluded that the Coarse Alluvium would be suitable for use on slopes of 300 feet or less. E&H determined that there was sufficient gravel and cobble content for the Coarse Alluvium to self-armor as the finer material eroded out of the soil mass.


365

A field test was then performed to provide an empirical basis for evaluating proposed erosion-resistant covers. Two test channels, each five feet wide and one foot deep were constructed on an existing tailings slope on a divider dike between two impoundments. The first test channel consisted of a 12 inch layer of Rock Armor over 12 inches of Random Alluvium over tailings. The second test channel consisted of 24 inches of Coarse Alluvium over tailings.

A PVC lined stilling basin was constructed at the top of each test channel. The PVC liner overlapped the upper 3 feet of the test channel to provide a smooth flow transition between the stilling basin and the channel. Each test channel had a basin at its lower end to collect and measure the amount of sediment generated during the tests. Each sediment basin had a spillway located to direct the discharge water onto the tailings surface and ultimately to the tailings basin. No water was discharged outside of the tailings facility.

No water pipeline was available at the test area, therefore water trucks were used to fill the stilling basins and flows were controlled by varying engine speeds for the PTO pumps on the trucks. The flow rate for the test was based on the peak flow generated during the 100 year, 24 hour storm event with a Type 2 distribution and assumed the flow would concentrate by a factor of 2 as it progressed down slope; therefore the full flow for the test was twice the peak flow estimated from the storm hydrograph . Each test pad was subjected to at least three flow regimes of approximately 1/3 flow, 2/3 flow, and the full flow from the design storm event. The staged flow tests modeled the peak flow resulting from the design storm at the toe of 300, 600, and 900 foot long slope segments, respectively.

The rock-armor test channel was stable throughout the test, and exhibited only very minor loss or displacement of material. There was no loss of average cover thickness and only trace amounts of sediment were generated. The rock armor surface following the test was essentially the same as that in place at the start of the test and no tailings were exposed during the test.

The Coarse Alluvium test channel eroded during all stages of the test. An average of fourteen inches of total cover section remained in place following the full flow test, and a correspondingly significant amount of sediment was generated. As predicted, the test channel surface developed a remnant surface layer of cobbles and gravel. No tailings were exposed or displaced during the test. Based on the results use of the Coarse Alluvium was considered a feasible method of providing erosion protection for the 1/3 full flow condition, which correlated to a maximum slope length of 300 feet.

Chemical and physical test work was performed on samples of rock from the Black Hills Quarry to support use of the Rock Armor in contact with storm water. Samples of rock were subjected to the Synthetic Precipitation Leaching Procedure EPA Method 1312 and subsequent chemical analysis of the leachate did not identify any constituents of concern. LA Abrasion testing of the Rock Armor indicated the Rock Armor to be an exceedingly durable material.

3.5 Emergency spillways

Emergency spillways were designed to route major flood events away from the dam crests to avoid overtopping. Emergency spillways were designed for the peak flows generated by runoff from upgradient catchments under PMP conditions. Spillway sizing for all dams except 1/2 Dam was based on the six-hour PMP event which generated a larger peak flow than the longer duration PMP events. Due to the limited catchment area for 1/2 Dam the six-hour PMP did not generate sufficient water to reach the elevation of the spillway; therefore the 72-hour PMP was used to analyze the 1/2 Dam spillway. All PMP analyses included flood routing through the impoundments, but conservatively assumed that the upstream diversion facilities fail.

3.6 Sediment control

All runoff from the faces of the closed embankments was routed through sediment control ponds to provide additional protection against release of tailings in the event of a localized failure of the erosion resistant cover. The sediment control ponds included both upgraded existing facilities and new structures.


366

3.7 Decant structure closure

The decant piping was closed by grouting the pipelines a distance of 1500 feet from the outlets at the toe of the embankments to eliminate the future potential to discharge water through the system or to create elevated pore water pressure within the embankments. In addition, exposed risers in the supernatant pond areas were backfilled with low slump concrete.

3.8 Revegetation

All soil cover was revegetated with a seed mix approved by the State of Arizona. Rock Armor covered areas were not seeded, but are expected to naturally reseed from adjacent ground.

4 SLOPE EROSION PROTECTION CONSTRUCTION

The tailings embankment faces cover a total area of 735 acres and slope 3H:1V—2H:1V. The downstream faces are 250 to 300 feet high and have slope lengths of 800 to 900 feet. Total crest length, though discontinuous, is on the order of six miles. Most of the slopes had remnants of intermediate cyclone benches and access roads, as well as significant erosion rilling. Construction of the slope erosion protective cover included the following:

• Crest regrading;• Slope rough grading;• Borrowing and placement of 1 million cubic yards of Coarse Alluvium (24 inches on upper

1/3 slope);• Borrowing and placement of 1 million cubic yards of Random Alluvium (lower 12 inches on

lower 2/3 slope); and• Production and placement of 1 million cubic yards of Rock Armor (upper 12 inches on lower

2/3 slope).

Regrading of the tailings crest included cut-to-fill of tailings material from the dam crest over the tailings header pipe access road and a lower catch bench. In addition, the crest regrading ensured that there would be no discharge of storm water over the crest onto the slope to initiate erosion rilling at the top of the slope.

Dozers rough graded the slopes to eliminate remnants of benches, access roads, and erosion rills that would interfere with the placement of the cover materials or promote concentrated flow on the final cover surface. The intent of the rough grading was to smooth the slopes rather than to significantly alter the topography.

Twenty-four inches of Coarse Alluvium was placed over the upper third of the slopes, to a maximum slope length of 300 feet. The material was hauled from the designated borrow area and stockpiled at the crest of the slope; it was dozed into position with lift thickness controlled with GPS units on each dozer.

The Random Alluvium and the Rock Armor were stockpiled on the outboard side of a road constructed across the slope at the upper limit of the placement area. The materials were dozed into place with lift thickness controlled by GPS the same as the Coarse Alluvium placement. With dozer pushes up to 600 feet, size segregation was observed in the Rock Armor, but this resulted in increasing rock size towards the lower part of the slope to counteract the increasing erosive potential in down slope areas.

5 SLOPE EROSION PROTECTION PERFORMANCE

During the two year tailings closure construction period the site twice experienced storms that resulted in around 2 inches of rainfall in approximately 20 minutes. Although these events resulted


367

in widespread damage to areas where construction was not yet complete, slopes with finished blanket armor covers suffered only minimal and damage which was easily repaired.

REFERENCES

SRK Consulting. 2005. BHP Copper San Manuel Plant Site, Pinal County, Arizona, Tailings Closure Design Report, submitted to the Arizona Department of Environmental Quality, Phoenix, Arizona



369


Case study: Site-wide water balance of the Pierina Gold Mine, Peru

L. GeorgeWater Management Consultants, Inc., Denver, CO, USA

W. LudwickSouth American Region, Barrick Gold, Lima, Peru

J. ChahbandourWater Management Consultants, Inc., Denver, CO, USA

ABSTRACT: Site-wide water management for the mining industry is becoming increasingly important because of the requirements for International Cyanide Management Code and internal water volume accounting. In locations where precipitation is intermittent for parts of the year, it is important to develop storage strategies for use during dry seasons. It is critical to have an accurate inventory of solution at all times in case of uncontrolled circumstances (e.g., extreme precipitation events or power outages). The following case study presents a water balance model that was cre-ated for a complex gold mine operation in Peru. Water Management Consultants (WMC) created a site-wide water balance model in GoldSim based on conceptualizations from mine visits and on the operational conditions of the process facility.

1 INTRODUCTION

Water Management Consultants (WMC) was commissioned by Minera Barrick Misquichilca S.A. to develop a site-wide water balance model for the Pierina Mine (Pierina). The location of the mine is presented in Figure 1. The location of the open pit mine is near the city of Huaraz in north-central Peru and produced approximately 45,000 tonnes per day of gold-silver ore in 2007. The open pit is within the Puca Uran watershed while all process facilities are within the adjacent Pacchac watershed. Both flow to the Rio Santa.

The model developed for this project includes water balances for individual process facilities and integrates them into a site-wide water balance and inventory. The model is based on concep-tualizations from mine visits and on the operational conditions of the process facility as commu-nicated to WMC through discussions with mine personnel (Ludwick, 2006).

The main purpose of the project was to provide Pierina with a tool for process facility solution management. The resulting model was used to assess implications of scenarios such as failure of pumping equipment and extreme precipitation events in order to size containment facilities, and to prevent unintentional releases. The overall goal of the proposed work was to develop a site-wide water balance model for Pierina to:

• provide leach operators sufficient information for daily solution management decision making in compliance with the International Cyanide Management Code (www.cyanidecode.org),

• assess implications of loss of pumping capacity with regards to solution drain-down, extreme precipitation events, and existing operational storage.

• make recommendations that reflect changes due to continued mine expansion regarding exist-ing containment volumes and pumping capacity,


370

• determine an estimate of make-up water or the volume of excess solution in the leach pad; excess solution would represent that volume necessary to be treated before discharge.

The model incorporates the following:

• general operational conditions such as ore stacking rates (varying monthly), solution applica-tion rates (varying daily), leach pad areas and leaching schedule,

• existing management of incident precipitation including exposed liner and raincoats,• up-gradient surface water controls (including roads and drainage networks),• current primary and secondary solution containment systems,• representative climate conditions,• current pumping capacity in primary and secondary containment systems.

2 PROJECT SETTING

Pierina began operations in 1998 and active mining will continue until 2011. The process facilities lie within the Pacchac watershed, which in turn flows into the Rio Santa. The following sections are intended to provide a brief overview of the climate data and site facilities.

2.1 Climate data

Daily data for precipitation and evaporation, along with temperature and wind speed/direction were collected at the Pacchac Meteorological Station (2000–2006). The total average annual pre-cipitation in the area is approximately 1,156 millimeters (mm) with average monthly values rang-ing from 2 mm to 240 mm. The wet season extends from October to April, and the dry season from May to September. Daily average rainfall ranges from about 4 to 8 mm during the wet season and 0.08 mm to 1.3 mm during the dry season. The total average annual pan evaporation is approxi-mately 1,112 mm with average monthly values ranging from 48 mm to 126 mm. The 100 year storm event, as determined by mine personnel, is 98 mm over a 24-hour period.

Figure 1. Location of Pierina Gold Mine.


371

2.2 Site facilities

The modeled portion of the site consists of facilities within the Pacchac Valley catchment that cov-ers an area of about 5.3 square kilometers (km2). The facilities included are:

• six holding ponds,• the leach pad and caisson platform,• the processing plant,• the Acid Rock Drainage (ARD) treatment plant,• the South Diversion Ditch (SDD),• and the North Diversion Ditch (NDD).

Currently, the leach pad covers an area of approximately 1.15 km2. The leach pad is lined with a HDPE geomembrane and has a series of under-drains that convey spring flows from beneath the pad. Portions of the leach pad surface are covered with “raincoats” (plastic liners positioned over large portions of the pad) to divert rainfall and prevent dilution of the pregnant leach solution (PLS). Under normal operating conditions there is no pond containing PLS exposed to the atmos-phere. The leach pad includes an internal dike and caisson platform that provides containment for the PLS. The PLS level is maintained below the surface of the caisson platform and the PLS is contained within the pore spaces of the ore, under saturated conditions.

The PLS is transferred to the processing plant via vertical turbine pumps located within the caisson platform.

The six holding ponds include the raincoat pond, sedimentation pond, two sludge ponds, collec-tion pond, and the polishing pond. The raincoat pond is used to collect the runoff from the raincoat liners placed over the leach pad. The sedimentation pond collects runoff from the waste dump area and some of the up-gradient area. The sludge ponds are used to store thickener solids from the ARD and cyanide destruction processes, and temporarily store barren leach solution from the processing plant. The collection pond stores water diverted from the sediment and raincoat ponds. The polishing pond is the final holding pond for the water on site.

The ARD Plant was originally designed for neutralization of ARD, but has since been converted to transfer water and solids from cyanide destruction to other locations. The processing plant han-dles all PLS solution that is pumped from the caissons. The NDD is used to divert water from the waste dump and other areas north of the site to the sedimentation pond. The SDD is used to divert runoff water around the process facility and return it to the Pacchac drainage. The SDD collects runoff from up-gradient catchment areas and can accept discharges from the sedimentation, col-lection, or polishing ponds.

3 TECHNICAL APPROACH

The technical approach for the water balance modeling included:

• conceptualization of hydrologic model for Pierina facilities within the Pacchac Valley,• creation of the model and its components within the probabilistic simulation program GoldSim

(GoldSim, 2007),• development and calibration of a water balance model for the entire site using historical data

from 2006,• development of a predictive model for use in evaluating potential impacts of various scenarios

requested by the mine.

The following sections discuss the different components of the water balance.

3.1 Conceptual hydrologic model

The final conceptual hydrological model is presented in Figure 2. The model incorporates all water storage facilities and flows within the process facility, including flows that are not used at


372

Figu

re 2

. Si

te-w

ide

wat

er b

alan

ce c

once

ptua

l mod

el f

or P

ieri

na.


373

this time but were required for predictive purposes. The model was run on a daily timestep because of the requirement that it be used as a daily solution management tool.

The arrows in Figure 2 indicate the direction of flow for solution/water on site. The main facili-ties are shown in bold. All flows indicated on this diagram are represented in the model. The flow on site is described in further detail in sections 3.1.1, 3.1.2, and 3.1.3.

The model includes water balances for individual site facilities and integrates them into a site-wide water balance and inventory. The facilities include six separate holding ponds, the leach pad and caisson platform, raincoat liners used to cover the leach pad, the processing plant, the acid rock drainage (ARD) plant, and two diversion ditches. Spring flow beneath the leach pad is collected by the under-drain system and re-circulated back into the leach pad. Waste rock and the surrounding natural and disturbed areas were incorporated as part of the water balance to calculate the runoff diverted into the holding ponds and diversion ditches.

3.2 Climate

Historical data were used directly in the model, when available, and were also used to calculate the statistics utilized for the predictive climate behavior. For predictive model runs, the model uses the mean and standard deviation for daily precipitation and evaporation values from 2000 to 2006 to create stochastic data sets that change with every iteration of the model. The average monthly precipitation and evaporation are presented in Figure 3. The data sets are based on a normal distri-bution about the mean for evaporation, and a log-normal distribution about the mean for precipita-tion using the means and standard deviations of daily data. These distributions produced predicted precipitation and evaporation data sets that closely resembled historical data. A 100 year storm event, previously calculated by mine personnel, was also incorporated into the model and can be inserted anywhere in the modeled time period for prediction of necessary storage capacity.

3.3 Surface water management

The Pacchac drainage area was split into seven sub-basins which were defined for each of the key facilities included in the water balance. The sub-basins were defined based on topography, surface water diversions, and roads and were classified based on surface and soil conditions.

The surface water runoff volumes for each pond were calculated using the National Resources Conservation Service (NRCS) curve number method (NRCS, 1972). This method is typically applied in the United States, but, as Peru does not have a defined method to estimate precipitation runoff, this method was assumed appropriate for use with the water balance model.

Pacchac Meterological Station Average Monthly Precipitation and Pan Evaporation (mm)

0

50

100

150

200

250

Avera

ge D

ep

th (

mm

)

Precipitation 134.3 176.1 239.7 119.3 37.5 8.8 2.2 10.5 51.5 100.3 103.3 173.1

Evaporation 105.4 59.4 48.3 68.1 96.8 87.0 123.7 126.4 92.5 119.9 111.9 72.8

Janurary February March April May June July August September October November December

Figure 3. Average monthly precipitation and pan evaporation values.


374

The NRCS approach uses climate data, and accounts for infiltration losses based on anteced-ent moisture conditions (AMCs). The following equation is used to calculate the runoff from a specific area:

QP I

P I Sa

a

= −−

( )

( )

2

+

where Q is the runoff depth, P is the precipitation depth, Ia is the initial abstraction depth including

surface storage, interception and infiltration prior to runoff, and S is a storage parameter deter-mined using the following equation:

S = (1,000/CN) − 10

where CN is the curve number. Curve numbers range from 0–100 and are a function of the soils ability to infiltrate water. The CN is determined by the AMC described below, and the land use. I

a

is typically approximated as 0.2*S, for a final runoff equation of:

QP 0.2*S

P S= −

+( )

( . )

2

0 8

Runoff curve numbers for each sub-basin area were chosen based on site soil conditions. Curve numbers as a function of AMC were assigned based on surface conditions and are explained below.

The AMC allows the model to account for water already in the soil when rainfall occurs. The AMC is based on dry (AMC I), moist (AMC II) or wet (AMC III) soil. The model handles this by looking at the previous five days of precipitation. If rainfall over that five day period totals less than 12.7 mm, the soil is at AMC I, if it totals between 12.7 mm and 27.9 mm, the soil is at AMC II, if the total is greater than 27.9 mm, the soil is at AMC III. The different curve numbers for each AMC used in the model are listed in Table 1 (Ward, 2004). Curve numbers were estimated from comparing surface conditions on site to the NRCS curve numbers listed for agricultural and commercial use for hydrologic soil group B, or C (NRCS, 1972). Soils classified as group B and C have moderate and slow infiltration rates when wet, respectively.

The roads and surface water diversions work similarly to convey water off site and to different holding ponds (Figure 2). The NDD and SDD serve as the major surface water diversions on site. The NDD conveys surface water from the waste dump/sedimentation pond sub-basin to the sedi-mentation pond, and the SDD conveys the surface water from its surrounding sub-basin, as well as some of the process water off site.

Table 1. Curve numbers for antecedent moisture conditions.

On site surface conditions/NRCS surface conditions/ hydrologic soil group AMC I AMC II AMC III

leach pads 0 0 0waste rock/cultivated with conservational tillage/B 52 71 86disturbed areas/cultivated with conventional tillage/B 63 81 92partially disturbed areas/cultivated/B 60 76 87undisturbed areas/thin stand forest/B 48 66 80haul roads/roads/B 66 84 96other dirt roads/roads/C 78 90 96liners 100 100 100


375

3.4 Process water management

All flows are modeled as instantaneous flow from one location to another except the flow through the leach pad where storage in the pore space was modeled to include time for pore-space wet-up and drain-down.

The downward movement of precipitation and leach solution applied to the leach pad is delayed during the “wet-up” period. That is, by the process of increasing the water content in the ore from the initial moisture content (average of 10.7% by volume, based on column test data received from the mine, Ludwick, 2006) to the moisture content where solution begins to flow directly through the ore (average of 18% by volume, also based on column test data received from the mine, Ludwick, 2006). The model uses the thickness of the ore and the leach application areas to calculate the volume of ore under leach. The outflow from the pad is then calculated as a function of the irrigation rate for each area.

When a new area is placed under leach, or irrigation, the volume of solution required to increase the volumetric water content from the average 10.7% to 15% represents the initial solution up-take. Solution will flow, or leak, from the new ore into the under-lying ore at a rate of 1% of the total volume applied per day until wet-up is completed. When the moisture deficit has been met, the volumetric water content will be 18% and steady-state operational conditions will result in the inflow rate equaling the leakage rate. The leakage rates were determined during model calibration. This process is illustrated in Figure 4.

When leaching stops, the solution drains out of the ore until a volumetric water content of 15% is reached. This is the estimated residual water content of the ore.

The model also includes a delay of one day to account for the time required for solution to move along the bottom liner of the leach pad.

The leach pad is directly connected to the caisson platform, as all flows of PLS from the leach pad reach the caissons. The solution contained within the caisson platform is pumped to the processing plant, and barren solution is pumped back to the leach pad. If cyanide destruction is

Figure 4. Schematic of the processes modeled for solution movement through the leach pad.


376

necessary to manage the water balance, barren solution is sent from the cyanide destruction circuit to the ARD plant for pH adjustment. ARD plant overflow can be sent to either the collection or polishing ponds. Thickener underflows are transferred to the sludge ponds.

Any solution stored in the sludge ponds can be transferred back to the caissons. There are two surface water diversions within the pad, one to divert precipitation falling on the exposed liner to the SDD, and a series of canals from the raincoats to divert precipitation to the raincoat pond.

The raincoat pond discharges to either the sedimentation or collection pond, depending on cyanide level. The sedimentation pond collects a large amount of surface water runoff and allows eroded sediments to settle out. Overflow from the sedimentation pond can go to either the SDD or the collection pond. The collection pond can also be diverted to the SDD, or to the polishing pond, and the polishing pond can then divert water to the SDD or off site. The polishing pond is also used as a source of make-up water to the processing plant and leach pad if necessary.

4 MODELING RESULTS

4.1 Calibration

The model calibration was conducted using site data from January to August 2006, while the site was under normal operating conditions. The purpose of this calibration was to make sure that the model matched the simulated water elevations to historical water elevations in three of the holding ponds and the solution elevation in the caissons.

4.1.1 Calibration approachCalibration of the water balance model involved use of historical climate data along with available historical facility process flow data. Some of the process flows were not measured or recorded on site; reasonable maximum flows were estimated based on site visits and evaluation of each facil-ity. Some reasonable adjustments (based on estimated error in the measurements of flow) were also made to the historical data in order to fully calibrate the model. The final calibration model matched the historical solution elevation in the caissons as well as the collection, polishing, and sludge ponds relatively well, as described below.

4.1.2 Calibration resultsTo calibrate the model to the historical caisson elevation, the historical irrigation and PLS pump-ing rates had to be adjusted. None of the precipitation onto the exposed liner was diverted to the SDD. Other data used during the calibration were historical values received from the mine includ-ing precipitation, irrigation areas, make-up water, and under-drain diversions.

To calibrate the water elevations in each of the holding ponds, assumptions of pumping rates from one facility to another were made. Pumping rates needed to match the pond elevations were limited to 150 m3/hr. This is the flow limit for most of Pierina’s pumps and conveyance systems at the site.

Figures 5 through 8 present the calibrated solution elevations versus historical solution ele-vations for the caissons and for the holding ponds where there were historical data available. Typically, the difference between the two data sets is less than one meter, except in the case of the sludge ponds, where the difference is up to two meters. The difference in water elevation is much larger in the sludge ponds than seen in the other ponds because the total volume of the sludge ponds is relatively small when compared to the other storage facilities. There were no historical data available to calibrate flows for the sedimentation and raincoat ponds.

4.2 Predictive runs

The main purpose of the predictive model is to provide Pierina with a tool for daily facility solution management. The model will be used to assess implications of the scenarios such as


377

Collection Pond Water Level (m)

3850

3852

3854

3856

3858

3860

12/31/2005 3/1/2006 4/30/2006 6/29/2006 8/28/2006

Date

Sta

ge

(m

) A

MS

L

Historical Stage (m) Modeled Stage (m)

Figure 6. Calibrated water elevation for the collection pond.

Caisson Water Level (m)

39723974397639783980398239843986

12/31/2005 3/31/2006 6/29/2006

Date

Sta

ge

(m

) A

MS

L


Figure 5. Calibrated water elevation in the caisson.


378

Sludge Ponds Water Level (m)

3920

3922

3924

3926

3928

12/31/2005 3/31/2006 6/29/2006 9/27/2006

Date

Sta

ge

(m

) A

MS

L


Figure 8. Calibrated water elevation in the sludge ponds.

Polishing Pond Water Level (m)

38233824382538263827382838293830

12/31/2005 3/31/2006 6/29/2006

Date

Sta

ge

(m

) A

MS

L


Figure 7. Calibrated water elevation in the polishing pond.


379

failure of pumping equipment, and extreme precipitation events to size containment facilities and prevent accidental releases. Two examples of predictive runs are explained in the following sections.

4.2.1 Loss of PLS pumping capacityA model run simulating loss of PLS pumping capacity was developed to predict the length of time required for the caisson platform to fill up to its maximum elevation of 3,990 m from a starting elevation of 3,981 m (the high elevation at normal operating conditions with a total available stor-age volume of 349,800 m3). Three scenarios were modeled: barren solution application continues as normal, barren solution application is reduced by 50 percent, and barren solution application ceases during the shutdown.

The model predicted that the solution elevation in the caissons would near the maximum eleva-tion within seven days if the leaching is continued at the normal rate. With the leach rate cut in half, it takes 12 days to reach the maximum elevation. With leaching shut off completely, the dike did not overtop within 20 days, it reached a solution elevation of 3,986 m.

4.2.2 Impact of extreme precipitation eventsA model run with a synthetic 100-year-storm event occurring during the wet season (98 mm over one day) was executed to predict what would happen to the caissons and pond water elevations. The starting elevations of water in the ponds were based on normal operating levels. The model predicted that the sedimentation and raincoat ponds will overflow.

In this scenario, the overflow from the raincoat pond goes to the sedimentation pond, and the sedimentation pond overflows to the collection pond. The sedimentation pond overflow is not transferred to the SDD, because there is likely to be a large amount of sediment involved with this size of storm. This sediment would be discharged to the collection pond. No discharge offsite would occur from the collection pond because there is enough volume in the storage facilities to hold the runoff water from this storm.

5 CONCLUSIONS

The Pierina site-wide water balance provides the information required to safely manage daily solution management on site. The model can also be used as a decision-making tool by entering all historical data into the model and creating predictive runs. Day-to-day management can be simulated easily by mine personnel using the Graphical User Interface (GUI) that comes with the GoldSim model. The results of the model runs provide the user with a summary of flows around the mine site. This will allow the user to determine the necessity for make-up water or excess solu-tion that will need to be put through cyanide destruction.

The objectives of this project were fulfilled through use of the site-wide water balance model. Mine personnel currently use the model for daily solution management decision making in com-pliance with the International Cyanide Code. With estimation of key model inputs for future oper-ations, mine personnel can evaluate extreme storm events and volumes required for make-up or cyanide destruction.

SPECIAL THANKS

It is important to acknowledge the vital contributions made by the people of the Pierina Process Department, especially Wesley Ubillus and Pedro Puente, and to thank them for their efforts to make the project a success.


380

REFERENCES

International Cyanide Management Institute, International Cyanide Management Code for The Manufacture, Transport and Use of Cyanide In The Production of Gold (Cyanide Code), http://www.cyanidecode.org, Updated Jan 9, 2007.

Ward, A. & Trimble, S. Environmental Hydrology, Second Edition, Lewis Publishers, London, pp. 132–133, 2004.

Natural Resources Conservation Service (NRCS, formerly Soil Conservation Service, SCS). 1972. National Engineering Handbook, Section 4, Hydrology.

GoldSim Technology Group L.L.C., GoldSim Probabilistic Simulation Environment, Version 9.50, 2007.Ludwick, W., Ubillus, W. & Puente, P. Minera Barrick Misquichilca, S.A., personal interviews and commu-

nication. 2006–2007.


381


Reclamation of the Panna Maria uranium mill site and tailings impoundment: A 2008 update

C.L. StrachanMWH Americas, Inc., Fort Collins, CO, USA

K.L. RaabePanna Maria Uranium Operations, Hobson, TX, USA

ABSTRACT: Panna Maria Uranium Operations began reclamation of the Panna Maria uranium mill site and tailings impoundment in south Texas, following mill shutdown in 1992. The closure plan and reclamation progress was documented by the authors in the 1998 Tailings and Mine Waste Conference. The closure plan for the tailings impoundment (and associated areas) was submitted to Texas regulatory agencies in 1993, with details of the plan finalized from agency review and site specific conditions in 1994. The closure plan was prepared to meet uranium mill tailings performance criteria under Texas and U.S. Nuclear Regulatory Commission regulations. These criteria include isolation and containment of radiological hazards under long-term conditions, with minimum reliance on post-reclamation active maintenance. Mill decommissioning and site reclamation began (under Texas agency approvals) in 1993, with tailings impoundment covering completed in 1996, and site revegetation initiated in 1998. Reclamation work since 1998 has included monitoring of tailings cover settlement, inspection of site-wide revegetation success and erosion control, re-permitting of a sediment control dam for wildlife considerations, and development of the West Mine lake. This paper will describe the work since 1998. The reclaimed mill site and tailings impoundment are scheduled for final review and acceptance by Texas agencies prior to property transfer to the U.S. Department of Energy.

1 INTRODUCTION

The Panna Maria uranium facility was operated by Panna Maria Uranium Operations (PMUO) as an open pit mining and conventional milling operation for recovery of uranium. The site is approximately 60 miles (96 km) southeast of San Antonio in Karnes County, Texas. The site is located within the south Texas uranium belt, a region of uranium mining and milling dating from the 1950’s.

A general description of the site and closure plan permitting issues are summarized in the following sections from Strachan and Raabe (1997 and 1998).

2 SITE DESCRIPTION

The site is in the south Texas Plains vegetation region, characterized by gently rolling hills covered with grasses and brushwood. The site is within a region of net evaporation, with annual precipitation averaging approximately 30 inches (760 mm) and annual pan evaporation averaging approximately 80 inches (2032 mm). The site is within the San Antonio River watershed, one of the primary rive systems in the region draining to the southeast into the Gulf of Mexico.


382

2.1 Geologic setting and mining

The site is underlain by southeast-dipping Miocene sediments, consisting primarily of poorly-consolidated siltstones, claystones, and fine-grained sandstones. In the site area, these sediments include the lower sequence of the Catahoula Formation, underlain by the Fashing and Tordilla Members of the Whitsett Formation of the Jackson Group. The Catahoula and Fashing Members are the units immediately beneath the site and comprise the non-mineralized overburden materials removed during mining. Stockpiled overburden materials provided the primary borrow materials for reclamation.

The uranium ore produced from the Panna Maria mines (named for the nearby historic Polish settlement of Panna Maria) came from mineralized areas of the Tordilla Member (primarily fine-grained sands). The uranium mineralization occurred as a roll-front deposit in marginal marine sediments. Five open-pit mine sites, covering approximately 1,200 acres (485 ha), were developed by PMUO starting in 1977. The mines were operated until 1985. Overburden removal and mining were conducted by both truck-and-shovel and scraper operations. Mine sites were reclaimed concurrently with mining by backfilling the open pit with overburden, and reveg-etating the surface. The final Panna Maria mine site was left open in anticipation of improved uranium market conditions, but was eventually reclaimed. Reclamation of the mine site was completed in 1994.

2.2 Milling and tailings disposal

The Panna Maria mill and associated facilities were constructed near the planned mine sites within a selected 560-acre (227 ha) area underlain by a clayey unit within the Catahoula Formation. PMUO started the Panna Maria mill in 1979 for recovery of uranium. The mill utilized an acid leaching and solvent extraction process, with the final product being dried U

3O

8 (yellowcake).

The mill was initially shut down in 1985 with the completion of Panna Maria mine production. The mill was converted to an acid neutralization process to treat tailings pond water and make it usable for dust control during mine reclamation. The mill was modified for renewed start-up in 1986 to process ore from the Mount Taylor underground uranium mine in New Mexico. Milling of Mount Taylor ore was stopped for economic reasons in 1988, when the mill was modified to process ore from the Rhode Ranch open-pit uranium mine in Texas. PMUO processed Rhode Ranch uranium ore until final mill shutdown in 1992.

From 1979 to 1992, the Panna Maria mill produced approximately 6.8 million tons (6.2 million tonnes) of tailings from the three ore sources described above. All of the tailings processed at the mill were discharged into the Panna Maria tailing pond (Figure 1).

2.3 Tailings impoundment description

The Panna Maria tailings pond site was located based on proximity to the mill and underlying geology (specifically the clayey unit within the Catahoula Formation). The impoundment was designed as a ring-dike structure, with tailings contained with a zoned earth embankment and above natural silts and clays. The tailings impoundment within the embankment centerline covered approximately 150 acres (61 ha) and was designed for an ultimate capacity for 10 million tons (9.1 million tonnes) of tailings.

The zoned earth embankment was constructed in one stage to a crest elevation of 375 feet (114 m). The embankment contained a central core that was keyed into underlying silts and clays. The embankment was constructed with 2:1 (horizontal:vertical) inside slopes and 3:1 outside slopes, with embankment heights ranging from 30 to 60 feet (9 to 18 m).

Tailings were discharged as a slurry by spigotting from the perimeter embankment. Final tailings thickness within the impoundment ranged from approximately 10 to 50 feet (3 to 15 m). Approximately 77 percent of the total tailings tonnage is from processing Panna Maria mine ore, followed in stratigraphy by the Mount Taylor ore (8 percent) and Rhode Ranch ore (15 percent).


383

3 CLOSURE PLANNING AND INITIAL WORK

The mill and tailings impoundment were permitted and operated under regulations administered by Texas regulatory agencies. Closure plan permitting and reclamation for the tailings impound-ment was permitted as a Radioactive Materials License amendment administered by Texas agencies responsible for radiation control.

3.1 Closure plan criteria

Closure criteria for the tailings impoundment are outlined in Texas regulations and are based on U.S. Environmental Protection Agency (EPA) reclamation standards (40 CFR 192) that are administered by the U.S. Nuclear Regulatory Commission (NRC) as technical criteria for recla-mation (10 CFR 40, Appendix A). The technical criteria provide control of radioactive materials by long-term isolation and stabilization of the tailings.

This is achieved by consolidation of tailings and associated site materials and covering with an erosionally stable soil cover. The release of radon gas (radon-222) from the cover over the tailings (generated by radioactive decay of radium-226 in the tailings) is limited to an average of 20 pico-Curies per square meter per second. Closure must be effective for a design period of 1,000 years, or at least 200 years, without reliance on active maintenance. These criteria are met by design for extreme storm events, and conservative evaluation of erosional stability.

In addition, agreement on ground water protection issues with the regulatory agency resulted in limiting infiltration through the soil cover to an equivalent saturated hydraulic conductivity of 5 × 10–8 centimeters per second (cm/sec) or 0.05 feet per year (ft/yr). In order to meet these criteria, the closure plan for the tailings impoundment included the following components:

1. Consolidating and covering the tailings with random fill.2. Constructing a multi-layered cover system over the tailings, which is designed to reduce the

rate of radon emanation and the rate of precipitation infiltration.3. Regrading the original tailings embankment slopes to a maximum slope of 5:1.4. Managing surface runoff from the reclaimed tailings impoundment with designed slopes, chan-

nels, and selected vegetation and rock protection for non-erosive conveyance of storm runoff.5. Regrading and excavating (where necessary) soils in the mill site, ore storage, and pond areas

with radium-226 concentrations above specified limits, with disposal of excavated materials in the tailings impoundment.

Figure 1. Operational site layout.


384

3.2 Operational planning prior to closure

In 1990 (two years prior to mill shutdown), PMUO started planning for site closure. This work included identification of potential slope configurations for the reclamation surface of the tailings impoundment and evaluation of reclamation cover systems. The planning provided guidance for water management and tailings consolidation for the final stages of tailings impoundment operation. This planning also resulted in discharge of tailings during the final years of operation in areas that were consistent with the planned reclamation surface, thereby reducing the amount of tailings regrading after mill shutdown.

Prior to mill shutdown, the water stored in the tailings impoundment was processed for recovery of uranium in the solvent extraction circuit in the mill, then neutralized with calcium carbonate from local sources of caliche. The neutralized water was stored in the solution storage ponds on site until the neutralization process was completed (the M&M ponds shown in Figure 1). This processing and neutralization treatment was conducted in 1985 and also prior to mill shutdown in 1992. The later stage of treatment neutralized approximately 100 million gallons (380,000 cubic meters) of water, which were discharged into the M&M ponds and then transferred back into the tailings impound-ment. The residual, neutralized tailings fluids were evaporated within the tailings impoundment.

3.3 Tailings regrading and covering

A key component of tailings pond closure was the method for initial covering of the tailings, providing a surface for earthmoving equipment and a foundation for subsequent cover system construction. As the first stage, causeways were constructed across the tailings impoundment with random fill. These causeways divided the impoundment into four cells and served as access roads for subsequent tailings regrading and covering (Figure 2).

In limited areas (primarily within the northwest cell), tailings were regraded to be consistent with the planned subgrade surface. Regrading was accomplished with low ground-pressure equipment in tailings areas that had experienced sufficient consolidation and drying to allow equipment travel.

From the causeways, a random fill or interim cover zone was constructed over the tailings using random fill soils from adjacent area cleanup and stockpiled soils not suitable for cover material.

Figure 2. Tailings impoundment covering layout.


385

The random fill zone was constructed with low ground-pressure or conventional earth moving equipment, depending on the consolidation and moisture conditions of the tailings. The random fill zone was an important component of tailings covering, providing a firm base for cover system construction, a zone of separation between tailings and uncontaminated cover materials, and a fill zone with capacity for materials from soil cleanup.

3.4 Tailings fluid evaporation

Due to the volume of water to be evaporated and the planned schedule for tailings pond closure, the rate of evaporation was optimized by spray systems and by construction of shallow evaporation ponds within the tailings impoundment. The shallow evaporation ponds were constructed on top of the compacted random fill surface with perimeter berms two to three feet high. The ponds were constructed at elevations that were consistent with the subgrade surface and planned cover systems (Figure 3).

3.5 Tailings settlement

Tailings settlement was expected from (a) initial settlement of the tailings slurry and (b) during tailings regrading and covering.

Actual tailings settlement was evaluated during the tailings regrading and random fill placement stage. Additional random fill was added to compensate for this settlement, and settle-ment was correlated with the volume of additional random fill that was required. This evaluation showed that actual settlement was similar to settlement predicted during design. Actual settlement was not significantly large due to: (a) discharge of the majority of tailings prior to 1985 (allowing sufficient time for consolidation of these tailings), (b) the relatively coarse grind of the tailings, and (c) the managed disposal of tailings in the later stages of operation. The managed disposal resulted in coarse-grained tailings being distributed throughout the impoundment, reducing drainage distances from the fine grained tailings.

Figure 3. Cover system detail.


386

4 COVER SYSTEM

The tailings cover system is shown in Figure 3. The portion of the cover system above the random fill zone consisted of the components summarized below. The sources of cover materials were primarily the stockpiled mine overburden and the M&M pond embankment shown in Figure 1. These materials were evaluated for compaction, hydraulic conductivity, radon attenuation, and dispersivity prior to use in construction. Topsoil was obtained from existing stockpiles on site and had previously been evaluated (along with establishment of vegetation species) as part of the mine site reclamation work completed by PMUO.

The total cover system thickness is a minimum of four feet (1.2 m), meeting design criteria for tailings isolation, radon attenuation, and reduction of infiltration. The four-foot cover thickness is also required to meet applicable Railroad Commission of Texas reclamation regulations. The cover system was keyed into the existing perimeter embankment (Figure 4).

4.1 Infiltration barrier

The lower layer of the cover system is the infiltration barrier, a one-foot (300 mm) thick zone of compacted soils consisting of selected materials designed to achieve an equivalent saturated hydraulic conductivity of 5 × 10–8 cm/sec (0.05 ft/yr). On-site clayey soils were evaluated for this layer and did not have sufficient clay content to consistently achieve the desired hydraulic con-ductivity. Two off-site clays were selected as additives to the on-site stockpiled overburden. The additives consisted of a sodium montmorillonite or “bentonite” from western Texas (added at two percent by weight) and a silty clay form a sand washing operation near San Antonio (added at a minimum of five percent by weight).

The barrier was placed, mixed, and compacted in maximum six-inch (150 mm) lifts with con-ventional earthmoving equipment. Standard compaction specifications for clay liner materials were used (95 percent of Standard Proctor density and one percent below the three percent above optimum moisture content). Material specifications for the infiltration barrier included plasticity index and fines fraction requirements for the stockpile material, as well as for the two clay addi-tives. Material specifications for the clay-amended mixture consisted of a percentage of fines of at least 45 percent and a plasticity index of at least 15. The minimum clay percentage by weight was converted to an equivalent volume per specified area for construction control. During construc-tion, PMUO collected relatively undisturbed samples of completed infiltration barrier for on-site permeability testing to confirm that the design value of less than 5 × 10–8 cm/sec was met.

4.2 Cover

The middle layer of the cover system is a two-foot (600 mm) thick zone of compacted, on-site materials to provide additional radon attenuation, a root zone for vegetation, and a zone of

Figure 4. Reclaimed impoundment cross-section.


387

protection for the infiltration barrier from biointrusion and desiccation. This layer was placed in a maximum six-inch (150 mm) lifts and compacted with conventional, earthmoving equipment to standard compaction specifications (95 percent of Standard Proctor density and one percent below to three percent above optimum moisture content).

Material specifications for the cover layer consisted of a percentage of fines of at least 30 percent to ensure desired radon attenuation characteristics.

4.3 Topsoil

The top layer of the cover system is a one-foot (300 mm) thick zone of topsoil for establishment of vegetation. The topsoil had been salvaged and stockpiled from initial mill site construction and overburden excavation. The topsoil was placed in loose lifts and prepared for revegetation.

5 DRAINAGE PLAN AND FINAL SURFACE

The overall drainage plan for the tailings impoundment area is shown in Figure 5. The cover sur-face over the reclaimed tailings impoundment slopes to the north at a 0.50 percent grade. Runoff from the north-sloping cover surface discharges in a controlled manner through an outlet channel down to the natural ground surface and into the reclaimed West Mine Lake. Where the outlet chan-nel bed slope has a 10 percent slope, the channel surface was covered with riprap. Other reaches of the outlet channel surface with less steep slopes were topsoiled and vegetated.

5.1 Drainage evaluation

Runoff from planned drainage was evaluated for erosional stability according to guidelines docu-mented in NRC (1990). This includes calculation of runoff velocities due to peak flow from the Probable Maximum Precipitation (PMP) event. The velocity of this peak flow is compared with permissible non-erosive velocities and acceptable tractive forces on the reclaimed surface under varying conditions of vegetation quality (ranging from full coverage to bare soil conditions).

Figure 5. Reclaimed site drainage plan.


388

Where runoff velocities were above acceptable values (such as the outlet channel) riprap protec-tion was designed for the channel surface. Riprap materials and particle size were selected based on criteria for durability in NRC (1990). A dolomitic limestone source from near Marble Falls, Texas was selected for the riprap, based on acceptable size and durability (SMI, 1997).

5.2 Surface reclamation

The area outside of the tailings impoundment was reclaimed in a manner that removed con-taminated soils and established drainage for runoff away from and around the reclaimed tailings impoundment. The mill site area was regraded for runoff to drain to the north and into the reclaimed West Mine Lake (Figure 5). Contaminated soils and other materials excavated from the

Table 1. Summary of reclamation construction materials.

Material specifications Compaction specifications

Max Passing Lift Size No. 200 Plasticity thicknessa Relative Water (in.) Sieve index (in.) compactionb contentc

Material Description (mm) (%) (%) (mm) (%) (%)

Random fill Fill Above tailings – – – – – –

Compacted random Upper surface of 6 (150) fill landfill 3 (75) >30 – max. >95 –1 to +3

Infiltration barrier Lower 1.0 ft of – – – – – – material cover• Stockpile material 3 (75) >30 – – – –• Sodium bentonite (2% by weight) 1 (25) >70 >60 – – –• Silty clay (5% by weight) 1 (25) >60 – – – –• Barrier mixture 3 (75) >45 >15 6 (150) >95 –1 to + 3 max.

Cover material Middle 2.0 ft of 6 (150) cover 3 (75) >30 – max. >95 –1 to + 3

Embankment and channel 12 (300) outlet fill Clean fill 6 (150) >12 – max. >90 –2 to + 2

Riprap Outlet channel surface• Type A D50

= 8 in. 20 (500) – – 15 (375) – – min.• Type B D

50 = 10 in. 28 (700) – – 21 (525) – –

min.• Filter material D

50 = 1 in. 8 (200) – – 6 (150) – –

min.• Bedding material 2 (50) <30 – 6 (150) – – min.

Topsoil Upper 1.0 ft of Cover and 12 (300) surface 6 (150) – – min. – –

a Lift thicknesses are listed as maximum, minimum, and nominal values.b Relative compaction is expressed as a percentage of maximum density from the Standard Proctor test.c Water content is expressed relative to optimum water content from the Standard Proctor test.


389

surrounding reclaimed area were placed in the tailings impoundment as random fill (as discussed above). Contaminated soils were determined by radium-226 activity concentration, at levels estab-lished by Texas regulatory agencies. Unsalvageable materials from mill de-commissioning were taken apart and buried in the tailings impoundment.

The initial tailings embankment outside slopes (constructed at slopes of 3:1) were regraded by cut-and-fill earthwork to a maximum slope of 5:1. The surface of the reclaimed embankment was designed to be several feet higher than the top of the cover system surface to ensure that runoff will not overtop the reclaimed embankment crest. Embankment fill was placed in lifts of 12-inch (300 mm) maximum thickness and compacted to standard compaction specifications (90 percent of Standard Proctor density and two percent below to two percent above optimum moisture content). The final embankment slopes were covered with a one-foot (300 mm) thick layer of topsoil for establishment of vegetation.

The material and compaction specification for the cover system and fill material are summa-rized in Table 1.

5.3 Establishment of vegetation

The reclaimed impoundment cover surface, reclaimed embankment slopes, and surrounding areas were covered with topsoil for establishment of vegetation. For all of the vegetated areas on site, species were selected for eventual establishment of a full, self-sustaining vegetative cover. These species consisted of King Ranch bluestem (Bothriochloa ischaemum Var. songarica), buffalograss (Buchloe dactyloides), sideoats grama (Bouteloua curtipendula), and kleingrass (Panicum col-oratum) on gently sloping areas; and kleingrass and medio bluestem (Dichanthium aristatum) on embankment slopes and channel surfaces. Maintenance of these species and discouragement of woody or brushy species will be conducted by fertilizing and mowing.

6 CLOSURE HISTORY

After mill shutdown in 1992, PMUO started site closure with mill decommissioning and sur-rounding site regrading and cleanup. The closure plan was submitted by PMUO to the Texas Natural Resources Conservation Commission (TNRCC) in 1993. Based on TNRCC comments, additional issues were evaluated and documented in an updated closure plan that was submitted to TNRCC in late 1994 (WWL, 1994). Many of the closure plan elements had been previously discussed with TNRCC prior to closure plan submittal. This allowed mill site decommissioning and reclamation to take place in late 1992 and 1993, and for M&M pond reclamation to start in 1993. Fill placement and regrading within the tailings impoundment was started in 1993, and was conducted concurrently with evaporation of remaining tailings water. Tailings impoundment reclamation work started in 1994 with regrading on the outside slopes of the tailings embankment. Cover system construction started along the southern side of the impoundment in late 1995.

Based on finalizing the clay-amended mixture design for the infiltration barrier, updated construction specifications were submitted by PMUO to TNRCC in 1995 (SMI, 1995). Adjustments to the outlet channel design on the north side of the impoundment were submitted by PMUO to TNRCC in 1996 (SMI, 1996). Responsibility for the radioactive materials license for the site was transferred from TNRCC to the Texas Department of Health (TDH) in July 1997, and was transferred to the Texas Commission on Environmental Quality (TCEQ) in 2007.

Through 1996, tailings pond closure work was conducted in a manner consistent with the closure plan and included embankment slope regrading, impoundment surface regrading, and cover construction. Outlet channel construction was started at the end of 1996. Completion of the cover system, outlet channel, and remaining closure plan components was scheduled for early 1998 with vegetation to be established by the end of 1998.

Performance monitoring of the site has continued since 1998, with results submitted to TDH and TCEQ. The monitoring results have shown that (a) rates of measured radon emanation though


390

the cover are within standards, (b) groundwater level and quality are within predicted values, (c) measured settlement of the cover surface is within expected low levels, (d) vegetation coverage is as expected, and (e) no erosion of soils from slopes has been observed.

6.1 West mine lake

As outlined in Section 2.1, the Panna Maria mines were developed as sequenced cut and fill operations, with mine overburden placed in previously mined areas. The resulting open mine area was the West Mine, northwest of the tailings impoundment. PMUO planned the reclamation of the West Mine for a lake. The upper slopes of the West Mine were regraded, and covered with topsoil, and vegetated. The catchment area draining into the West Mine allowed gradual filling with runoff that was unaffected by tailings reclamation activity. The West Mine lake reached full capacity in 2005. During the period of lake filling, PMUO developed habitat and initial benthic species for successful development of the lake as a bass fishery (Figure 5).

6.2 RK pond

PMUO constructed a sediment retention basin in the late 1970’s west of the tailings impoundment during initial site construction work. During the period of mill operation and site reclamation, this basin (the RK pond) has become a valuable wetland. In 2003, PMUO received permission form the Railroad Commission of Texas to keep the RK pond a permanent structure for wildlife habitat. This permission was granted after PMUO conducted the appropriate flood routing and embankment slope stability analyses.

7 CONCLUSION

The reclaimed tailings impoundment described in this paper is in the final stages of performance monitoring prior to property transfer to the U.S. Department of Energy. Tailings impoundment closure has been designed to meet the design criteria administered by TNRCC, TDH, and TCEQ for tailings isolation, radon emanation, infiltration, and erosional stability. The closure plan has been developed around existing site conditions, and closure costs have been reduced as much as possible by: (a) designing the closure surface to match the tailings surface at mill shutdown, (b) using on-site materials as much as possible, and (c) carefully selecting off-site materials (when necessary). Post-closure performance monitoring of the reclaimed tailings impoundment and surrounding areas has been conducted for nearly 10 years. The monitoring results have shown good vegetation success with minimal erosion in all of the reclaimed site areas. Reclaimed areas of the site have been successfully used for grazing and wildlife habitat, with environmental monitoring showing key air quality and water quality constituents within accepted values. Cover settlement over the tailings has been relatively low, with values within predicted ranges.

REFERENCES

Shepherd Miller, Inc. (SMI) 1995. Specifications of Cover System Construction. Prepared for Panna Maria Uranium Operations and submitted to Texas Natural Resources Conservation Commission.

Shepherd Miller, Inc. (SMI) 1996. Description of Modifications to the Closure Plan for the Panna Maria Tailings Pond. Prepared for Panna Maria Uranium Operations and submitted to Texas Natural Resources Conservation Commission.

Shepherd Miller, Inc. (SMI) 1997. Assessment of Riprap to be Used for Closure Plan of the Panna Maria Tailings Pond. Prepared for Panna Maria Uranium Operations and submitted to Texas Natural Resources Conservation Commission.


391

Strachan, C.L. & Raabe, K.L. 1997. “Design, Permitting, and Construction Issues Associated with Closure of the Panna Maria Uranium Tailings Impoundment,” Proceedings of 14th Annual Meeting, American Society for Surface Mining and Reclamation, Austin, Texas, May 10–15, pp 601–608.

Strachan, C.L. & Raabe, K.L. 1998. “Reclamation of the Panna Maria tailings impoundment: A case history,” Tailings & Mine Waste ’98, Balkema, pp 825–834.

Texas Natural Resources Conservation Commission (TNRCC) 1990. Texas Regulations for Control of Radia-tion Part 43, “Licensing of Uranium Recovery Facilities.” Texas Register, April.

U.S. Nuclear Regulatory Commission (NRC) 1994. Final Staff Technical Position on Design of Erosion Pro-tection Covers for Stabilization of Uranium Mill Tailing Sites.

Water, Waste & Land, Inc. (WWL) 1994. Description of Closure Plan for the Panna Maria Tailings Pond. Prepared for Panna Maria Uranium Operations and submitted to Texas Natural Resources Conservation Commission.



393


Mining impacts: A case study

R.K. Will & W.E. MotzerTodd Engineers, Alameda, CA, USA

ABSTRACT: A historic underground gold mine in the Sierra Nevada Mountains of central California was re-opened as an open-pit surface mine and operated for several years until the ore body was mined-out. A mine closure plan was submitted and accepted by regulatory agencies fol-lowed by extensive reclamation of all site features except the mill tailings pile. The mine owners developed an alternate capping plan based on performance criteria, but the revised plan was not initially accepted by the state. As a result of delays the entire mine site was re-examined. Concerns re-addressed included off-site migration of dissolved metals in groundwater, stormwater sediment transport, the formation of a pit lake containing dissolved constituents, and other complex issues. Reclamation is now in progress following negotiated closure specifications.

1 INTRODUCTION

The Jamestown Mine is in Tuolumne County in the historic “Mother Lode” gold mining district in the western foothills of the Sierra Nevada Mountains in central California (Figure 1). Following the discovery of gold in 1849 many prospectors and miners came to the area initially develop-ing the placer deposits in the shallow alluvial sand and gravel sediments along creeks and rivers. As these deposits were depleted, underground mines were developed along veins in the bedrock.

Location MapTODD ENGINEERSAlameda, California

June 2008

Yosemite

National

Park

120

49

108

49

JamestownMine

Pinecrest

Sonora

TuolumneJamestown

Tuolumne County

0 10 20

Scale (Miles)

N

KEY MAPNot to Scale

Figure 1


394

The Jamestown Mine was one of these underground mines and was operated until World War II when Executive Order L-208 terminated gold mining. The increasing price of gold in the 1970’s spurred renewed exploration and development. The Jamestown Mine was re-opened as an open pit operation in 1986 by the Sonora Mining Company.

2 MINING FACILITIES

2.1 Mine pits

Three primary open pits were developed: Harvard Pit, Crystalline Pit and South Crystalline Pit as shown in Figure 2. The Harvard Pit is 2700 feet long, 1500 feet wide and 600 feet deep, shown in Figure 3. The other two pits are much smaller. Production of ore totaled 17,000,000 short tons yielding 660,000 troy ounces of gold.

2.2 Milling facility

Most of the gold occurred as microscopic particles (30–150 microns) attached to surfaces of iron pyrite crystals. Ore was processed in a 7,000 ton-per-day concentration mill erected adjacent to the Harvard Pit. Coarse gold was extracted by gravity and fine grained gold ore processed by froth flotation. The sulfide concentrate, containing the fine grained gold, was shipped by truck to a cya-nide treatment plant in Nevada for extraction of the gold. It is interesting to note that a 42 pound gold nugget was recovered in the mill, the largest nugget recovered in California. The nugget is on display in the nearby town of Murphys.

2.3 Tailings Management Facility (TMF)

Tailings from the mill were piped to a 120 acre lined tailings impoundment behind a containment dam (Figure 4). The TMF was constructed with a leachate collection system and a groundwater

CrystallinePit

SouthCrystalline

Pit

DP -5

R SA

DP -6Harvard Pit

DP-1

TMF

N

ScaleinFeet

1,0000

Based on vertical airphotography taken June30,1993.

TODD ENGINEERSAlameda, California

June 2008

Aerial Photograph

Figure 2


395


June 2008

Harvard Pit

Figure 3

TODD ENGINEERSAlameda,California

June 2008

Tailings PondFigure 4

underdrain. Below the TMF containment dam, a process water retention pond (PWRP) was con-structed to collect leachate and allow recycling of process water.

2.4 Rock Storage Area (RSA)

The rock storage area was for storage of overburden and waste rock stripped from the mine pits to uncover the ore-grade rock is shown in the aerial photograph (Figure 2) and after reclamation (Figure 5).


396

2.5 Stormwater Detention Ponds (DP-1)

Four stormwater runoff detention ponds remained following mine closure.

3 MINE CLOSURE

3.1 Ore depletion

Sonora Mining Company ceased operations in 1994 due to declining ore grade and lower market prices of gold. A mine closure plan was submitted by the Sonora Mining Company to the State of California Mine Reclamation Board and was approved and accepted in 1994. Reclamation work was completed consistent with the closure plan with the exception of the tailings impoundment capping.

3.2 Mine pit reclamation

The Crystalline pits were backfilled with waste rock, re-contoured, covered with topsoil, and revegetated. The Harvard Pit remained open to perform as a groundwater sink to collect ground-water seepage from the mine site as depicted in Figure 6. As the pit filled with ground water and direct precipitation, the size of the pit lake was predicted to increase over many years until the evaporation rate from the pit lake surface equaled the groundwater and precipitation inflow rate and thus “equilibrate”. A cross-section through the Harvard Pit Lake is shown in Figure 7.

3.3 Mill site

The processing equipment such as crushers, ball mills, flotation cells, etc. were salvaged and sold. The mill buildings were demolished and the some of the steel and other metals were recycled.


June 2008

Reclaimed WasteRock Pile

Figure 5


397

3.4 Tailing Management Facility (TMF)

The tailing impoundment dam was re-vegetated. The leachate collection and recovery system was managed by pumping and recirculating the leachate. The excess water that collected in stormwater detention ponds evaporated by aerial spraying or used for irrigation of the vegetation. Two of five stormwater detention ponds similar to the one shown in Figure 8 remained to collect stormwater runoff and to serve as wetland wildlife habitats.

RockStorage

Area

Crystalline

TailingsManagement

Facility

LEGEND

Monitoring well

Ground water contour (f msl)

Ground water flow direction

Key mining area

Surface Water Drainage Flow

HarvardPit

DP-5

Source: MWH (2003).

SouthCrystalline

Pit

( 1204 )

12000 600

N

Scale in Feet

TODDENGINEERSAlameda, California

June 2008

Direction,Jame stown Mine

TD MW 22

TD MW 18

TD MW 12TD MW 3

TD MW 4

TD MW 16

TD MW 15

T DM W9

T DM W6

TD MW 14

TD MW 19

TD MW 14

RSM W9A

RSM W8A

TD MW 7

Ground water FlowFigure 6

Cross SectionThrough Harvard Pit

Lake and Wood’sCreek


June 2008

A’A

January2001Harvard Pit Lake

Elevation

WoodsCreek

January 2001Ground water Elevations

600

800

1000

1200

1400

1600

600

800

1000

1200

1400

1600

Scale in Feet2x vertical exageration

0

0

225

Figure 7


398

4 ALTERNATIVE CAP DESIGN

4.1 Alternative cap design

The mine operators developed a lower cost alternative capping plan in 1994 for the tailings stor-age pile based on performance criteria as allowed by state regulation. The revised plan was not initially acceptable to Tuolumne County because of disagreement over the hydraulic conductivity of the capping materials.

5 LITIGATION

In 2001, as a result of delays in implementing the cap, the California Attorney General’s Office on behalf of the Department of Water Resources filed an injunctive relief and civil liability complaint against the Sonora Mining Corporation, the County and past mine operators. The complaint alleged that the mine violated a Cleanup and Abatement Order (CAO) and Waste Discharge Requirements and that the mine waste materials were polluting groundwater and surface waters of the state.

5.1 Revised mine closure plan

The defendants entered into negotiations with the state and a team of mining consultants was assembled by the defendants to work with the state’s technical staff in developing revised mine reclamation plans.

5.1.1 Water monitoring program and new monitoring wellsThe revised closure plan included updated surface water and groundwater sampling and analysis criteria. Additional groundwater monitoring wells were proposed to improve delineation of poten-tial contaminant migration. In addition, a plan to sample adjacent domestic wells down gradient


June 2008

Detention PondFigure 8


399

of the mine was developed. Constituents of concern and ranges of concentration in groundwater and surface water were:

Arsenic 0.002–0.025 mg/lSulfate 150–2000 mg/lNitrate 0–170 mg/lMagnesium 20–160 mg/lSodium 0–150 mg/lCalcium 40–220 mg/lTDS 200–4000 mg/l

Performance criteria were developed along with a contingency plan to pump water from desig-nated wells to create a capture zone along the mine perimeter.

5.1.2 Stormwater controlAdditional surface water diversion and collection channels were designed to supplement the exist-ing detention ponds. Erosion control measures were increased in high runoff areas. Regrading of the low grade ore stockpile followed by soil capping was conducted to prevent generation of acid rock drainage. In-situ treatment alternative methods were used to precipitate select constituents in the Harvard Pit Lake.

5.1.3 Tailing management facilityA redesigned cover for the tailings impoundment was specified by performance criteria and water was diverted from one of the retention ponds into the Harvard Pit Lake.

5.2 Litigation settlement

The litigation was settled in July 2006 and the Jamestown Mine Reclamation Project was awarded by the Jamestown Mine Trust to a private contractor: Shaw Environmental Liability Solutions.

6 RECLAMATION PROGRESS

The following reclamation progress report was provided by Shaw Environmental Liability Solu-tions on June 13, 2007.

6.1 Tailings management facility

Dewatering of the 50 acre pond on the tailings impoundment was performed by siphoning 300 mil-lion gallons of water to the Harvard Pit. Following a drying period the 130 acre surface was cov-ered with on-site low permeability soils creating a cap meeting California Title 27 requirements.

6.2 Rock storage area

An investigation is being conducted to further define the areas potentially requiring regrading to reduce slope gradients and characterization of the rock to determine leaching potential. Consid-eration is being given to selling rock for rip-rap or as feed stock to produce aggregate.

6.3 Hydrogeologic evaluation

The movement of groundwater at the site is being monitored and re-evaluated to refine effective transmissitivity and migration to the Harvard Pit.


400

6.4 Treatment technologies

Treatment of the Harvard Pit Lake water to precipitate or remove metals and salts may be required and alternative technologies are being evaluated. Enhanced evaporation of the pit lake using dye or water cannons is being studied. The use of existing detention ponds as a bioreactor for pit water remediation is also being considered.

7 SUMMARY

The Jamestown Mine is slowly being transformed by careful planning, design and implementation of selected reclamation plans. The long term challenge is to prevent migration of toxic metals into the surrounding groundwater.


401


Design & construction of an evaporation pond at a historic uranium mining facility

T.A. ChapelSenior Geotechnical Engineer, Tetra Tech, Inc., Fort Collins, CO, USA

C. WoodwardEnvironmental Coordinator, Denison Mines (USA) Corp., Denver, CO, USA

R. JolleyJones and Demille Engineering, Richfield, UT, USA

ABSTRACT: In September, 2007, Denison Mines (USA) Corp. received permits from the Utah Division of Oil, Gas, and Mining to begin mining operations at the Tony M Mine, an underground uranium mine that was originally developed by Plateau Resources in the 1980’s. Previous recla-mation included breaching the dam that had been constructed to retain water pumped from the mine. Following geotechnical investigation by Tetra Tech during 2006, the dam and appurtenant facilities were re-designed, and construction occurred during 2007. The pond subgrade provided a cost effective, low permeability native barrier between the mine water and the underlying sedi-ments. The liner consisted of reconditioned and densely compacted native clay overlying Brushy Basin Clay. Seepage through the pond liner and at the toe of the embankment was analyzed for steady-state and transient seepage conditions using a finite element computer model.

Construction of the evaporation pond involved some interesting challenges, including process-ing and compacting the clay soils, which had liquid limits in excess of 100. Laboratory permeability test results did not initially correlate well with field infiltration test results, however when alternative processing and testing methods were adopted the regulatory specifications were achieved. The native liner design and construction permitted a high quality, low permeability barrier to be constructed at a reasonable cost compared to the improbable and infeasible cost of a synthetic liner system in this remote area. Work is currently underway to investigate if similar techniques can be used to develop a larger evaporation facility for phase 2 of the mining operation at Tony M.

1 INTRODUCTION & HISTORY

The Tony M Mine is an underground uranium mine that was developed by Plateau Resources and operated between approximately 1982 and 1983. Uranium was produced from sandstones within the Salt Wash Member of the Morrison Formation. Mining operations ceased in 1986. The mine is located in Townships 34 and 35 South, Ranges 11 and 12 East, Garfield County, Utah. The general project location is shown on Figure 1.

The mine site is in an arid environment in south-central Utah, approximately 15 miles north of Lake Powell, 10 miles south of Mt. Pennell in the Henry Mountains, and west of Utah Highway 276 (see Figure 1). The location of the existing dam and evaporation pond is on Bureau of Land Management (BLM) land about 5000 feet northwest of the existing Tony M mine portals and sur-face facilities. The surrounding land is undeveloped, aside from the mine workings, and is used for cattle grazing during some parts of the year. The ground consists of sparsely vegetated hills and a shallow valley in the Brushy Basin Member of the Jurassic Morrison Formation.


402

Groundwater was encountered during mining, and a method of discharge for water pumped from the mine was needed. Because of the arid location, the mine site is well suited for evapora-tion of mine water from a surface pond. An earthen embankment dam was designed and con-structed in an ephemeral drainage to retain water pumped from the mine workings. When Denison Mines (USA) Corp. (then operating as International Uranium Corporation (IUSA)) made plans to re-open the mine, the dam consisted of two embankment segments that had been separated by a breach channel during reclamation. The original dam was a homogeneous, clay embankment with a maximum section (measured adjacent to the breach location) of 33 feet, a crest width of 25 feet, and a crest length of approximately 630 feet. The crest elevation was 4,879 feet. The upstream face of the dam had a slope of 3.5h:1v and the downstream face had a slope of 2.5h:1v, with a 74 foot wide buttress located approximately 15 feet below the crest. The breach channel was approxi-mately 20 feet wide with 2.5h:1v side slopes was excavated through the dam and down to an eleva-tion of approximately 4,847 feet. The material that was removed from the dam was placed on the buttress and downstream slopes, resulting in a top width of the embankment of approximately 40 feet and a slope on the downstream face of about 5h:1v. Additional material from the breach was placed on the floor of the pond in an area immediately upstream from the breach.

The pond area was reported to have retained water to a maximum elevation of 4,874 feet. At that level, five feet of freeboard would have been maintained below the dam crest and approximately 18.2 acres would have been inundated. Previous construction in the pond area included an emer-gency spillway at an elevation of 4,976 feet on the south perimeter of the pond (see Figure 1).

Denison Mines (USA) Corp. retained Tetra Tech to assist in permit applications with the State of Utah. The permitting process was complicated by the fact that no new uranium mining permits had been issued in the State of Utah for many years. Staff had moved on, been reassigned, or had begun their careers after the previous mines had been closed. Table 1 (below) shows the permits and dates they were issued for the Tony M Mine.

Figure 1. Location and topography of the Tony M Evaporation Pond.


403

The Dam Safety Permit issued by the Utah State Engineers Office and the Stormwater Diver-sion Construction Permit, Evaporation Pond Construction Permit and Groundwater Discharge Permit by Rule issued by the Utah Department of Environmental Quality all apply specifically to the construction of the evaporation pond; most of the permits outlined in Table 1 include approvals for the construction and use of the evaporation pond.

2 GEOTECHNICAL INVESTIGATION & DESIGN

Tetra Tech investigated the feasibility of reconstructing the earthen embankment and constructing a native clay liner to retain water pumped from the Tony M mine. It was proposed to store water in the pond during evaporation. The construction of the evaporation pond began in September of 2007 and was completed in January of 2008. Water was pumped into the pond in February of 2008 and mine dewatering is currently ongoing.

Subsurface conditions at the evaporation pond site were investigated by drilling three test holes in the existing dam, eight holes in the pond area, and three borings south of the pond area. Hand samples were also collected from the pond area near the emergency spillway and from a former borrow area. The approximate locations of the borings and the hand sample location within the pond are shown on Figure 1.

Soils encountered in borings in the earthen embankment were up to 51 feet of silty, moder-ate to high plasticity clay fill overlying claystone and sandstone bedrock of the Brushy Basin Member of the Jurassic Morrison Formation (Davidson, 1967). In the un-compacted fill removed from the breach and placed on the downstream slope of the dam, standard penetration resistance tests indicated blow counts of 8 and 9 blows for 12 inches. In the compacted fill, samples had a range of penetration resistance values (N) of 22 blows for 12 inches to 49 blows for 12 inches. Laboratory testing on comparatively undisturbed samples included natural moisture content and dry density, swell-consolidation, unconfined compressive strength, direct shear, Atterberg limits, water-soluble sulfates, and permeability tests.

Laboratory dry densities ranged from 91 pcf to 119 pcf. The average dry density was 101 pcf. Moisture content of the fill ranged from 7 to 31 percent and averaged 21.3 percent. A sample tested for Atterberg limits had a liquid limit of 55 and a plasticity index of 33, with 64.6 percent silt and clay sized particles (passing the number 200 sieve). Falling head permeability tests were

Table 1. Permits for the Tony M Mine.

Permit Date received

Air Order July 25, 2007Groundwater Discharge Permit by Rule March 22, 2007Water Rights Confirmation December 13, 2007Dam Safety Permit April 20, 2007Evaporation Pond Construction Permit June 27, 2007Large Mine Permit/Plan of Operations/Reclamation Plan September 5, 2007Septic System Construction Permit July 13, 2007Stream Channel Alteration Permit April 16, 2007Stormwater Diversion Construction Permit February 14, 2007Cultural Clearance December 22, 2006Concurrence on Wildlife Protection January 8, 2007County Conditional Use Permit August 22, 2007Utah Pollution Discharge Elimination System Permit May 9, 2008BLM Decision Record /Finding of No Significant Impact September 5, 2007 and November 23, 2007

(from Denison Mines (USA) Corp. 2007).


404

performed on a comparatively undisturbed sample obtained from a California liner at 19 feet, and from a sample remolded to 95 percent of standard Proctor maximum dry density at optimum moisture content. Results were1.40 × 10–8 cm/sec for the California sample and 2.8 × 10–9 cm/sec for the remolded sample. An unconfined compressive strength of 6,832 psf was measured for a sample of the fill from a depth of 24 feet, and an unconfined strength of 6,364 psf was meas-ured for a sample of the fill that was remolded to 95 percent of standard Proctor maximum dry density at optimum moisture content. Direct shear tests on two samples of the embankment fill showed a range of cohesion of 1,450 psf to 1,750 psf and angles of internal friction (phi) of 20 and 31 degrees. A sample remolded to 95 percent of standard Proctor maximum dry density at optimum moisture content had an angle of internal friction of 39 degrees and a cohesion of 240 psf. A sample from a depth of 19 feet swelled 2.4 percent and had a swell pressure of 6,370 psf when wetted after application of a 2,000 psf pressure.

Additional testing was performed on samples of materials proposed for use as embankment materials, including standard Proctor compaction, swell/consolidation, remolded permeability, pinhole dispersion, unconfined strength, and direct shear.

The soils encountered within the existing pond area were classified as low and high plasticity clays (CL and CH) according to the Unified Soil Classification System (USCS). The index prop-erties of the soils tested are summarized below in Table 2. The percentage of particles passing the number 200 sieve ranged from 51 to 95 percent. Atterberg limits tests indicated a range of liquid limits from 35 to 102, and a range in Plasticity Index from 18 to 63.

A bulk sample was compacted according to the ASTM D 698 method (standard Proctor) to determine the maximum density and optimum moisture content. An additional sample was then compacted to 95 percent of the maximum dry density at optimum moisture content and tested to determine the hydraulic conductivity using the falling head method. A coefficient of permeability of 5.4 × 10−9 cm/sec was determined for the composite sample tested.

Natural, sandy clay and thin, discontinuous sand lenses were encountered below the fill in the pond area. A sample of the sand had 33 percent silt and clay sized particles, a liquid limit of 22 and a plasticity index of 6. Soil observed and sampled in the borrow area was sandy clay. The clay samples had generally similar engineering properties to the fill soil.

Laboratory test results and observations indicate the in-place materials in the pond area and borrow soils located south of the pond site were similar to the existing embankment materials and were suitable for use as liner material for the pond or as fill for the embankment.

Comparatively unweathered claystone and sandstone bedrock were logged at depths of 2 to 4 feet below the current ground surface in the pond area, and at depths of 18 feet to 51 feet below the existing embankment. Although clay and claystone were the predominant soils encountered, sandstone lenses 1.5 to 8.5 feet thick were encountered in two of the exploratory borings drilled into the dam abutments. Test results indicated these materials are suitable as a dam foundation.

Table 2. Pond soil index properties.

Sample Percent passing Liquid Plastic Plasticity USCS Boring I.D. interval (ft) No. 200 Sieve limit limit index classification

MFG-4 2 59.4 33 13 20 CLMFG-5 0–2 66.9 53 23 30 CHMFG-6 2 93.9 77 33 44 CHMFG-7 0–2 94.6 65 25 40 CHMFG-8 4 51.3 34 16 18 CLMFG-9 2 93.1 102 39 63 CHMFG-10 2 91.3 91 30 61 CHMFG-11 2 72.6 36 18 18 CLMFG-11 4 58.7 35 14 21 CL

(from Chapel, 2006).


405

Free water was not encountered in any of the borings. Moisture content of the shallow laboratory samples was generally several percentage points below the anticipated optimum moisture content of the clay soils, but moisture content increased to near or above optimum in the deeper samples. Bedrock samples were logged as slightly moist to moist.

2.1 Slope stability analysis

The stability of the previously designed, as-constructed, and proposed embankment was evaluated at the maximum dam section. The SLOPE/W limit equilibrium program was used to evaluate the factor of safety against failure with the reservoir operating at an elevation 5 feet below the crest of the embankment. Table 3 summarizes the material properties for the existing embankment. The design soil strength parameters used in the stability analysis and described in Table 3 were developed by applying conservative engineering judgment to evaluate the soil properties and test results.

Bishop’s Simplified Method was chosen for the stability analysis for consistency with the previ-ous slope stability studies conducted during initial design. Iterative searches were performed to determine the most critical failure surface of the cross-section that was analyzed. Both static and pseudo-static conditions were analyzed. Using the effective strength parameters estimated from the results of the laboratory testing (design parameters), we calculated minimum factor of safety for the existing dam under normal operating (long-term) conditions to be 3.0.

Pseudo-static analysis was used to model slope stability under seismic conditions. USGS seis-mic hazard mapping and International Building Code (IBC) methods were used to determine seis-mic coefficients for the subject area that ranged from 0.06 g to 0.2 g. Seismic coefficient factors of 0.1 g and 0.2 g were selected for the pseudo-static analyses. With this condition, the minimum factors of safety were 1.8 and 1.3 for the coefficients of acceleration corresponding to 0.1 g and 0.2 g, respectively.

A failure surface that intersects the crest or the upstream slope of the embankment would likely cause significant damage to the dam. Because of this, a rapid drawdown condition in which critical

Table 4. Summary of minimum factors of safety.

Factor of safety Factor of safety (Tt laboratory (Design parameters)Condition Slope type Load condition parameters)

Long Term Downstream Slope Static 3.2 3.0Long Term Downstream Slope Pseudo-Static (0.1 g) 2.0 1.8Long Term Downstream Slope Pseudo-Static (0.2 g) 1.5 1.3Rapid Drawdown Upstream Slope Static 3.0 2.8

(from Chapel, 2006).

Table 3. Summary of material properties used in the slope stability analyses.

Total unit weight Effective strength parameters

Material (pcf) Φ' (deg) c' (psf)

Embankment Clay (Tt Laboratory Parameters) 115 20 1000Embankment Clay (Design Parameters) 100 18 0500Sand/Gravel Drain Material 130 35 0000Claystone (Bedrock) 115 20 1000

(from Chapel, 2006)


406

surfaces intersected the crest and upstream slope of the embankment was evaluated. The minimum factor of safety for these critical surfaces was 2.8. Results indicated the minimum factors of safety for the specified conditions exceeded typical requirements for the design slopes.

2.2 Seepage analysis

Seepage analyses using the computer program SEEP/W, Geostudio (GEO-SLOPE, 2004) were used to determine the seepage through the pond liner and at the toe of the embankment.

Two-dimensional flow was analyzed for steady-state and transient conditions. Using water contents measured at the site, a steady-state analysis was conducted to determine the input con-ditions for the transient analysis. The values of saturated hydraulic conductivity of each mate-rial were selected based on results from the laboratory tests and values in the GEO-SLOPE soils database. The ratio of horizontal to vertical permeability (Kh/Kv) was specified to be 10 in the claystone for the analyses. Table 5 summarizes the material properties used for the seepage analyses.

A transient analysis was conducted to determine the seepage rate through the pond liner over a 20 year period. The model showed that the infiltration rate over the 20 year time period ranged from 5 × 10−9 cfs to 3.5 × 10−8 cfs. Infiltration increased from 0.75 × 10−8 cfs to 3.5 × 10−8 cfs over approximately 11 years and then decreased again to 5 × 10−9 cfs. The volumetric water content in the pond liner increased until it became saturated at the top. The water content in the claystone remained relatively constant.

A transient analysis was also conducted to determine the seepage rate into the toe drain at the bottom of the embankment. The output at the toe drain from the embankment was calculated to be 0.463 cfs.

The modeling indicated that for a dam and liner constructed according to the recommendations in the geotechnical report, seepage characteristics of the liner and earthen embankment would be acceptable.

3 CONSTRUCTION

As a result of the geotechnical and geologic investigations and the computer modeling success, Denison Mines (USA) Corp. decided to re-construct the dam in much the same configuration as it was originally designed and constructed. In January 2007 Denison Mines (USA) Corp retained Jones and DeMille Engineering to incorporate the findings from the Geotechnical Investigation prepared by Tetra Tech into construction drawings and specifications. Plans and specifications were prepared and submitted to the Utah Department of Environmental Quality (UDEQ) Divi-sion of Water Quality and to the Utah State Engineer for review. Both agencies reviewed and approved the drawings and specifications and issued construction permits. Following approval the project was put out to bid. Jackson Excavation from Bicknell, Utah was selected to complete

Table 5. Summary of material properties used in the seepage analyses.

Saturated permeability1 Saturated volumetric Residual volumetric Soil material (cm/s) water content2 water content2

Embankment Clay 1.4 × 10–10 0.34 0.16Sand 9.1 × 10–3 0.39 0.03Claystone 5.1 × 10–8 0.38 0.10Pond Liner 5.4 × 10–9 0.34 0.16

(from Chapel, 2006). Notes: 1Results from Laboratory testing, 2Value from GEO-SLOPE database.


407

the reconstruction of the evaporation pond. Construction of the pond began in September, 2007. Reconstructing the evaporation pond consisted of two major elements, each described below.

3.1 Liner construction

The first step was to recondition the entire pond bottom to ensure that the permeability of the compacted clay pond liner met the UDEQ requirement of a field measured coefficient of perme-ability less than 1×10−7 cm/s.

Quality assurance and quality control testing requirements were developed to guide the contrac-tor in meeting the design intent and documenting the work to fulfill the requirements of UDEQ and the Utah State Engineer. The project specifications outlined the testing requirements that must be achieved throughout the construction process. An independent 3rd party construction quality assurance and quality control (CQA/QC) contractor was retained by Jackson Excavating. Central Utah Testing completed the CQA/QC testing and submitted a CQA/QC certification report for review and approval by UDEQ. Jones and DeMille Engineering (Jones and DeMille, 2007a) pro-vided the acceptance testing as required. Tetra Tech provided additional consulting engineering services during compaction of the liner.

The CQA/QC contractor performed compaction testing in accordance with ASTM D2922 and determined maximum laboratory density in accordance with ASTM D698. Test frequencies for CQA/QC were as described below:

• Compacted Dike Embankment Repair: A minimum of 1 random density test for each 1,500 square yards was required, and each layer placed must be tested. Approximately 100 tests were anticipated.

• Pond Bottom Reconditioning: A minimum of 1 random density test for each 1,000 square yards of reconditioned surface were required. Approximately 80 tests were anticipated.

Acceptance testing included density requirements that ranged from 90 percent to 95 percent of standard Proctor maximum dry density, depending on the location of the tests. Moisture content was specified to be within two percent of optimum water content, determined in accordance with ASTM D 698.

The primary performance specification of the pond liner was that a coefficient of hydraulic con-ductivity of 1x10−7 cm/s or less was achieved in filed tests using Single Ring Infiltrometer tests. Tests on the reconditioned pond bottom were required at a minimum of 1 test per 7,500 square yards. (approximately 12 tests were required). The density and moisture requirements were a means of achieving the required permeability.

The first several days of actual construction were spent determining a suitable method for reconditioning the pond bottom. Several sealed single ring infiltrometer tests were completed to determine the permeability of the reconditioned liner. After several sealed single ring tests and additional laboratory permeability tests it was determined that the process used to recondition the liner was meeting the required permeability.

The clay liner reconditioning process consisted of windrowing the top 12˝ of clay soil, adding water as needed by spraying the windrows using a water truck, mixing and spreading a six inch layer of moisture conditioned clay liner, then compacting the layer using a combination of several passes with a sheep’s foot vibrating roller and then several passes using loaded rubber tire loaders. The process was repeated for the upper layer of clay liner soil.

No additional soil was imported for the liner construction. Where the liner was not 12˝ thick, clay material from an area within the pond that was in excess of 12˝ thick was moved and used in the areas that were lacking.

The high plasticity nature of the soils used in the liner required additional processing to achieve uniform moisture conditions in the fill.

The miners were testing the dewatering system in the mine while the construction was ongoing. For construction use, water from the mine was pumped into a small pond within the larger evapo-ration pond. Water trucks were filled from this location and used for soil conditioning.


408

3.2 Dam construction

The second portion of the project was to repair the breach in the evaporation pond embankment dam. Approximately 14,000 cubic yards of fill was necessary to fill the breach channel in the dam. Approximately 11,000 cubic yards were gathered from the downstream side of the dam. The remainder was obtained from a hillside located south of the dam and within the former pond area. Soil was placed and compacted in six inch thick lifts. Each lift was keyed into the side of the breach channel by benching into the undisturbed portion of the dam on each side of the breach. The embankment repair and recompaction process included:

• moving enough loose soil into the breach area for the required lift;• placing, moisture conditioning, and compacting the lift using a sheep's foot roller and rubber

tired loaders; and• preparing the new base of the fill by cutting into existing dam material on each side of the

breach using a motor grader.

After the pond bottom had been reconditioned and the breach was repaired, riprap was placed on the upstream face of the dam. The riprap was primarily obtained from locations where it was stockpiled during the reclamation efforts. Additional riprap was removed from the original rock quarry located ¼ mile southeast of the pond.

The entire construction process took approximately 60 days to complete and cost approximately $300,000 dollars (Jones and DeMille, 2007b). The short construction time and low construction costs were made possible because of the materials being in or within a short distance of the actual pond.

4 CONCLUSIONS

The arid climate in the vicinity of the Tony M Mine results in evaporation that is greater than the precipitation. Geotechnical modeling showed that infiltration through a native clay liner over a 20 year period would range from 5 × 10−9 cfs to 3.5 × 10−8 cfs. Infiltration would increase from 0.75 × 10−8 cfs to 3.5 × 10−8 cfs over approximately 11 years and then decrease again to 5 × 10−9 cfs. This corresponds to the time at which the pond liner would become saturated on the surface of the liner.

The Tony M evaporation pond was designed and constructed in a small, confined drainage area that is geologically well suited for this application. The naturally occurring, high plasticity clay soil and bedrock made an excellent liner that was both protective of the environment and economi-cally feasible to construct. An additional benefit of this liner is the aesthetic impact of the pond and liner on the surrounding area and the environment. The visual impact of the pond and native liner are much more appealing than would be a synthetic liner. The presence of similar soils nearby for use as construction materials made construction in this remote area more feasible.

The project success demonstrated the benefits that can be realized by a multi-disciplinary approach to engineering problems. Work is currently underway to investigate if similar techniques can be used to develop a larger evaporation facility for phase 2 of the mining operation at Tony M.

REFERENCES

Chapel, T.A. 2006. Geotechnical Investigation: Evaporation Pond Embankment, Tony M Mine, Garfield County, Utah Tetra Tech, Inc., prepared for Denison Mines (USA).

Davidson, E.S., 1967. “Geology of the Circle Cliffs area, Garfield and Kane Counties, Utah, U.S. Geological Survey Bulletin 1229.

Denison Mines (2007). Permits to Design and Construct Evaporation Facilities, various internal files.Jones and DeMille Engineering. 2007a. Construction Plans and Specifications, Tony M Mine.Jones and DeMille Engineering. 2007b. As Built records, Tony M Mine.


409


Gold quarry North Waste Rock Facility slide investigation and stabilization

R.J. Sheets & E.E. BatesNewmont Mining Corporation—Carlin Surface Mine Operations, Carlin, NV, USA

ABSTRACT: On February 5th, 2005, approximately 9.1 Megatonnes of the Gold Quarry North Waste Rock Facility (NWRF) collapsed. The slide covered over 460 meters of Nevada State Route 766 and came within 10 meters of a nearby creek. The slope had been at or near its final design configuration since the early 1990’s (3H:1V). Following an extensive geotechnical inves-tigation, it was determined that the NWRF slide resulted due to strain-softening, fine grained plastic clayey silt that composed the lower 2/3 of the height that failed along the contact with the native ground surface. Instability developed as the clayey silt passed from its peak to residual strength. Stabilization measures to achieve a 5.1H:1V design slope have included unweighting of approximately 11 Megatonnes and constructing a 1.4 Megatonnes buttress. Movement recognized during remediation required design modifications based on current assumptions that the strength properties of the slide surface are two standard deviations less than the average residual strength determined from laboratory tests.

1 INTRODUCTION

Newmont Mining Corporation’s Carlin Surface Mine Operation is located on the Carlin Trend in Northern Nevada. Primary mining activity is centered around the Gold Quarry Open Pit, 11 km northwest of Carlin, Nevada. Large scale mining began at this location in the early 1980’s. During 2007, roughly 56.8 Megatonnes (62,640,000 tons) of material were mined out of Gold Quarry of which 12.6 Megatonnes (13,847,000 tons) were ore that produced over 37,425 kilograms (1.2 million troy ounces) of gold. Current open pit dimensions are approximately 2.1 kilometers north-south by 1.6 kilometers east-west with slope heights varying from 350 to 490 meters. Interramp slope angles vary between 25° and 49° and are dependent on material type and geologic structures. Active support facilities in close proximity to Gold Quarry include: two waste rock storage facili-ties, two heap leach process facilities, one active tailing storage impoundment dam, and one water reservoir.

On February 5th, 2005, approximately 9.1 Megatonnes of material from the North Waste Rock Facility (NWRF) collapsed. The slide covered over 460 meters of Nevada State Route 766 with the toe displacing roughly 183 meters, stopping within 10 meters of a nearby creek bed east of the highway. Prior to the failure, the northeastern facing slope of the NWRF was at a final configura-tion of nearly 3H:1V over a height of 128 meters.

This paper presents the post-failure study to examine the construction history of the NWRF and geotechnical investigation conducted to determine the contributing factors that resulted in the slide. It includes a discussion of material strength properties determined from samples obtained during drilling the lower lifts of the NWRF, the slide mass, and failure surface and the resulting remediation design based on slope stability analysis. Additionally, it details instability issues that developed after the primary remediation activities were complete and how the final slope configu-ration was redesigned based on continued movement.

410

2 BACKGROUND

2.1 Relevant local geology

Mining activity in the Gold Quarry Open Pit has uncovered three distinct rock sequences. The lower plate section is comprised of Paleozoic sedimentary rocks. This includes two Devonian age units: the Popovich Formation which is composed of micrite and calcarenite at the base transition-ing upward to a silty, bioclastic limestone, and the Rodeo Creek Unit consisting of decalcified limey siltstone, siliceous mudstone, and cherty siltstone. The lower-plate Paleozoic rocks are the dominant host of gold mineralization in Gold Quarry.

Volcanoclastic and lacustrine sediments that originated in the Tertiary Period unconformably overlies the Paleozoic rocks. The Tertiary Carlin Formation has been divided into 14 distinct units that are summarized here as four main sections. The bedrock contact is characterized by basal gravel units with swelling clays in some areas. Variably indurated tuffaceous sedimentary rocks that contain gravel lenses are deposited upon the basal gravels. The lower tuff units contain a significant amount of clay and montmorillonite altered tuff. The middle units are composed of partially indurated, interbedded siltstone and sandstone, with minor tuffaceous and gravel lenses. The upper layers consist of variably calcite-cemented sand and gravel debris flows. The Carlin Formation was the major source of waste rock material that was placed at the base of the NWRF; therefore, it is the unit of primary interest to understand in interpreting the cause of the slide. (Harlan et al. 1999; Regnier 1960).

2.2 Construction and reclamation timeline

Construction of the NWRF began in the mid-1980’s. The northeast slope area that encompassed the NWRF slide can be characterized as being constructed in three 24 meters lifts of highly plastic tuffaceous Carlin Formation waste material, for a total height of 72 meters, and a sulfide stock-pile that was built for future processing, but entirely converted to waste. The sulfide waste was between 52 to 55 meters in height. The overall waste rock facility thickness was roughly 128 meters. A brief synopsis of construction and reclamation progress from the 1980’s through to the NWRF slide is discussed in the following paragraph.

The base of the NWRF, Lift 1, was completed to final limits by the middle of 1989, with con-current reclamation commencing in 1990. Lift 2 and Lift 3 were established to their outer limits by Spring 1992; the initial sulfide stockpile was being developed at the same time. Slope reclamation

Figure 1. Gold Quarry NWRF following the slide of the northeast facing slope.


411

of Lifts 2 and 3 was occurring by August 1994, with vegetation developing by Summer 1995. The sulfide waste lift was near completion by September 2001; reclamation activity to cover the sulfide material was conducted in Summer 2003. Vegetation was developing through the cover material by Summer 2004. Figure 2 shows the NWRF as it looked prior to the February 2005 slope failure.

Construction of the NWRF was based on guidelines provided by Newmont’s geotechnical con-sultant in 1989. The original NWRF design provided for an embankment slope that had a slope of 3H:1V with an overall height of 168 meters. The associated slope stability analysis only assumed a facility composed of Carlin Formation material as opposed to one containing dense hard rock waste from the Popovich Formation or Rodeo Creek Unit. The slope configuration at the time of failure did not appreciably differ from consultant recommendations and was approximately 37 meters lower than the theoretical maximum design height.

3 POST NWRF SLIDE FAILURE MECHANISM STUDY

3.1 NWRF slide site investigation

Following the NWRF Slide, an extensive geotechnical investigation was performed to further evaluate the extent of the collapse surface, determine the material properties, assess pore water pressure conditions, and monitor on-going surface displacement. Data collected and interpreted from the investigation was necessary to characterize the conditions that resulted in the slide and back analyze the failure mechanisms as well as investigate long term stability options. A map that indicates borehole and monitoring locations is shown in Figure 3. Eight boreholes were drilled into the failure mass and surrounding area with detailed engineering logs developed for each bore-hole and samples collected for material strength testing. In addition to drilling, six test pits were excavated within and around the slide mass to obtain bulk samples of Lift 1 type material.

Vibrating wire piezometers were installed in boreholes to model the pore water pressure in and around the slide. Piezometers located from 2.4 to 3.7 meters below the contact between the NWRF and native ground did not measure pore pressure. Likewise, piezometers installed within Lift 1 at 9.1 to 15.2 meters above the foundation contact measured no pore pressure. There were mea-surable pressures at the base of Lift 1 equivalent to 0.25 b-bar. This method to model pore pres-sure was utilized because its presence was related to the change in vertical stress and subsequent

Figure 2. Gold Quarry NWRF in Summer 2004. Reclamation activities for all Lifts have been completed. This photograph shows the NWRF configuration prior to the February 2005 slope failure.


412

consolidation of the slide mass rather than upwelling into the NWRF base. Additional data that supported this conclusion was the water levels recorded in nearby monitoring wells indicating that the water table is 9.1 to 12.2 meters below surface topography. In order to establish the loca-tion of the failure plane, inclinometers were installed into some boreholes. Inclinometer surveys confirmed that the failure surface developed in the lowest portion of Lift 1, immediately above the native ground contact. Potential on-going surface displacements were measured by establishing a survey prism network that was manually monitored until the automatic survey equipment was installed. Data from these surveys were used to ensure safety along State Route 766 throughout stabilization and remediation activities.

Samples recovered from the sulfide waste, the three lifts of Carlin Formation material, and the foundation were characterized through standard soil classification tests. These included: grain size distribution analysis; Atterberg Limits, in situ densities, in situ water contents, specific gravity, and x-ray diffraction to identify major and moderate mineral types. Direct shear tests were con-ducted to determine residual shear strengths. Tests were conducted at a low shear rate of 0.00127 cm/min, which required over a day to carry out a single trace, to prevent the build up of excess pore pressure during the experiments. Some samples required four such traces to develop the residual strength. Due to time constraints, a limited number of shear tests were performed; however, those that were performed fell within the strength range reported by Mesri (2003) for samples with similar plasticity indices.

3.2 NWRF slide material properties

The lower two lifts of the NWRF were found to be composed of montmorillonite-rich clays derived from the lower tuffaceous members of the Carlin Formation. This significantly weak,

Figure 3. Plan view of the NWRF. The slide area topography is overlain on the pre-failure surface. The map shows borehole locations for sampling and monitoring installation as well as trench locations. The section lines identify critical cross-section lines for the pre-failure, post-failure, and remediation design stability analyses.


413

high- plasticity, clay-silt material was the predominate factor that resulted in the NWRF slope failure. Comparison of this material to “typical” waste rock dump construction indicates that this material is susceptible to eventual slope instability.

Standard grain size analysis was conducted on material samples collected from the investiga-tion. The average grain size distribution for each lift is shown in Figure 4; the sulfide material that composed the upper portion of the slope is within the range typical of waste rock material. Conversely, the material in Lifts 1, 2, and 3 are predominately silts and clays. This is considerably finer than material found in a typical waste rock facility. Usually, these material types are prone to be weaker and must be analyzed accordingly for slope stability.

Atterberg limits were performed to characterize the plasticity of the fines as well as the sul-fide and foundation material. Results for Lifts 1, 2, and 3 have been plotted on a Casagrande chart as shown in Figure 5. The chart indicates that the majority of the samples classify as high plasticity silts. The average values, including those for the sulfide lift, have been summarized in Table 1. The sulfides lift, being composed of more characteristic rock waste, exhibits a repre-sentative-type liquid limit and plasticity index. High plasticity soils can exhibit strain-softening behavior, a property that reduces materials strength during continual deformation until it is lowered to an inherent residual strength value which is independent of the initial peak strength condition.

Results of a direct shear test carried out on sample 846_Pit collected from the NWRF is shown in Figure 6; the first pass indicates that under primary consolidation the peak shear strength is achieved, whereas, the residual shear strength is reached by the third trace. It is readily evident that there can be a dramatic decrease in strength associated with continual deformation of high plasticity fines. The average Mohr-Coulomb residual strength parameters for Lift 1 material were cohesion of 26.2 kPa and friction angle of 10.9 degrees; the standard deviations were 6.89 kPa and 1.4 degrees respectively.

1000 100 10 1 0.1 0.01 0.001 0.0001

10

0

20

30

40

50

60

70

80

90

100

% F

iner

Gravel Sand Silt Clay

Average Lift 1

Average Lift 2

Average Lift 3

Average Sulfide

Average Rock Pile

Sand Pile

Grain Size (mm)

Figure 4. Grain size distribution graph for material samples collected from the NWRF. The graph depicts that material in the waste rock facility is actually significantly finer than material that composes typical rock or sand-type waste rock facilities.


414

Lift 1Lift 2Lift 3

0 10 20 30 40 50 60 70 80 90 100 110Liquid Limit

Plas

ticity

Ind

ex

10

20

30

40

50

60

70

80

CH

CL

CL-ML

MH or OH

ML or OL

A-Line

Figure 5. Casagrande chart that shows the Atterberg limit test results for Carlin Formation waste material that composes the three lowest lifts of the NWRF. The majority of the samples plot in the lower right are of the chart which indicates that the material can primarily be classified as a high plasticity silt.

Table 1. Summarized atterberg limits of NWRF material.

Liquid limit Plasticity index

Material Mean STD Mean STD

Sulfide lift 11.4 16.5 1.8 2.5Lift 3 73.1 12.5 27.8 8.6Lift 2 80.1 13.2 37.5 8.2Lift 1 83.8 14.9 34.9 11.3Foundation 36.5 19.4 11.8 7.5

Peak Shear Resistance First Pass

Residual Shear Resistance Third Pass

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2Horizontal Displacement (cm)

100

200

300

400

500

600

700

800

Shea

r Fo

rce

(N)

2.98 cm355 N

0.51 cm797 N

Figure 6. Direct shear test results for sample 846_Pit. As a result of continued deformation, the material exhibits significant strain-softening between the first pass and third pass.


415

4 INITIAL STABILITY ANALYSES REVIEW

4.1 Back-analyses of NWRF slide and pre-failure geometry

Based on limited information, a preliminary stability analysis of the post-slide condition was conducted to develop immediate stabilization recommendations that would permit State Route 766 to be re-opened to traffic. As the more thorough site investigation was conducted and residual strength results became available, additional stability analyses of the pre-failure slope configura-tion and the post-failure slope configuration were carried out to determine if the residual strengths were representative of the conditions exhibited by the NWRF.

A long section through the centerline of the failure mass and parallel to the predominate displacement vectors was developed for the post-failure analysis (Fig. 3). Two scenarios were reviewed to evaluate potential variation in residual strengths through the lower lifts and the esti-mated pore water conditions to define a limiting equilibrium condition: 1) assuming all three lifts exhibited the residual strength, the measured water level in the collapse zone, and a pore pressure of 0.25 b-bar exhibited in lower intact Lift 1; and 2) Lifts 2 and 3 still had near their remolded peak strengths, Lift 1 had degraded to its residual strength, and water was present along the backplane (to account for higher Lift 2 and 3 strengths). The stability analysis for Condition 1 resulted in a limit equilibrium condition (i.e. factor of safety approximately 1.0); while Condition 2 required approximately 91 meters of pore water on the back plane. Since the results of Condition 1 were more representative of known conditions and conservative assumptions, it was concluded that the strengths and pore water condition were reasonable for use in stability analyses going forward. A typical cross-section through the NWRF slide area is shown in Figure 7. The diagram identi-fies the pre-failure, post-failure, and remediation design topography (which will be discussed in subsequent sections) as well as identifying the lifts and pore water location.

Stability analyses were carried out for two cross-sections representing pre-failure geometry through the area of the slide and the area north of the slide that exhibited movement; however, it did not fail (Fig. 3). The goal of the analyses was to determine the residual factor (R) (Skempton 1964) which has been defined as the ratio of the difference between peak and average strength to the difference between peak and residual strength. In summary, it represents the portion of the failure surface that would be necessary to have activated residual strengths allowing failure to occur. This was accomplished by iteratively increasing the residual strength distribution along the

Distance (m)

Ele

vatio

n (x

100

0m)

Original Ground Surface

Lift 1Lift 2

Lift 3Sulfides

Cover

Remediation TopographyPost-Failure Topography

Pre-Failure Topography

13o

51 m15o

14o

52 m

Equivalent Water Pressure in Lift 1

0 100 200 300 400 500 600 700

1.5

1.6

1.7

Figure 7. Typical cross-section through the NWRF utilized for stability analyses and remediation design reviews. The pre-failure topography is shown with the dashed line, the post-failure topography is shown with the solid line, and the remediation topography is shown with the dot-dash line and inter-bench angles and bench locations. The water level and lifts are also identified.


416

basal portion of the failure plane. An analysis of the section developed through the failure mass, a residual factor of 61% was determined to develop a limit equilibrium condition. When this section was analyzed with the assumption that residual strengths had developed along the entire failure plane the factor of safety was 0.79; if no pore water was present the factor of safety only slightly improved to 0.85. The section slightly north of the failure mass required a residual factor of 68%. The factor of safety for the fully developed residual strength case was 0.84.

4.2 Forward stability analyses and remediation recommendations

The stabilization of the NWRF required both the failure mass and a portion of the slope to the north of the slide (which had exhibited movement but did not fail) to be remediated with a fac-tor of safety of at least 1.2 and only allow for tolerable displacement resulting from potential earthquake loads. Slope design analyses were performed based on the results obtained from the geotechnical investigation and material testing program as well as the back analysis.

Two necessary unweighting cuts, totaling 1.72 Megatonnes were completed in the months immediately following the slide. Nearly one-half of the material removal was necessary to stabi-lize the area to the north of the failure area. The remainder was required to offset the removal of material at the toe of the slide to re-open State Route 766. Between 9.1 and 22.8 meters of height was removed from the top of the NWRF. This activity increased the factor of safety from the equi-librium condition of 1.0 to 1.1, and raised it significantly higher than the factor of safety of 0.79 prior to failure when assuming complete residual strengths. However, as this unweighting effort had not reached the necessary 1.2 factor of safety additional remediation was necessary.

Following thorough design analyses, specific guidelines were developed to stabilize the area of the slide mass and the unstable area to the north. The failure area was required to have an overall slope angle of 11 degrees (5.1H:1V) with inter-bench slope angles of 15 degrees (3.7H:1V) with no inter-bench heights greater than 30.5 meters. Additionally benches were required to have a 3% grade toward the toe to prevent ponding water. The crest was to be unweighted to stabilize the NWRF behind the slide mass and achieve the overall 11 degree slope angle. These remediation measures would improve the factor of safety to 1.27 assuming the likely pore pressure condition.

The NWRF slope to the north required additional consideration as the existing configuration was more unstable than the slide area. Since achieving a 11 degree overall slope angle along this portion of the NWRF would require significant material removal, it was determined to develop design that required a combination of unweighting and a rock fill buttress across the toe. The buttress was designed to have dimensions of 63 meters wide at the base, 115 meters wide at the crest, and 46.9 meters in height. The buttress was designed to contain approximately 1.36 Megatonnes of oxide rock material. The analysis predicted a final factor of safety of 1.26.

Both areas of the NWRF required additional unweighting to accomplish a stable slope. In order to meet the remediation design specifications approximately 8.16 Megatonnes of additional mate-rial would be removed from the top of the NWRF. Combined with the initial unweighting cut dis-cussed earlier in this section, a total of 9.88 Megatonnes of material were required to be removed to remediate a 9.07 Megatonnes slope failure. The figure below (Fig. 8) depicts the NWRF design that satisfied the initial recommendations to remediate the slide area. The angle of perspective is looking toward the west.

5 RE-ACTIVATION OF THE FAILURE SURFACE

5.1 Additional slope movement during remediation

Primary remediation work on the NWRF began shortly after the slope failure in 2005 and con-tinued through 2006 and into 2007. This included the unweighting, re-sloping and contouring the slide mass, and constructing the buttress. Remediation efforts along the crest and slide scarp were completed in the middle of Summer 2006. However after several weeks of no activity, cracks in


417

the upper portion of the previous slide area were recognized in September 2006. A survey of the observed cracks confirmed that they were located along the scarp trace and slide surface of the February 2005 failure. Slope monitoring prisms were replaced in certain areas, as previous prisms had been removed or lost during remediation work, to define the extent and rate of observed movement.

The visible evidence occurred only along the scarp within the sulfide material in the upper remediated slope. Initial observations and interpretations concluded that the cracks were likely due to continued consolidation and settlement within the slide mass. Monitoring continued through the remainder of 2006 and into 2007, during this time the cracks become more promi-nent along the failure scarp and propagated down the slope following the limits of the February 2005 slide.

During the 1st and 2nd quarters of 2007, there was a significant increase in the displacement and velocity as measured with the robotic slope monitoring system, as shown in Figure 9. Veloci-ties increased from 0.3 cm/day in January to 0.9 cm/day by the end of May. In early February, an inclinometer (85 mm diameter) and additional piezometers were installed within the slide mass to determine the extent of movement and the pore water condition. By the beginning of March, the inclinometer casing had sheared off near the foundation contact. However, the piezometers did not indicate there was excessive pore water building up at the location.

As a result of the accelerated movement rate and identifying that the movement was shearing along the original failure plane, a subsequent review of the NWRF slide area and design param-eters was conducted. Three potential scenarios were investigated: 1) the residual strengths used in the design and slope stability analysis over-estimated the actual strength; 2) the water level within the slide area is overall higher than immediately following the February 2005 slide; or 3) the water infiltrating the slide mass is saturating the sulfide mass and increasing the dead load thereby driv-ing the renewed movement.

Based on recommendations additional water level monitoring was installed in and behind the slide area to characterize the water level. An installation of three piezometers behind the failure plane did not measure significant pore water build up in the native ground below the contact, or above the contact in Lift 1 material. Within the slide mass standpipe installations were completed to measure the water level within the foundation and Lift 1; these also indicated that there had been no significant build-up of water within the area during remediation. These results indicated that the major factor behind renewed movement would likely be related to residual material strengths along the failure surface.

A stability analysis was completed on the slope geometry at that time using strength test results from the post-failure material testing program. When the third lowest residual strength test results

Figure 8. The view looking west into the NWRF slope design to stabilize and remediate the February 2005 failure area. The outline of the slide toe would still be visible on the opposite side of State Route 766.


418

(cohesion = 27.1 kPa and friction angle = 8.96 degrees) were utilized in the analysis it produced a factor of safety of approximately 1.0. Based on these results it is possible to have continued displacements along portions of the original failure surface due to the extreme sensitivity of the residual strength values to the stability of the NWRF.

5.2 Modified NWRF stabilization design

The renewed movement of the February 2005 slide area and the subsequent stability analysis using weaker residual material strengths resulted in modifications to the existing remediation design. Iterative stability analyses were completed based on unweighting the active block along the failure plane with successive 6 meter benches. The purpose was to determine the total amount of material that would need to be removed to increase the factor of safety to acceptable levels. The analyses were carried out using the lower strength parameters than those used in the initial remediation design review. This was done to account for the observed variability of strengths along the failure plane and ensure, based on known information, a stable configuration that there was confidence would remain stable. Along the failure surface, the lower residual strength parameters were cohe-sion of 9.6 kPa and friction angle of 8.8 degrees. These values are approximately two standard deviations less than the average Lift 1 residual strength values from the original study.

Following the review of potential unweighting heights, the analysis results indicated that an additional 1.13 Megatonnes of material should be removed from along the original failure scarp area to increase the factor of safety to 1.26, based on weakened residual strengths. Unweighting activities began in the middle of June 2007. Only a couple days into the unweighting efforts, the slope monitoring system saw an immediate decrease in movement rates. Nearly all prisms installed across the slide mass indicated zero to minor creep within two weeks after unweighting started; however, the unweighting progressed to completion to prevent and mitigate the potential

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

9/1/

06

10/1

/06

10/3

1/06

11/3

0/06

12/3

0/06

1/29

/07

2/28

/07

3/30

/07

4/29

/07

5/29

/07

6/28

/07

7/28

/07

8/27

/07

9/26

/07

10/2

6/07

11/2

5/07

12/2

5/07

1/24

/08

2/23

/08

3/24

/08

4/23

/08

5/23

/08

6/22

/08

Date

ND-30

ND-31

ND-36

ND-29ND-28

ED

M C

umul

ativ

e D

ispl

acem

ent (

m)

Figure 9. Graph showing the cumulative movement of five slope monitoring prisms installed on the NWRF. The graph indicated the acceleration observed through the first half of 2007; and the immediate stabilization with the additional unweighting cut that was initiated in June 2007.


419

for reactivation of the slide area again in the future. The post-additional unweighting NWRF slope configuration is shown in Figure 10. The area has been contoured to achieve an approximately 3H:1V slope which was determined to be acceptable in the stability analysis as this portion of the slope is predominately in sulfide material that was behind the limits of the February 2005 failure surface.

5.3 Localized movement of the NWRF Toe

Throughout the unweighting project, the majority of the monitoring prisms indicated that the slope was being stabilized. However, two prisms located in the toe of the slope continued to measure a slight creep of approximately 1.25 cm/month (Fig. 9). During Spring 2007, cracks and heaving of the toe area had been observed and assumed to be primarily related to the movement of the entire mass. Upon observing continued movement after unweighting activities had ended, it was apparent that a portion of the toe was creeping independently of the entire slide mass. Four trenches were excavated along the observed limits of the portion of the toe exhibiting movement to examine if it was surficial, or displacing along the failure plane. The upper trenches exposed a moderate to steeply dipping backplane while the trenches excavated through the toe indicated that there was still movement along the failure contact surface.

A review of the current geometry showed that the lowest inter-bench had not yet been recon-figured to meet provided recommendations; it was still slightly steeper and higher following the efforts to clear and re-open State Route 766. Currently, the toe area has a height of 43 meters and an angle of 17 to 18 degrees. Both exceed the 30.5 meters and 15 degree limits. Slope stability analyses suggested that, based on the reduced failure surface strengths, the toe area had a local factor of safety of 0.99.

Three options were reviewed to stabilize the lower portion of the slope: 1) a localized unweight-ing cut above the toe area; 2) a surficial buttress along the toe of the slope; or 3) a small keyed buttress along the toe of the slope. Because of the location of the slide area in relation to the high-way, construction of a large buttress to stabilize the February 2005 slide was not an option due to limited space. In this case, since a potential buttress would only be stabilizing a small portion of the overall slope, it was deemed a viable option. The options were required to achieve a factor of safety approaching 1.3 and not encroach on the right-of-way along the highway.

The first two options were determined to be non-viable as they would not reach the desired factor of safety without compromising the remainder of the NWRF. The localized unweighting

Figure 10. The view looking west into the NWRF slope following the Summer 2007 unweighting cut along the February 2005 slide scarp. The benched portion of the slope in the middle of the picture where the unweighting occurred has been contoured to an approximately 3H:1V geometry.


420

could stabilize the toe; however, this removal of material would destabilize the upper portion of the February 2005 failure. The surficial buttress would not consolidate the majority of the buttress on the toe to resist movement. An analysis that incrementally increased the surficial buttress found that at a height of 12.2 meters the buttress starts adding a surcharge load to the creeping toe area and would tend to induce failure if the buttress was constructed higher.

Because of the restriction to design a buttress that would fit within the existing limits of the NWRF in the area of the slide, the option to construct a keyed buttress actually would require the removal of existing material at the toe of the failure area. The design also assumed that the key must be excavated to 1.5 meters to ensure it is well into the foundation. Similar to prior stability analyses, the study was conducted by increasing the height of the buttress in increments (3 meters for this case) to identify the optimal configuration. The review determined that a buttress height of 12.2 meters above native ground and average width of 18.3 meters would increase the factor of safety to 1.33.

Currently construction of this buttress along a 250 meter span of the NWRF toe is planned to commence and finish in Summer 2008. The critical factors in construction will be to only excavate and expose a limited slot at a given time, presently expected to be 45 meters. The slot then would be backfilled to construct the buttress before excavating the next slot. This requirement is neces-sary due to the strength sensitivity of the material along the failure plane. Removing a portion of the toe could inherently destabilize that portion of the NWRF; therefore, it will be critical to keep the exposure time to a minimum. Over the past several months a stockpile of oxide rock waste has been constructed to ensure that there will be quality material available in sufficient volumes to complete the construction in a timely manner.

6 CONCLUSIONS

The slope failure of the NWRF was an unexpected event as the facility had been constructed following the provided geotechnical recommendations and was well below the theoretical final design height. The fact that the area had been recontoured and reclaimed almost provided a per-ceived sense of security that conditions did not exist that would develop into a potential instability. At the time of the NWRF slope failure, portions of reclaimed facilities were not included as part of a more regular, routine geotechnical inspection program.

As the preliminary findings of the post-failure investigation began to focus on the presence of high-plasticity, clayey silt derived from the Carlin Formation at the base of the NWRF, there was an urgency to evaluate any other facility that contained the Carlin Formation as a significant con-struction material. This not only included other waste rock facilities, but also tailing storage and heap leach facilities as well. However, upon review it was determined that there was an important difference in the manner that this type of material had been utilized in the construction of tailing and heap leach facilities. These were constructed following detailed engineering and design speci-fications that typically called for short, compacted lifts of fine material. Waste rock facilities are

Figure 11. Photo looking southeast along the strike of the NWRF slope.


421

generally constructed in larger 15 to 30 meter lifts. Material is end-dumped to extend a lift without specified compaction, other than the heavy equipment traffic.

Additionally, facilities that are not active or in reclamation were now added to a comparatively (to previous practices) more regular inspection schedule with documentation. Slope failure does not occur without prior warning or evidence. Since these facilities typically do not require more rigorous or real-time slope monitoring methods typically applied in open pits, visual inspections become an even more critical tool for continually evaluating slope performance.

The NWRF February 2005 failure was the result of non-typical waste material being used in the construction of the lower lifts. Furthermore the high-plasticity clay silt material from the Carlin Formation that was being placed had inherent properties that would have made it certainly susceptible to instability. Thorough geotechnical investigations and material strength testing are necessary to understand the engineering properties and behavior of this type of material when it is used in construction.

REFERENCES

Harlan, J.B., Harris, D.A., Mallette, P.M., Norby, J.W., Rota, J.C., & Sagar, J.J. 1999. Geology and Mineral-ization of the Maggie Creek District. Gold Deposits of the Carlin Trend. Nevada Bureau of Mines and Geology Bulletin 111: 115–142.

Regnier, J. 1960. Cenozoic geology in the vicinity of Carlin, Nevada. Geological Society of America Bulletin, 71(8): 1189–1210.

Mesri, G. & Shahien, M. 2002 Risidual Shear Strength Mobilized in First-Time Slop Failures. Journal of Geotechnical and Geoenvironmental Engineering ASCE, 129(1): 12–31. [Jan. 2003]

Skempton, A.W. 1964. Long-Term Stability of Clay Slopes. Fourth Rankine Lecture. Geotechnique 14: 77–101.

Barkley, R.C., Pryor, P.R., & Nicholas, D. 2006. Gold Quarry Slide Geotechnical Investigation Failure Mech-anisms Report. Call & Nicholas, INC.



423


An overview of the Grouse Creek Mine tailings impoundment closure

Falma MoyeHecla Mining Company, Challis, ID, USA

Steve Rogers & Don PoulterWater Resource Engineer, Water Management Consultants, Inc., Denver, CO, USA

Brant TritthartGrouse Creek Unit, Hecla Mining Company, Challis, ID, USA

ABSTRACT: The Grouse Creek gold-silver deposit in central Idaho operated from 1994 to 1997. Site reclamation activities began in 2000 after permanent closure was announced. Since 2000, approximately 60% of the site has been reclaimed. Tailings impoundment reclamation began in 2006 with consolidation studies. The tailings impoundment reclamation plan includes placement of base fill over tailings, encapsulation of tailings with a synthetic liner, and placement of 4 foot earthen cover over the liner. Surface water will be managed through an energy dissipator, and the reclaimed impoundment surface will serve as a floodway for storm waters metered through a permeable weir to receiving streams. In addition, a clay cover to reduce infiltration will be placed over the south embankment surface.

1 INTRODUCTION AND BACKGROUND

The Grouse Creek mine, an epithermal gold silver mine operated from 1994 to 1997 in the Yankee Fork District of central Idaho (Figure 1). Mine production began at 7,500 tons of ore per day in October 1994. Through 1997, mining operations generated approximately 4.4 million tons of

Figure 1. Site location map.


424

tailings and 15 million tons of waste rock. Crushed ore from the mine was delivered to the mill, where it was further reduced in size by grinding. The milled ore was then processed using a two-stage wash circuit and a Carbon-in-Pulp cyanide leaching process. The tailings and water from the ore processing were placed in the tailings impoundment. Waste rock was placed in a storage facility located on the northwest side of the tailings impoundment. The site went into temporary shutdown in 1997 and permanent closure in 2000. The site has been approximately 60% reclaimed with the tailings impoundment being the last major component of reclamation.

1.1 Tailings impoundment overview

The tailings impoundment was constructed on top of an historic landslide complex and was to be constructed in three phases. Geotechnical investigations led to the design of a key trench for the South Embankment to ensure long-term stability of the tailings impoundment at final build-out. Phase I of the tailings impoundment construction occurred from 1993 through 1994 and com-pleted the North and South Embankments to the 7,212 elevation. Phase 2 construction occurred from 1995 through 1996 and completed the facility to the current 7,250 elevation. Phase 3 con-struction, which did not occur, would have completed the facility to the 7,290 elevation. Both embankments are composed of locally excavated minus 24 inch borrow materials and waste rock. The maximum height of the fill materials is about 135 feet at the North Embankment crest to over 200 feet at the South Embankment crest. Following placement, the face of each embankment was compacted to reduce infiltration. The Idaho Department of Water Resources (IDWR) requires that the impoundment be capable of containing the runoff from the probable maximum flood (PMF), plus additional freeboard for wave action.

The tailings impoundment covers a surface area of approximately 67 acres. Tailings were ini-tially placed into the impoundment using perimeter cyclones to segregate the tailings and deposit coarser-grained cyclone underflow above a basal gravel drainage layer to create a blanket drain. The basal gravel blanket is 14-inches thick and was installed up to the 7,150 elevation. The fine slimes, or cyclone overflow, were deposited above the cyclone underflow to form a seal over the higher permeability drainage layer. The cycloned tailings underflow was placed against the inte-rior perimeter of the southern half of the tailings impoundment to form a beach.

Prior to the placement of the tailings impoundment liner, a drainage system consisting of fifteen (15) underdrains plus feeder underdrains was constructed to capture groundwater beneath the lined impoundment. Groundwater and tailings seepage that enter the underdrains are conveyed to a lined collection pond located below the South Embankment.

Pore water from the tailings drainage layer above the liner system was collected in a sump in the southeast corner of the impoundment. The drainage layer sump, or floor drain, was designed to continuously pump pore waters from the drainage layer to the surface of the impoundment to reduce the head on the liner, minimize seepage to the underdrains and promote consolidation. Floor drain pumping was discontinued in January 2006 because self-weight consolidation of tail-ings had effectively reduced the volume of pore water that could be pumped.

The tailings impoundment was constructed with a composite liner system consisting of a mini-mum of 12-inches of a low-permeability soil liner (maximum permeability of 1 × 10–6 cm/s) over-lain by a 60-mil very low-density polyethylene (VLDPE) synthetic liner. Geomembrane seams were secured with double fusion welds with an air pocket between the welds to allow for pressure testing of the seam.

The final placement of tailings into the impoundment was in 1997. The impoundment contains approximately 4.4 million tons of tailings. On July 1, 2003 there were approximately 522 million gallons of water with the surface level at the 7,231.5 elevation. As a result of a dewatering program implemented in June 2003, more than one billion gallons of primarily storm water and snow melt water had been discharged by May 2008 when the water surface was at approximately the 7,204 elevation.

Surface flows and storm water upgradient of the impoundment are diverted around the impound-ment for discharge to receiving streams.


425

1.2 Goals of final reclamation

The goals of the final reclamation for the Grouse Creek Unit tailings impoundment include the following:

• Decommission the jurisdictional embankment and impoundment based on design parameters provided by Hecla, dam safety criteria under IDWR applicable rules and regulations, and site-specific criteria.

• Route the 500-year, 24-hour runoff event from upgradient drainages through or around the tail-ings impoundment to Jordan Creek.

• Convey the 6-hr PMF peak flow across the tailings impoundment final cover surface while protecting the North and South Embankments from overtopping.

• Attenuate the 96-hr PMF total volume on the tailings impoundment final cover surface through Pinyon Hill porous rock berm spillway.

• Provide for possible passive treatment in the north end of the tailings impoundment.

Reclamation of the tailings impoundment began in 2006 with construction of a land bridge to test consolidation. In 2007, additional base fill material was placed and a second land bridge was constructed. An engineered work plan was prepared to comply with agency requirements (Hecla and Water Management Consultants, 2007 a,b,c) and was recently approved by regulatory agencies. Figure 2 provides a photographic comparison of the tailings impoundment between 2004 and 2008.

2 ENGINEERING EVALUATIONS

Engineering analyses, design studies, and evaluations for the tailings impoundment included the following:

• Tailings characterization,• Tailings cover grading plan development, and• Surface water management through the impoundment area to Jordan Creek.

2.1 Tailings characterization and consolidation

Tailings characterization and consolidation properties were completed to determine the current properties of the tailings and the impact of constructing the cover over tailings. The primary goal was to estimate the amount of settlement/consolidation that would occur in the tailings due to the

Figure 2. 2004 and 2008 comparison photos.


426

base and surcharge fill required for grading. Factors that contribute to the settlement/consolidation analysis include:

• Physical properties, depositional history, and thickness of the tailings,• Magnitude of cover loads,• Current in-place tailings density, and• Amount of consolidation remaining in tailings due to self-weight loading conditions.

The analyses were divided as follows:

• Tailings properties and densities (average and versus depth) estimates based on tailings deposi-tion history and laboratory consolidation tests on tailings samples.

• Data calibration and model development to simulate current consolidation.• Current tailings consolidation estimates and amount of settlement remaining under self-weight

consolidation conditions.• Settlement/consolidation of tailings estimates due to weight of additional fill and cover

placement.• Settlement estimate refinement and projection of time versus settlement based on the final

design and material balance of the cover.

These analyses and evaluations resulted in the proposed tailings cover grading plan.

2.1.1 Tailings characterizationBased on historic data and testing for this design, the tailings deposited into the impoundment are characterized as three materials:

• Whole tailings—the tailings particle gradation resulting from the crushing and grinding of the ore for processing,

• Cyclone sand tailings—the coarse-grained fraction of material separated from the whole tail-ings by use of cyclones (referred to as the cyclone underflow), and

• Fine tailings—the fine-grained fraction of material separated from the tailings by use of cyclones (referred to as the cyclone overflow).

Design studies for the tailings considered only whole tailings consisting of approximately 20% fine sand-sized particles. The near surface fine tailings were tested for gradation and essentially 100% (by weight) of the material passed the #200 sieve size (0.074 mm). A seepage-induced con-solidation test (SICT) was also performed on the fine tailings to determine the typical consolida-tion characteristics for comparison with the whole tailings.

The compressibility parameters of the whole and fine tailings are similar indicating that densi-ties resulting from self weight and additional loading consolidation would be similar. The differ-ence is in the estimated permeability coefficient. The fine tailings exhibit a coefficient one order of magnitude greater than whole tailings. This indicates that while the two materials may exhibit similar densities, it will take a longer time for the fine tailings to consolidate.

2.1.2 Consolidation modelThe consolidation model is based on the primary assumption that the whole tailings deposited dur-ing initial operations and the fine tailings deposited from the cyclone overflow govern the consoli-dation characteristics of the in-place tailings mass and the magnitude of settlement/consolidation over the majority of the impoundment area. This includes the thickest deposits of tailings along the central axis of the impoundment.

Deposition and tailings depths data obtained from the annual surveys were used to reconstruct the height-time field data for a column of tailings in the impoundment. The field data during the filling stage were then used to find the amount and type of material deposited in the area of the selected column (the column was selected based on maximum depth and minimal influence of sand deposition from the cyclone tailings). Figure 3 shows the correlation between field data and the filling history used in the analyses.


427

The predicted and actual settlement history after 994 days in Figure 3 indicates that while the settlement rates are different, the overall agreement is reasonable. This supports both the material quantities and the material characteristics used in the analysis. Based on the initial model, the tail-ings will continue to consolidate under self-weight loading conditions. The additional self-weight consolidation is estimated to be approximately 4.5 feet over a nine year period following the August 2006 tailings surface survey, with up to 0.5 feet additional consolidation over the follow-ing eight years. Review of the surface surveys to date and extrapolation of the data points reveals a reduced self-weight consolidation of the tailings to the point of ‘completion’ and the average settled density of the tailings is approximately 82.5 per cubic foot (pcf).

2.1.3 Consolidation estimatesThe primary assumptions set in estimating the consolidation curves are:

• The consolidation model based on the fine and whole tailings characteristics and the annual tailings surface surveys is valid.

• Deposition of sands was such that sand tended to displace fines along the liner contact and cre-ate a wedge of sand at the perimeter of the tailings in the area of discharge; therefore, a delta of sand did not extend over the fine tailings and impact settlement/consolidation of the fines.

• A separate model for estimation of sand consolidation is not required for design of the tailings cover. The extent of the cycloned tailings sand is minimal and will not significantly influence the overall settlement/consolidation of the proposed cover. Therefore, use of the single whole/fine tailings consolidation model is reasonable.

• Settlement of measurable significance will only occur in the tailings underlying the cover. The tailings cover material will be placed in a manner such that the materials placed over the tailings will be sufficiently dense at the end of construction and not exhibit measurable settlement.

Due to the shape of the impoundment and distribution of tailings in the impoundment, it was necessary to develop consolidation curves that represented a range of tailings depths versus added loads from the proposed tailings cover. Initially, the tailings thickness was modeled in ten columns with each increasing in equal increments up to the maximum depth of the tailings. A series of loads was then applied to the columns in increasing increments up to 40 feet of fill (placed at a unit

0

20

40

60

80

100

120

0 1000 2000 3000 4000 5000 6000 7000 8000

Time (days)

Heig

ht

(ft) Field data

Model prediction

Figure 3. Field data vs estimated consolidation model.


428

weight of 120 pcf). Settlements were then estimated in each column for each loading increment based on the tailings consolidation characteristics obtained from the fine tailings SICT. Results indicate that reasonable linear interpolation existed between the sets of data such that only five increments of tailings depth for the Load vs. Settlement were required to estimate settlement under the load of the proposed tailings cover. Figure 4 shows the resulting curves used to estimate the tailings settlement and develop the tailings cover design.

Tailings settlement was then estimated based on the surcharge load applied by the fill required to achieve the design grade over the existing tailings surface. Additional fill was then added over the areas of settlement to reestablish the design grade elevation, and settlement from the increase in fill thickness was estimated. This iteration was continued until settlement versus fill thickness was within +/− 0.1 feet of the design grade. These estimates were based on a grid of points set over the impoundment area relative to the design grade and specific areas of concern based on the tailings depth and estimated fill thickness that did not fall within the grid pattern. The final fill estimates were converted to elevations to create a fill and grading plan.

2.2 Tailings cover design

Cover design criteria related to the tailings embankment stability and safety under the IDWR rules and regulations for tailings embankments include:

• Protection from the PMF runoff from the catch boundary above the impoundment and other areas adjacent to the embankment structure.

• Maintain a minimum allowable Factor of Safety (FOS) against slope failure in the embankment and adjacent abutment under static and Maximum Creditable Earthquake (MCE) loading condi-tions as follows:

– Static Loading FOS = 1.5 minimum allowable value, – Pseudo-Static Loading FOS based on site bedrock acceleration from the MCE = 1.0 mini-

mum allowable value, and – Crest deformation from MCE loading conditions not to result in a breach of the crest and

discharge of the tailings in the event the minimum allowable pseudo-static FOS cannot be met, or as otherwise directed by IDWR regardless of the pseudo-static FOS.

• Design the final cover and spillway outlet such that the residual pool remaining at the end of the ‘dry’ season is less than 50 acre-feet.

The tailings impoundment cover was designed to allow for settlement/consolidation of the tail-ings. The final grade of the cover in the southern pond area is set to approximately 0.5% draining

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

10.00

0 1000 2000 3000 4000

Fill Load (pcf)

Sett

lem

en

t (f

t) Tailings = 96.6 ft

Tailings = 79.6 ft

Tailings = 62 ft

Tailings = 42 ft

Tailings = 20ft

Tailings = 7 ft

Figure 4. Fill load vs settlement.


429

east toward the Pinyon Hill spillway near the right abutment of the North Embankment and 0.5% draining east towards the controlled outlet in the north pond area.

2.3 Stability evaluations

The stability analyses for the design and construction of the North and South Embankments meet or exceed all applicable requirements and minimum FOS stability estimates required by the IDWR for tailings embankment safety (Golder 1992a). The design report and the Phase I and II as-built reports for the two embankments were reviewed as part of the project work (Golder 1994 and 1997 respectively). A site reconnaissance conducted in October, 2005 found no design issues or site conditions that indicated a need to re-evaluate the stability of the North and South Embankments. The current design modifications increase the embankment stability, and the spillway location and discharge channel alignment do not impact either embankment; therefore embankment stability analyses were not recomputed to demonstrate compliance with IDWR rules and regulations.

2.4 Hydrologic/hydraulic evaluations

Runoff water control facilities such as channels and energy dissipation structures were designed for the 500-year, 24-hour storm event. The reclaimed impoundment and associated spillway were designed to attenuate and discharge the runoff generated from the Probable Maximum Precipita-tion (PMP) event.The reclaimed tailings surface is divided into two areas:

• The north pond area will potentially be used as an aerobic polishing pond for passively treated waters to be discharged through a permitted NPDES outfall.

• The south pond area will serve as a floodway to route surface flows from contributing water-sheds across the impoundment surface and through the Pinyon Hill porous rock berm spillway, a controlled outlet into a channel that discharges to Jordan Creek.

2.4.1 Hydrologic modelThe Hydrologic Modeling System (HEC-HMS 3.1.0) was used to estimate peak discharge rates and flows for the design events evaluated. The unit hydrograph method incorporating the SCS Type II storm distribution with a total rainfall depth of 4.7 inches for the 500-year, 24-hour fre-quency event was simulated. The 6-hour local storm PMP event was used to estimate the peak design flow for diversion purposes, and the 96-hour general storm PMP event was used to estimate the peak design volume of runoff from a PMP storm event for storage purposes on the reclaimed impoundment surface.

2.4.2 Design levelsUsing the HEC-HMS model results, the appropriate design values for each component of the proposed surface water diversion plan system are presented in Table 1.

Table 1. Hydraulic structure design levels.

Surface water diversion structure Design level

Washout Creek Energy Dissipation Inlet Channel 500-yr 24-hr peak flow = 927 cfsEnergy Dissipation Structure at the Tailings Facility 500-yr, 24-hr peak flow = 1,446 cfsSouth Pond Area 96-hr PMP peak volume = 549 acre-feetPinyon Hill Spillway Spillway was designed to drain the 96-hr general PMP storm volume over approximately 30 daysOutlet Channel to Jordan Creek Peak flow through spillway ∼12 cfs


430

3 DETAILS OF CLOSURE PLAN

The major elements of the Closure Plan are outlined and described below. These elements consist of:

• Place base and surcharge fill and grading plan over existing tailings,• Encapsulate tailings and base fill with synthetic liner and four feet of earthen material,• Manage upstream runoff and runoff from reclaimed impoundment surface,• Embankment modifications including regrading and cover, and• Construction of seasonally wet meadow and potential passive treatment cell at the north end of

the reclaimed surface.

3.1 Base and surcharge fill over tailings

Material from on-site borrow sources will be placed over the tailings surface to provide a working foundation for equipment. Surcharge fill will compensate for additional settling and consolidation of tailings to ensure that the design surface contours are established after reclamation is complete. In addition, a clay or fine bedding layer will be placed over the top of fill to cushion the overlying synthetic liner. The cushion layer will become part of a composite cover along with synthetic liner over tailings.

3.1.1 Borrow material sourceBorrow material for construction of the base fill and surcharge fill will be obtained mostly from on-site (local) borrow areas. Some materials will require screening or selective borrow excava-tion in order to obtain the specified material. Table 2 contains a summary of the borrow sources, the material use, and estimated quantities available for borrow. In addition, some tailings will be removed from the perimeter of the impoundment and placed within the base fill. However, the quantity of tailings will be on the order of 33,000 cubic yards (yds3) which is not a significant volume compared to the total base fill required.

3.1.2 Materials placementProcedures for material placement will vary depending on conditions in the area being worked. Two critical areas include exposed and submerged tailings surfaces.

Base fill materials will be placed to minimize buildup of ‘mud waves’. Some tailings distur-bance is expected as the initial lifts of base fill are placed; however, the generation of significant mud waves in the tailings will impact the amount of tailings settlement/consolidation, the amount of fill required to achieve the target fill elevations, and the overall contour of the finished grade. Mud wave generation will be a function of the tailings characteristics, rate of fill placement/advancement over the tailings surface, thickness of the fill layer (or lift) placed over the tailings surface, method of fill placement over the tailings surface, and type of equipment used. Rapid fill placement usually generates large mud waves that create construction problems and impact the settlement/performance of the fill.

Table 2. Borrow materials and quantities.

Borrow source Material use Estimated volume (yds3)

Pinyon Hill Cover fill, base fill, low permeability 1,500,000 compacted fill, small diameter riprapSouth Embankment crest Base fill 165,000and lower slopeNorth Embankment Base fill 250,000Energy Dissipator Excavation Cover fill, base fill 75,000Clay Stockpile Low-permeability compacted fill 60,000


431

Base fill placement plans and procedures were developed in conjunction with the contractor. The plan includes methods of fill placement over exposed and submerged tailings, a sequence for fill placement over the tailings surface, and the type of equipment to be used for fill placement. Once the tailings are covered and a firm surface is achieved, fill placement has been achieved as outlined in the specifications for the different materials and work areas.

3.2 Tailings encapsulation, liner placement and earthen cover

The approved cover to isolate tailings from surface water and oxygen includes using a syn-thetic liner overlain with a 4-foot-thick soil cover. The proposed cover design meets the intent of the approved plan and provides safety factors for the performance of the proposed cover including:

• A cover design that accommodates settlement of the tailings and maintains positive drainage following long-term consolidation of the tailings.

• A low-permeability base fill material serves as a liner cushion against punctures and tears and forms a composite liner system with the overlying synthetic liner.

• The selected synthetic liner will be 60-mil-thick low-linear density polyethylene (LLDPE), which is resilient to punctures and tears during handling, installation and overlying fill placement.

• A thorough and detailed quality assurance plan, which ensures good-quality installation of the synthetic cover liner, suitable materials above and below the liner, and careful placement of fill over the synthetic cover liner.

A buried drain will be constructed below the cover liner as a precautionary outlet to remove any build up of water in tailings following construction of the cover liner. The phreatic water level in the tailings will remain at the level present at the time of fill placement and construction of the final cover. The level will remain constant assuming the impoundment will function as a closed system and no infiltration occurs into the lined tailings.

3.2.1 Existing linerThe existing impoundment liner will be left in place and serve as the lower liner to contain the tailings. Construction specifications highlight preservation of the liner integrity and include direc-tions for protection and repair in the event the cover or liner becomes damaged. Once the base fill is placed, fill subgrade established and the top surface prepared for liner placement, the liner extending above the fill contact will be carefully removed from the impoundment slopes and folded over onto the fill subgrade surface.

3.2.2 Synthetic cover linerThe upper one foot of the base fill will be prepared and machine-traffic compacted to a smooth surface suitable for liner placement. This upper one foot of material will consist of a soil-like material obtained from designated borrow sources or tailings excavated from the impoundment perimeter as required in the grading plan. The intent is for the upper one foot layer to consist of a silty to clayey soil matrix that contains no particles greater than 2.5-inch diameter, 80% or less no greater 1-inch diameter, and at least 50% (by dry weight) no greater than the No. 100 sieve size (U.S. Standard Sieve Size). It is anticipated that such material will act as a low-permeability material directly below the overlying synthetic liner.

The synthetic cover liner will be deployed over the prepared fill subgrade surface and folded liner, and the seams field welded to form a continuous membrane over the area of the tailings and fill subgrade. The cover liner will extend about 3 to 5 feet beyond the limits of the folded liner and secured in an anchor trench with compacted fill. The cover liner will extend into a cutoff trench, which will block migration of groundwater from the surrounding slopes towards the impound-ment. Once the soil is placed over the cover liner, the weight of the fill will serve to press the contact surface between the liners to function as a seal between the liners.


432

3.3 Upstream runoff management

Runoff from upgradient drainages will enter an energy dissipation structure through a constructed channel before discharging to the south pond (Figure 5). This structure will dissipate the runoff energy and transition the flow from a super-critical to a sub-critical flow regime during the 500-yr, 24-hr storm event. A surface swale will convey discharge from the energy dissipation structure across the south pond to a spillway designed to provide temporary storage and controlled release of flood surges but would function as a shallow flow-through pond during normal conditions. This design allows any sediment and debris transported during all flow conditions to settle out on the reclaimed surface, minimizing sediment delivery to Jordan Creek. Below the spillway, runoff will flow in an engineered conveyance structure to Jordan Creek.

The south pond will be sized to store the 96-hour PMF runoff volume and the Pinyon Hill spill-way will be designed to slowly drain the retention pond to Jordan Creek. The spillway structure consists of a porous rock berm designed to release the 96-hr general storm volume over approxi-mately 30 days at an approximate max flow rate of 12 cfs.

The spillway consists of gravel core zone topped by a rock berm designed to release the maxi-mum flow rates at the required levels (Figure 6). The berm will be constructed in bedrock on the right abutment of the North Embankment. The cut will be narrowed in the vicinity of the core of the rock berm in order to minimize the flow rate through the porous gravel and rock fill. The upstream slope will consist of graded rock riprap at a slope of approximately 2H:1V (horizontal:vertical). The crest width will be a minimum of 15 feet and include a minimum 8-inch compacted aggregate base course. The downstream slope will consist of larger rock riprap on a 3H:1V slope.

Figure 5. Cross section of energy dissipator area.

Figure 6. Cross section of porous rock berm spillway.


433

3.3.1 Outlet channel designThe outlet channel from the porous rock berm spillway will be constructed within bedrock. The side slopes will generally be 0.5H:1V with a bottom width of approximately 12 feet. The longitu-dinal slope of the channel will be approximately 2.4 percent. A 36-inch culvert will convey flows from the outlet channel and from existing waste rock channels across the existing access road to the outlet channel to Jordan Creek. A riprap-lined channel constructed on the upslope side of the access road will convey flows from the outlet channel (as well as local runoff from upslope areas) to Jordan Creek. A grouted rock riprap energy dissipation structure will reduce the velocity and erosive force of the water prior to entering Jordan Creek.

3.4 Embankment modifications

The North and South Embankments will be modified to obtain fill material for construction of the tailings impoundment base fill. An infiltration cover will be placed over the downstream face of the South Embankment. The modifications meet or exceed tailings dam safety standards for tailings impoundments and include:

• Lowering the crests of the embankments to elevation 7,230 feet,• Flattening the overall downstream slope of the South Embankment from 1.5H:1V to 2.6:

H:1V, including one catch bench to intercept and route surface runoff from the face of the embankment,

• Constructing a catch bench on the South Embankment at approximately elevation 7,100 feet to capture runoff from the upper face of the embankment

• Constructing a French drain system below the lower catch bench to capture seeps that emerge from the toe of the South Embankment, and

• Constructing an earthen cover over the downstream slopes of the South Embankment to reduce infiltration.

The material excavated from the lower South Embankment slope (approx. 165,000 yds3) will be used as base fill over the tailings surface. The lower embankment crests at elevation 7,230 feet were based on the south area topography at the completion of the final grading and the estimated 96-hour PMF of 549 acre-feet that enters the south pond. The south pond stage-storage curve shows that the embankment crests can be lowered to elevation 7,230 feet with approximately one foot of dry freeboard above the peak design water level. Within five years of construction, the tailings are estimated to settle such that the dry freeboard during of the 96-hour PMP event will be approximately two feet.

Following excavation and regrading of the lower embankment face, a 2-foot-thick layer of silty-clay will be placed and machine traffic compacted on the face of the embankment, the terraces and the crest. This material will then be protected with a 12-inch layer of minus 3 inch angular rock, compacted in place. The two terraces, which will be constructed at a 0.5 percent slope towards the right abutment, will feed into a runoff conveyance channel. The drainage channel will consist of a trapezoidal open channel with a bottom width of 3 feet, 2H:1V side slopes and a finished depth of 18 inches. Geotextile fabric will be placed in the base of the channel, overlain by a 6-inch-layer of 3-inch median size crushed angular rock. This will then be covered with a 2-foot-layer of 12-inch median size crushed angular rock.

The North Embankment crest will be excavated down to a finished elevation of 7,230 feet. In addition, fill material will be removed from the north face and the crest will be rounded leaving a final crest width of approximately 30 feet. The existing downstream toe elevation of the dam is approximately 7,200 feet, and the final maximum embankment height will be approximately 30 feet. As the slope length of the finished embankment will be less than 150 feet, no cross-cut terraces are proposed for the face.

The North Embankment will be covered with approximately 1 foot of growth media and hydro-seeded using an approved seed mix.


434

3.5 North Pond seasonally wet meadow and outlet pipeline

The north pond area will encompass approximately 25 acres with a minimum 30 million gallon surge capacity and will potentially serve as a polishing pond for passively treated waters. The fill subgrade or design surface grading plan is based on the overall impoundment area grading plan. A berm to separate the north pond area from the south pond area is proposed with the crest set at elevation 7,230 feet. The berm will have a synthetic liner placed to create a hydrologic barrier between the north and south ponds.

The north pond will have an outlet designed for a maximum flow rate of 1,200 gallons per minute (gpm). A 12-inch high density polyethylene (HDPE) pipe with a longitudinal slope of approximately 0.3% will carry the peak flow rate to the WTP or a NPDES Outfall.

The outlet structure will consist of a 36-inch diameter HDPE slotted riser structure with an 18-inch diameter uncontrolled opening at elevation 7,208 feet. Flows within the pipe will be con-trolled by a 12-inch stainless steel valve. A trash rack will be installed on the HDPE riser at approximately elevation 7,213 feet to serve as an emergency overflow structure.

The outlet pipe will be a gravity flow system and located within the existing tailings facility as much as possible in order to reduce the required excavation. In order to provide gravity drainage, a portion of the pipeline will either need to be excavated, or a casing bored through the South Embankment abutment. The final elevation of the South Embankment will be at approximately elevation 7,230 feet and the invert of the pipeline in this location will be approximately elevation 7,200 feet (a cut of approximately 30 feet). A porous rock-fill berm will support the pipe as it crosses the approach to the Pinyon Hill porous rock berm spillway and serve as a sediment trap prior to water passing through the Pinyon Hill spillway.

3.6 Estimated quantities

The estimated quantities for the reclamation design are listed in Table 3. These approximate quan-tities demonstrate the order-of-magnitude of earthworks required during construction. However, actual quantities placed may vary from these estimates.

4 SUMMARY OF WORK COMPLETED

The Work Plan (Hecla and Water Management Consultants, 2007 a,b,c) was approved by the IDWR Dam Safety Division, the Environmental Protection Agency (EPA), and USDA Forest Service in the Spring 2008. To date, approximately 500,000 yds3 of base fill have been placed within the impoundment. Final reclamation is expected to be completed in 2010.

Table 3. Cover design quantities.

Material Estimated Quantity Units

Growth Media Topsoil (6 in) 45,000 yds3

Inert Cover 312,000 yds3

60-mil LLDPE Liner 2,400,000 ft2

Liner Subbase (12 in) 89,000 yds3

Tailings Cut 33,000 yds3

Base Fill* 1,278,000 yds3

South Embankment—24 in clay 55,000 yds3

South Embankment—12 in gravel 27,500 yds3

*Net volume—tailings cut placed as base fill

cy—cubic yards ft2—square feet.


435

REFERENCES

Golder Associates Inc. 1992a. Cyanidation and Dam Safety Permit Application Updates, Grouse Creek Project.

Golder Associates, Inc. 1992b. Plan of Operations, Hecla Mining Company Grouse Creek.Golder Associates Inc, December 1994. Certification Report, Quality Assurance/Quality Control Services,

Grouse Creek Tailings Impoundment, Stage 1, Stanley IdahoGolder Associates Inc, February 1997. Certification Report Grouse Creek Project Construction Quality Assur-

ance Monitoring and Test Results 1995 and 1996 Phase II Tailings Impoundment Raise, Stanley Idaho.Hecla Limited and Water Management Consultants, Inc. September, 2007a. Hecla Grouse Creek Unit Tailings

Impoundment Work Plan, Volume 1-Site background and Historical Information. Hecla Limited Grouse Creek Unit for USDA/EPA Administrative Order in Consent Scope of Work and Idaho Department of Water Resources.

Hecla Limited and Water Management Consultants, Inc. September, 2007b. Hecla Grouse Creek Unit Tailings Impoundment Work Plan, Volume 2-Design Report. Hecla Limited Grouse Creek Unit for USDA/EPA Administrative Order in Consent Scope of Work and Idaho Department of Water Resources.

Hecla Limited and Water Management Consultants, Inc. 2007c. Hecla Grouse Creek Unit Tailings Impound-ment Work Plan, Volume 3-Technical Specifications. Hecla Limited Grouse Creek Unit for USDA/EPA Administrative Order in Consent Scope of Work and Idaho Department of Water Resources.



437


Stabilized upstream tailings dam and converted into a filtered tailings facility

M.F. Veillette & T.E. MartinAMEC Earth & Environmental, Burnaby, BC, Canada

S. Alvarado LarretaGoldcorp, Durango, Mexico

ABSTRACT: The Cupias Tailings dam is located near the community of Tayoltita at the western extent of the Mexican state of Durango within the Sierra Madre Occidental mountain range. The Cupias tailings dam was initially constructed with a compacted starter dam made of locally bor-rowed material. The dam was raised using spigotted tailings via the upstream method and has been in operation for more than 30 years with a current height of about 70 m. Supernatant water was reclaimed via a series of vertical towers that are connected by a concrete decant drainage tunnel beneath the impoundment. It was determined that existing static stability did not meet an accept-able factor of safety. As a result, phased stabilization berms were designed to mitigate the dam stability issue and tailings management technology was changed from the existing slurry disposal method to a tailings filtration plant, thus converting to a filtered tailings facility. The stabilization berms were to be constructed from filtered tailings, however due to timing issues with the filtra-tion plant, the stabilization berms were mainly constructed out of locally borrowed material. This paper describes the project setting and requirements, construction, performance of the dam and the expansion of the filtered tailings facility design and construction components. The filtered tailings start-up facilities were constructed in 2008.

1 BACKGROUND

The San Dimas mine is located near the village of Tayoltita at the western extent of the Mexican state of Durango within the Sierra Madre Occidental mountain range. The project location is shown in Figure 1. San Dimas is a gold/silver producer, with a current daily ore throughput of approximately 2100 tonnes per day (tpd). Prior to 2002, Minas Luismin S.A. de C.V. (Luismin) owned and operated the San Dimas Mine and surrounding facilities including the Cupias tailings facility. In 2002, Wheaton Minerals Ltd. acquired Luismin and in 2005 Wheaton Minerals Ltd. merged with Goldcorp Inc. Goldcorp is the current owner of the San Dimas Mine.

The mine has been in operation for more than 30 years and tailings were previously stored in a number of small upstream constructed impoundments located in the vicinity of the mill. Since 1985 tailings have been deposited in the Cupias tailings impoundment, which is approximately 1.4 km north-east and approximately 65 m higher than the mill. The Cupias tailings dam was constructed initially with a compacted starter dam made of locally compacted borrowed material and had been subsequently raised with tailings by the upstream construction method until the end of 2007. The current crest elevation of the tailings dam is approximately at 715 m, with a maxi-mum height of 70 m. Water was reclaimed via a series of vertical towers that are connected by a concrete and steel drainage tunnel beneath the impoundment. The concrete (box culvert) and steel pipe drainage tunnel was used to discharge reclaim water into a concrete basin located at the toe of the dam from where the water was recycled to the mill.

A site investigation program and a stability assessment of the tailings dam were undertaken in 2003. One of the key conclusions of the study was that embankment stability considering potential


438

mobilization of undrained shear strengths within the tailings comprising the dam did not meet acceptable factors of safety. Stabilization berms were designed to mitigate the dam stability issue and recommendations were made to replace the existing slurry disposal method with a tailings filtration plant, to progressively convert the facility from a conventional upstream-constructed impoundment to a filtered tailings stack.

Design of the filter plant commenced in early 2004. Prior to the completion of the filtration plant, the first portion of the stabilization berm (Stage I) was constructed with locally compacted borrowed materials up to an elevation of 668 m. Stage I of the stabilization berm was completed in May 2005. A delay in the construction of the filter plant meant that filtered tailings would not be available as originally scheduled for use in stabilization berm construction. As such, Stages II and III of the stabilization berms (∼400,000 m3) were constructed from locally compacted borrowed material mixed with filtered tailings (approximately 25% by volume) in 2007. San Dimas began construction of the filtered tailings start-up facilities in January 2008 and most structures have been completed as of June 2008.

2 SITE INVESTIGATION AND RESULTS

2.1 General

A site investigation program comprising of test pits, drill holes and electric cone penetration testing (CPT) was carried out in 2003. The locations of the various soundings are shown on Figure 2. Test pits were excavated to characterize the downstream foundation area of the tailings impoundment and to identify potential borrow areas. Two (2) drill holes were executed downstream of the starter dam to characterize the bedrock foundation. The colluvium foundation material downstream of

Figure 1. Site location.


439

the tailings dam consists of approximately 2 to 3 m of well graded sand (SM), with gravel, silt and cobbles, trace boulders and clay. The colluvium is underlain by highly to completely weathered granodiorite bedrock.

2.2 Piezocone investigation

A total of nine piezocone (CPT) soundings were executed along two sections of the existing impoundment. CPT-2, 3, 4, 5 and 7 were performed along the maximum section of the dam as shown on Section A-A of Figure 3. CPT-1 and 9 were located at an offset north of the estimated

Figure 2. Site plan conditions in 2003 with site investigation layout.

Figure 3. Tailings dam cross section A-A for 2003 conditions with critical failure surfaces.


440

starter dam location as shown on Section B-B of Figure 4. CPT-6 was performed offset south of the main section in the tailings beach and CPT-8 was performed through the dam shell between CPT—7 and CPT—9 as shown on Figure 4. CPT–2 and CPT–8 encountered obstructions at shal-low depth and did not provide useable data and no dissipation data for CPT-1 was provided so peak and residual strengths for that sounding had to be inferred based on pore pressure conditions indicated by other CPT soundings.

2.3 Piezocone interpretation chart

In recent years, as first proposed by (Plewes et al. 1992), and expanded by (Davies, 1999), critical state theory has been incorporated into interpretation of appropriately stress-normalized piezo-cone data to allow, at a preliminary, screening level, and semi-empirical estimation of the state parameter (ψ) for the more silty materials typical of tailings. Rather than a rigorous definition of the critical state line via undisturbed sampling (always a problematic proposition in tailings) and laboratory testing, this method uses the state parameter concept to compare piezocone data for a given site against liquefaction (and non-liquefaction) case records, in much the same way as the Standard Penetration Test (SPT) has for many years been used to evaluate the liquefaction suscep-tibility of sandy soils against known case records of seismic liquefaction.

In this work, ψ is estimated from piezocone data as follows:

ψσ

=

−+

⎡

⎣⎢

⎤

⎦⎥

−=

−=

�nQ B

F

FF

f

qB

uq

s

t voq

( )

( . . / )

( . . ) ( )

(

1

3 6 10 2

1 33 11 9

−−−

= −−

u

qQ

q

uo

t vo

t vo

vo o

)

( )

( )

( )σσ

σ

where:q

t = cone tip resistance u

0 = equilibrium (static) porewater pressure

σvo

= total vertical stress Bq = stress normalized cone pore pressure parameter

σvo

’ = effective vertical stress Q = stress normalized tip resistanceu = dynamic porewater pressure f

s = sleeve friction

F = stress normalized sleeve friction.

Figure 4. Tailings dam cross section B-B for 2003 conditions with critical failure surfaces.


441

The above relationships were used to estimate ψ profiles for representative piezocone soundings from at the two sections of the tailings dam. The method has been used on more than 50 tailings facilities and calibrated with several static and earthquake induced liquefaction events. Cali-bration of the piezocone state parameter with frozen-core samples indicates an approximate degree of accuracy in Ψ of about +/− 0.02 to 0.03 (accuracy in void ratio). For additional per-spective, the piezocone data were also used to derive equivalent (N

1)

60 values (uncorrected for

fines content, and corrected to equivalent clean sand values) based on the methods proposed by (Davies, 1999).

Figure 5 is a plot that shows the inferred state parameter lines for ψ = 0, ψ = −0.1, ψ = −0.2, ψ = −0.3, and the (N

1)

60-ECS (ECS = equivalent clean sand) lines for (N

1)

60-ECS = 10, (N

1)

60-ECS =

20, (N1)

60-ECS = 30. To use the chart, for each piezocone data point, the Q(1-B

q) and F values are

plotted. If a majority of the points fall to the positive side of the ψ = 0 line, it indicates that the tailings are in an initially loose state and are therefore susceptible to contractant behaviour under shear strain, which can lead to liquefaction. This chart also shows material index zones based on piezocone-interpretation of the material types penetrated. These zones are as follows:

Zone 2—Organic soils/peatZone 3—Clays—clay to silty clayZone 4—Silt mixtures—clayey silt to silty clayZone 5—Sand mixtures—silty sand to sandy siltZone 6—Clean sand to silty sand.

Figure 5. State parameter (ψ ) profile and liquefaction screening charts at section A-A.

State Parameter (Ψ)

0

2

4

6

8

10

12

14

16

18

20

22

24

26

28

30

32

-0.2 -0.1 0.0 0.1 0.2

De

pth

(m

)

Liquefaction Susceptibility

ψ>0 High (static & dynamic)-0.1<ψ<0 Low (static)-0.1<ψ<0Intermediate (dynamic)ψ<-0.1Low

Liquefaction Screening Chart for Section A-A'

1

10

100

1000

F (%)

Q(1

-B))

Liquefaction Susceptibility

ψ>0 High - static & dynamic-0.1<ψ<0 - Low (static)-0.1<ψ<0 Intermediate (dynamic) ψ<-0.1 Negligable

ψ = -0.1

(N1)60ECS = 30

(N1)60ECS = 20

ψ = -0.3

ψ = -0.2

ψ = 0

(N1)60ECS = 10

Zone 6

Zone 7

Zone 2

Zone 5

Zone 4

Zone 3

Piezocone datanear two tailingsdam liquefactionslumps

t

CPT 3 CPT 5CPT 4 CPT 7

CPT 3 CPT 5CPT 4 CPT 7


442

In Figure 5, the boxed areas represent regions for two separate static liquefaction events, one at the Suncor oil sands tailings facility and the other at the Sullivan Mine tailings facility (Davies, 1999). Data from piezocone soundings taken from adjacent sections were evaluated as described in the previous sections and points were plotted as described above. As shown, for each of the cases, the majority of points fall to the positive side of the ψ = 0 line and indicates that the poten-tial for liquefaction is predicted (Davies, 1999).

The piezocone soundings taken at the Cupias tailings impoundment in 2003 have been incorpo-rated into the screening chart to assess the liquefaction potential of the tailings.

2.4 Static pore pressure conditions

Fundamental to the interpretation of piezocone data, and indeed to the safety of upstream tail-ings dams, is the estimation of the equilibrium (static) porewater pressure. This is best done by pausing the piezocone penetration and conducting a series of pore pressure dissipation tests at various depth intervals within a single sounding. In upstream dams, pore pressure conditions that are not hydrostatic usually represent the rule rather than the exception (Martin, 1999). Given the stress-normalization inherent in the estimation of the state parameter from piezocone data, failure to correctly characterize the static porewater pressure will lead to erroneous results. The dissipation results from all of the CPT soundings in 2003 showed a marked downward gradient in the tailings, with pore pressures well below hydrostatic, a favorable condition in terms of the stability of the dam. The strong downward gradient is most likely facilitated by the relatively per-vious foundation soils underlying the tailings deposit.

Fourteen (14) vibrating wire piezometers (VWP) were installed along Sections A-A and B-B in May 2006 to provide continuous pore pressure monitoring of the tailings dam in order to allow ongoing monitoring of pore pressure conditions and hydraulic gradients (See Figure 6).

3 STABILITY ANALYSIS

3.1 General

Limit equilibrium analyses were carried out to evaluate the stability of the tailings dam under static loading conditions. Pore pressure conditions, determined from piezocone soundings as described in the previous section, were input as a grid of pressure heads into SLOPE/W. Two sections were analysed for stability, Section A-A and Section B-B as shown on Figure 2. Both sections are considered critical in terms of the extent and strength of the tailings zone. The shell tailings and tailings zones for these sections were approximated using the results from piezocone soundings as described in the previous sections.

The state parameter was used to delineate dilatant and contractant zones within the tailings, and shear strength parameters, drained and undrained, were assigned appropriately to these zones. Two distinct zones were identified along the CPT profiles: a shell tailings zone and a tailings zone. Within the tailings through section A-A and B-B, three different strength profiles were evident. The shear strength parameters for these zones were assumed as follows:

• Shell tailings—assumed to be unsaturated and/or dilatant under shear, so use of drained shear strength parameters is appropriate. The drained shear strength of the tailings was characterized using an effective friction angle (φ’) of 30°. This is a typical φ’ value for tailings supported by much case history experience.

• Tailings—these zones were assumed to be contractant under shear for both static and dynamic loading conditions, so use of undrained shear strength parameters is appropriate. Through Sec-tion A–A undrained shear strength ratios (Su

/p’) of 0.2, 0.15 and 0.1 were selected based on the piezocone data, to characterize the peak undrained shear strength of this zone. Through Section B–B undrained shear strength ratios of 0.3, 0.2 and 0.1 were selected. For both sections the zones were assigned an effective friction angle (φ’) of 30° to characterize their drained shear strength.


443

Figures 3 and 4 give the major sections used for stability analysis and show the shell tailings and tailings zones as inferred from the CPT profiles. Table 1 summarizes the shear strength param-eters for slope stability used for each of the materials in the stability analysis.

The stability of the two dam sections was estimated for three conditions:

1. Drained, static loading conditions or effective stress analysis (ESA), wherein there are no shear-induced pore pressures, or any shear-induced pore pressures are dissipated as quickly as they are generated;

2. Peak undrained shear strength conditions (P-USA), wherein there are positive shear-induced pore pressures, but the level of shear strains is insufficient to cause a substantial reduction in undrained shear strength (i.e. non-brittle behaviour); and

3. Residual undrained shear strength conditions (R-USA), wherein the tailings liquefy and assume a post-liquefaction residual undrained shear strength.

The drained, static analyses modeled the stability of the dams under steady state conditions representing the long-term stability of the dams in the absence of triggering of undrained condi-tions. A minimum factor of safety of 1.5 was used for these analyses. For the analyses incor-porating peak undrained shear strength (P-USA) values in the tailings, a minimum factor of safety of 1.3 was used. Were residual undrained shear strengths (R-USA) to be triggered, this would be a short-term condition, representing the worst (i.e. minimum) possible scenario in terms of stability. As such, a minimum factor of safety of 1.1 was adopted for this shear strength condition.

3.2 Stability condition in 2003

Stability analyses were carried out on the 2003 geometry for Sections A-A and B-B to determine the factors of safety for drained conditions in the tailings as well as for peak and undrained shear strength conditions. Once these factors of safety were determined, an analysis was performed to determine the size of berms required to bring the factors of safety up to the target levels. Figures 3 and 4 show the factors of safety for the geometry of the tailings impoundment as it existed in 2003 for each of the strength conditions for the tailings.

A factor of safety of less than one, as implied for the USA stability analyses for both dam sections, implies failure, yet the dam has not failed. This does not mean that the undrained shear strength analyses are incorrect. Rather, it means that the dam had not been subjected to a triggering

Table 1. Cupias impoundment material properties.

Bulk unit

EffectiveMaterial weight,

Undrained shear strength Friction angle, Pore pressure

description γ (kN/m3) Peak (Su/σ

v’) Residual (S

r/σ

v’) φ' (degrees) conditions

Starter Dam 18 33 Zero (drained)Berm 18 33 Zero (drained)Shell tailings 18 30 approximatedTailings zone A 18 0.2 00.1 from CPTTailings zone B 18 0.15 00.15 dissipationTailings zone C 18 0.1 00.05 testsTailings zone D 18 0.15 00.1Tailings zone E 18 0.3 00.3Tailings zone F 18 0.2 00.1Tailings zone G 18 0.1 00.05


444

mechanism sufficient to induce undrained loading within the dam and hence mobilize the lower undrained shear strengths. Therefore, rather than expressing the safety of the dam in terms of factors of safety derived using undrained shear strengths, the safety of the dam might be better described in terms of the probability of undrained loading, with subsequent mobilization of peak or residual undrained shear strengths, being triggered within the tailings. Unfortunately, quantifi-cation of the annual probability of such triggering mechanisms is very difficult if not impossible. A generic summary of potential triggering mechanisms that could lead to static liquefaction is outlined in (Martin and McRoberts, 1999).

Most potential triggering mechanisms can be avoided through operational controls and moni-toring efforts on the part of the mine operator. However, some mechanisms such as strong earth-quakes are obviously beyond the operators control and pose an ongoing risk.

3.3 Stability berms

As shown in Figures 3 and 4, the existing factors of safety (FoS) under undrained loading condi-tions for the embankment geometry as it existed in 2003 were below acceptable levels, thus sta-bilization berms were designed. The construction of the stability berm to achieve the target FoS consists was carried out in three stages. Results of the Stage III stability analyses are summarized in Table 2.

Since the stabilization berm is sized to meet the appropriate FoS for peak undrained condi-tions (FoS > 1.3), the FoS for the static drained conditions is well in excess of the minimum target of 1.5. Although the FoS for residual undrained FoS is lower than 1.1 for Section B-B, it was deemed acceptable given that for other two-dimensional sections the FoS of > 1.1 is achieved.

The required staged stabilization berms configuration for Section A-A is shown in Figure 6. Filter and drainage zones were incorporated within and below the stabilization buttress to reduce the potential of porewater pressure build up.

Figure 6. Stabilization berm for section A-A for peak undrained shear strength conditions.

Table 2. Factors of safety for existing geometry with Stage III stabilizing berm.

FoS

Strength of tailings Section A-A Section B-B

Peak drained (ESA) 2.8 2.2Peak undrained (P-USA) 1.3 1.3Residual undrained (R-USA) 1.1 1.0

* Criteria is FoS ≥ 1.3 for peak undrained conditions.


445

4 FILTERED TAILINGS FACILITY EXPANSION

An options study was undertaken in 2003 to determine which alternative would be best suited for future tailings deposition in the Cupias valley. Six (6) different options were analyzed; raising the existing tailings dam (upstream raise method) with stability berms utilizing different alignments, constructing a new tailings dam downstream (centerline raise method) utilizing different align-ments or filtered tailings. Filtered (dewatered) tailings was chosen to be the most appropriate on the basis of economics and environmental issues.

Following completion of the stabilization berm, construction began on the filtered tailings start-up structures. Start-up structures for the filtered tailings facility involved constructing a toe berm, an under-drain, a diversion ditch and sedimentation berms prior to placement of any fil-tered tailings downstream of the stabilization berm. Ultimately, the filtered tailings facility will be expanded to fill the entire valley with filtered tailings. The ultimate filtered tailings stack has been conceptually estimated to store approximately 10 million m3 giving the facility a projected life of approximately 15 years at the current production rate of 2100 tpd. Figure 8 shows the ultimate configuration of the filtered tailings facility expansion.

4.1 Filtered tailings start-up components

4.1.1 Rock under-drain and inverted filterA rock flow-through under-drain was designed and constructed to convey seepage water from the impoundment’s decant system, and to provide under-drainage below the filtered tailings. rock under-drain was overlain by a filter sequence to prevent tailings ingress into the under-drain. Figures 7 and 8 illustrate the rock under-drain.

The upstream end of the decant line was plugged and an inverted filter downstream of the exist-ing decant outlet was constructed. This was done because the decant conduit could not be effectively filled. With the inverted filter weighed down by the filtered tailings, and the water pond eliminated from the tailings impoundment, there are redundant defenses against any potential conduit rupture and water loss that could transport tailings through the conduit.

4.1.2 Toe bermA starter rock fill toe berm was constructed at the toe of the ultimate filtered tailings facility, immediately upstream of the western sedimentation berm. As the compacted filtered tailings are raised above the toe berm, durable, non-acid generating waste rock or borrowed rock material will be placed over the tailings slope, founded initially on the toe berm crest. Figures 7 and 8 show the location of the toe berm.

4.1.3 Sedimentation bermsThe primary function of the western sedimentation berm just downstream of the toe berm is to retain sediments that are transported downstream from the filtered tailings facility as a result of erosion likely to occur during significant rainfall events. A pumping system to pump impounded water from the sedimentation pond over the berm for discharge will be installed. A spillway will be constructed along the left abutment of the berm to prevent overtopping of the berm during storm inflow events that yield inflows beyond its storage capacity. The sedimentation pond will be cleaned out during the dry season on a yearly basis.

The eastern sedimentation berm (at the upstream end of the tailings impoundment) was con-structed to provide a means of collecting runoff and contact water prior to release. The proposed pond expansion was developed to hold contact water that could potentially contain elevated con-centrations of suspended solids or other contaminants that may result after peak rainfall events. During normal operations, contact water from the eastern sedimentation pond is recycled back to the mill. However during intense storm events, excess water from the eastern sedimentation pond will be pumped into the eastern treatment pond for treatment prior to being discharged. Figures 7 and 8 shows the locations of the western and eastern sedimentation berms.


446

4.1.4 Diversion ditch and damA diversion dam was designed and constructed to route the bulk of the runoff from the south catchment area into the diversion ditch as shown in Figure 8. The diversion dam and ditches were sized to route a 24-hour duration, 200-year rainfall event corresponding to 193 mm of rainfall. The typical diversion ditch cross section has a trapezoid shape with a 2.5 m wide base, 1H:1V side slopes and is a minimum depth of 2 m constructed into bedrock. The tailings facility has negligible catchment areas towards the east and north and no diversions ditches were deemed necessary in these areas.

Figure 7. Filtered tailings expansion facility with key start-up structures and slope stability.

Figure 8. Ultimate filtered tailings facility expansion layout.


447

4.1.5 Tailings facility spillwayThe spillway will be located at the eastern ridge of the tailings impoundment as shown in Figure 8. The spillway has been sized to route the 24-hour duration Probable Maximum Precipitation (PMP) of 554 mm occurring across the entire 42 ha catchment. The diversion dam and ditches are assumed, for spillway design purposes, to have failed during the PMF event. The spillway cross section has a 4 m wide base, 2H:1V side slopes, a depth of 2 m, and will be lined with geotextile and 450 mm minus rip rap since the spillway is founded on native overburden.

4.2 Tailings facility expansion design

The performance of the filter plant is such that the water content of the filtered tailings is several percent above optimum (as defined by the standard Proctor compaction test). Due to the poten-tially saturated nature of the tailings once placed, a compacted shell was deemed to be required to provide a high shear strength zone, non-susceptible to liquefaction, and provide stability and containment of the non-compacted tailings which would be placed in a generally loose state. The compacted shell is to be constructed of unsaturated filtered tailings with moisture content ranging +/− 2% of the optimum moisture content with an average benched downstream slope of 2.5H:1V. The compacted shell will be constructed to a crest elevation of 725 m. This will yield a maximum vertical height of 110 m for the final embankment slope. The interface between the compacted tailings and the non-compacted, potentially saturated tailings will comprise a granular filter/drain 3 m wide (2 m perpendicular to the slope). The intent of the filter material is to facilitate drainage for the outer shell, via interception of any water expelled from the high water content tailings due to consolidation as the tailings stack is progressively raised.

Given the uncertainty regarding the degree of saturation and shear strength of the uncom-pacted filtered tailings, a conservative analytical design approach was taken for definition of the geometry of the compacted shell section. This involved the very conservative assumptions that: a) the uncompacted tailings could be saturated and could liquefy; and b) the post-liquefied shear strength of the uncompacted tailings could be represented by an undrained shear strength ratio (S

u/σ

v’) of 0.1. The shell configuration illustrated in section on Figure 7 was recommended until

more information can be gathered with respect to the behaviour of the uncompacted tailings over the long term. For stability analysis purposes, hydrostatic pore pressures were assumed for the uncompacted tailings. To facilitate an understanding of actual conditions within the uncompacted filtered tailings, piezometers will be installed. As more information is made available, the slope may be revisited and further optimization of the impoundment can be made.

The filtered tailings are typically left to dry for at least 3 to 4 days prior to being compacted with a dozer and a smooth drum roller for the downstream shell area. The compacted shell will be raised a minimum height of 5 m during the dry seasons. To facilitate the construction of the compacted shell prior to the placement of the saturated tailings, small upstream berms will be constructed in vertical stages of 5 m. The upstream berms will be constructed of similar material as that of the compacted shell. Tailings deposition will be upstream of the compacted shell area during the rainy season when filtered tailings cannot be suitably dried for trafficking and compac-tion. This is illustrated in Figure 7.

The downstream face of the compacted shell will be covered by a minimum 0.5 m thick layer of erosion protection material consisting of 300 mm minus graded rock or non mineralized waste rock. Additional 600 mm minus rip rap will be placed along the south abutment contact to prevent erosion should flows be larger than a 200 year, 24 hour storm event.

5 CONCLUSIONS

The Cupias tailings impoundment was developed as a conventional upstream-constructed tailings impoundment. Piezocone investigations demonstrated the tailings to be in a very loose and poten-tially liquefiable state. The tailings dam thus was susceptible to flowslide failure should undrained


448

shear strengths be mobilized via an appropriate triggering mechanism, a strong earthquake being but one. With this condition identified, stabilization of the dam to achieve adequate factors of safety under both drained and undrained loading became imperative.

At the same time as stabilizing the upstream dam, the San Dimas mine also required expanded tailings storage capacity.

The most cost-effective solution to meeting both objectives was to convert the upstream con-structed, slurried tailings impoundment to a filtered tailings stack. This was and continues to be achieved in a staged manner. The San Dimas filter plant results in tailings that are significantly wet of optimum for compaction, whereas typically filter plants are intended to achieve a moisture content near optimum in order to facilitate both trafficability of the tailings, and compaction. This unique aspect at San Dimas resulted in the need for a defined, outer compacted shell. Ongo-ing monitoring (including piezometers) and operating experience as the facility develops will be employed to optimize its design and construction.

ACKNOWLEDGEMENTS

The authors wish to thank the San Dimas mine and Goldcorp Ltd. for permission to publish this unique case record illustrating the conversion of a conventional, upstream-constructed tailings impoundment to a filtered tailings facility.

REFERENCES

Been, K. & Jefferies, M.G. 1985. A State Parameter for Sand. Géotechnique, Volume 35, Number 2, pp. 99–112.

Canadian Dam Association, 1999. CDA Dam Safety Guidelines.Davies, M.P. 1999. Piezocone Technology for the GeoEnvironmental Characterization of Mine Tailings. Ph.D.

Thesis, University of British Columbia.GEO-SLOPE International, Ltd. GeoStudio 2004 (Version 6.22). Calgary, Alberta.Martin, T.E. 1999. Characterization of pore pressure conditions in upstream tailings dams. Proceedings, Tail-

ings & Mine Waste ’99, Fort Collins, Colorado, pp. 303–313.Martin, T.E. & McRoberts, E.C. 1999. Some considerations in the stability analysis of upstream tailings

dams. Proceedings, Tailings & Mine Waste ’99, Fort Collins, Colorado, pp. 287–302.Plewes, H.D., Davies, M.P. & Jefferies, M.G. 1992. CPT based screening procedure for evaluating liquefac-

tion susceptibility. Proceedings 45th Canadian Geotechnical Conference, pp. 41–49.


449


Addo, B.A. 315Altobello, J.A. 125Amorim, N.R. 199Athanassopoulos, C. 189Azevedo, I.D. 199Azevedo, R.F. 199

Banta, L.E. 125Bates, E.E. 409Bedell, P.M. 17Belem, T. 139Benzaazoua, M. 139Boxill, L.E. 291Boye, A. 25Breckenridge, L. 243Bryan, R. 269Bussière, B. 139

Chahbandour, J. 360Chapel, T.A. 401Cherry, J.C. 35, 49Childers, H.M. 325Conroy, K. 281Cooke, R. 95Cullen, V. 165

Doerffer, M. 101Donohue, S.V. 35, 49

Eary, L.E. 221Elliott, J. 305Eshleman, J. 221Eykholt, G.R. 35, 49

Ferreira, O.R. 199Fisseha, B. 269Fourie, A.B. 3Fredland, J.W. 345Fundingsland, S. 253

Author index

Gardiner, E. 337George, L. 369Gladwin, D. 337Gusek, J. 281

Hazen, G. 253Heinrich, R. 101Henderson, M. 189Henriquez, J. 269Hight, P. 253Hudson, A. 243Hudson, A.L. 73

Isham, J.C. 207

Jakubowski, R. 221Johnson, J. 305Jolley, R. 401

Kaul, J. 189Ketellapper, V. 253Kohlman, A. 189Kunkel, J.R. 63

Larreta, S.A. 437Li, A. 83Ludwick, W. 369Lupo, J.F. 153, 177

Manchester, J.B. 35, 49Martin, T.E. 291, 437McKenna, G. 165Merkel, B. 233Meyer, T. 73Monok, B. 305Morrison, K. 305Motzer, W.E. 393Moye, F. 423Myles, P. 325

Naamoun, T. 233Nelson, M. 253Nyame, F. 315

Ortman, D. 361

Palombo, S. 113Poos, S. 243Poulter, D. 423

Quaranta, J.D. 125, 325

Raabe, K.L. 381Ribeiro, A.G.C. 199Rogers, S. 423Rutkowski, T. 281

Sheets, R.J. 409Simms, P. 269Strachan, C.L. 381

Thompson, D. 243Thorpe, M.B. 315Trautwein, A. 265Tritthart, B. 423

van Zyl, D. 355Varela, J. 113Veillette, M.F. 437

Watson, A. 221Will, R.K. 393Willis, K.S. 17Wislesky, I. 83Woodward, C. 401

Yilmaz, E. 139Young, J. 25

Date post:	15-Jan-2016
Category:	Documents
Upload:	indra-setiawan-sunarja
View:	177 times
Download:	13 times

Handbook Tailings and Mine Waste 2008

Documents