Quantifying Submarine Groundwater Discharge and Nutrient ...Anthony J. Smith, Jeffrey V. Turner,...

C S I R O L A N D a nd WAT E R

Quantifying Submarine Groundwater Discharge

and Nutrient Discharge into Cockburn Sound,

Western Australia

Anthony J. Smith, Jeffrey V. Turner, David E. Herne and Wayne P. Hick

CSIRO Land and Water, Perth

Technical Report 1/03, June 2003

A Technical Report to Coast and Clean Seas Project WA9911:

Quantifying Submarine Groundwater Discharge and Demonstrating

Innovative Clean-Up to Protect Cockburn Sound from Nutrient Discharge

Quantifying Submarine Groundwater Discharge and Nutrient Discharge into Cockburn Sound, Western Australia

A Technical Report to Coast and Clean Seas Project WA9911: Quantifying Submarine Groundwater Discharge and Demonstrating Innovative Clean-Up to Protect Cockburn Sound from Nutrient Discharge Anthony J. Smith, Jeffrey V. Turner, David E. Herne and Wayne P. Hick CSIRO Land and Water, Wembley, Western Australia 6913 Technical Report No. 01/03 ISSN 1446-6163 January 2003 ISBN No. 0 643 061134 CGS Report No. 104

© CSIRO Australia 2003

To the extent permitted by law, all rights are reserved and no part of this publication (including

photographs, diagrams, figures and maps) covered by copyright may be reproduced or copied in any

form or by any means except with the written permission of CSIRO Land and Water.

Important Disclaimer:

CSIRO Land and Water advises that the information contained in this publication comprises general

statements based on scientific research. The reader is advised and needs to be aware that such

information may be incomplete or unable to be used in any specific situation. No reliance or actions

must therefore be made on that information without seeking prior expert professional, scientific and

technical advice. To the extent permitted by law, CSIRO Land and Water (including its employees

and consultants) excludes all liability to any person for any consequences, including but not limited to

all losses, damages, costs, expenses and any other compensation, arising directly or indirectly from

using this publication (in part or in whole) and any information or material contained in it.

Technical Report - January 2003

Quantifying Submarine Groundwater Discharge i and Nutrient Discharge into Cockburn Sound

TABLE OF CONTENTS

Table of Contents...........................................................................................................i

List of Tables ...............................................................................................................iv

List of Figures ..............................................................................................................vi

Executive Summary ......................................................................................................xi

Acknowledgements.....................................................................................................xx

1 Introduction ............................................................................................................1

1.1 Background.....................................................................................................1

1.2 This Study.......................................................................................................2

1.3 Report Format.................................................................................................2

2 Hydrogeology.........................................................................................................5

2.1 Regional Setting .............................................................................................5

2.2 Subcrop Formations......................................................................................10

2.2.1 Kardinya Shale Member........................................................................10

2.2.2 Pinjar and Waneroo Members...............................................................10

2.2.3 Rockingham Sand .................................................................................10

2.3 Superficial Formations .................................................................................14

2.3.1 Tamala Limestone.................................................................................14

2.3.2 Cooloongup Sand..................................................................................14

2.3.3 Becher Sand ..........................................................................................15

2.3.4 Safety Bay Sand ....................................................................................15

2.4 Groundwater Occurrence and Flow..............................................................15

2.4.1 Artesian Aquifer System .......................................................................15

2.4.2 Rockingham Aquifer.............................................................................16

2.4.3 Superficial Aquifer................................................................................16

2.5 Groundwater Bores.......................................................................................19

2.6 Groundwater Levels......................................................................................26

2.6.1 Regional Groundwater Table ................................................................26

2.6.2 Influence of Sea Level Variation...........................................................26

2.6.3 Influence of High-permeability Limestone ...........................................27

2.7 Seawater Intrusion ........................................................................................31

3 Field Investigations ..............................................................................................37

3.1 Northern Harbour..........................................................................................37

3.1.1 Hydrogeology........................................................................................39

3.1.2 Groundwater Levels ..............................................................................40

3.1.3 EC Surveys of Submarine Porewater ....................................................40

CSIRO Land and Water

ii Coast and Clean Seas Project No. WA9911

3.1.4 Groundwater Chemistry and Isotopic Composition..............................43

3.1.5 Porewater Chemistry.............................................................................43

3.1.6 Benthic Flux Measurements .................................................................50

3.1.7 Continuous Radon Monitoring in Northern Harbour............................57

3.1.8 Seismic Survey......................................................................................57

3.1.9 Groundwater Temperature and EC Profiling ........................................62

3.2 Challenger Beach..........................................................................................67

3.2.1 Hydrogeology........................................................................................67

3.2.2 EC Surveys of Submarine Porewater....................................................69


3.3 BP Refinery ..................................................................................................73

3.3.1 Hydrogeology........................................................................................73

3.3.2 EC Surveys Submarine Porewater ........................................................76


3.3.4 Benthic Flux Measurements .................................................................78

3.4 Cockburn Sound Submarine Porewater Survey ...........................................81

3.4.1 Ionic Chemistry.....................................................................................81

3.4.2 Calculations Of Groundwater Nutrient Concentrations........................81

3.4.3 Synthesis Of Results .............................................................................82

4 Quantifying SGD into Cockburn Sound ..............................................................95

4.1 Conceptual Models of SGD .........................................................................95

4.2 Previous Estimates of SGD ........................................................................100

4.2.1 Flownet Calculations ..........................................................................101

4.2.2 CTD Profiling .....................................................................................102

4.3 Direct Measurement of SGD......................................................................102

4.4 Estimation of SGD by Inshore Water Balance ...........................................103

4.4.1 Groundwater Flow Model...................................................................103

4.4.2 Groundwater Recharge .......................................................................105

4.4.3 Calibration ..........................................................................................105

4.4.4 Results.................................................................................................106

4.4.5 Conclusions.........................................................................................108

4.5 Estimation of SGD Based On Saltwater-Freshwater Relations..................115

4.6 Summary.....................................................................................................117

5 Quantifying Nutrient Discharge into Cockburn Sound......................................119

5.1 Mechanisms for Nutrient Discharge...........................................................119

5.2 Nitrogen Concentration in Groundwater ....................................................120

5.3 Significance Of Measured Concentrations .................................................123


Quantifying Submarine Groundwater Discharge iii and Nutrient Discharge into Cockburn Sound

5.4 Previous Estimates Of Nutrient Discharge .................................................125

5.5 Estimation Of Nutrient Discharge Based On SGD Rates From The Inshore

Water Balance Model And Measured Nutrient Concentrations In Submarine

Porewater ....................................................................................................127

5.5.1 Method ................................................................................................127

5.5.2 Results.................................................................................................128

5.6 Estimation Of Nutrient Discharge Based On SGD Rates From The Inshore

Water Balance Model And Groundwater Nutrient Concentrations Measured

By Appleyard (1994) ..................................................................................135

5.6.1 Method ................................................................................................135

5.6.2 Results.................................................................................................136

5.7 Synthesis of Results....................................................................................141

5.7.1 Uncertainty..........................................................................................144

5.7.2 Composite Estimate ............................................................................146

References.................................................................................................................149

Appendix A: Aquifer Tidal Propagation at Northern Harbour.................................153

The Tidal Method...............................................................................................154

Fourier Analysis .................................................................................................156

Discrete Fast Fourier Transform (FFT)..............................................................157

Fremantle Tide Data...........................................................................................157

Fremantle Tide Data...........................................................................................158

Aquifer Tidal Propagation..................................................................................168

Appendix B: Benthic Flux Meters ............................................................................173

Lee-type..............................................................................................................174

Ultrasonic ...........................................................................................................174

Heat-pulse...........................................................................................................175

Appendix C: Models of Tide and Wave Affects on Benthic Flux............................179

Effect of Sea Level Dynamics ............................................................................180

Tide-Forced Model.............................................................................................181

Wave-Forced Model...........................................................................................182


iv Coast and Clean Seas Project No. WA9911

LIST OF TABLES

Table 1: Estimated hydraulic properties of the Superficial Formations inshore

from Cockburn Sound ..............................................................................18

Table 2: Relation between aquifer thickness and water table elevation at the

position of the saltwedge toe based on the Ghyben-Herzberg

approximation...........................................................................................32

Table 3: Electrical conductivity of groundwater samples from saltwater

monitoring bores.......................................................................................33

Table 4: Completion details of intercomparison monitoring bores ............................39

Table 5: Bulk-ground EC surveys by Steiglitz (2001b); measurement times and

ocean water levels.....................................................................................41

Table 6: Bulk-ground EC surveys by CLW; measurement locations and times ........42

Table 7: Groundwater chemistry and Isotopes ...........................................................44

Table 8: Northern Harbour porewater chemistry........................................................45

Table 9: Benthic flux measurements; SCOR SGD intercomparison..........................51

Table 10: Bulk-ground EC surveys; measurement locations and times .....................69

Table 11: Challenger Beach porewater chemistry......................................................71

Table 12: Bulk-ground EC surveys; measurement locations and times .....................76

Table 13: BP Refinery porewater chemistry...............................................................77

Table 14: Porewater chemistry ...................................................................................85

Table 15: Estimates of Submarine Groundwater Discharge into Cockburn

Sound and related sites ...........................................................................100

Table 16: Cockburn Groundwater Area model: conductance and external head

values for the general head coastal boundary cells .................................105

Table 17: Cockburn Groundwater Area model: calibrated hydraulic

conductivity values .................................................................................106

Table 18: Cockburn Groundwater Area model: model water balance for low-

recharge scenario ....................................................................................107

Table 19: Cockburn Groundwater Area model: model water balance for high-

recharge scenario ....................................................................................107

Table 20: Summary of nutrient concentrations in groundwater in Cockburn

Sound area and related sites....................................................................121

Table 21: Default trigger values for physical and chemical stressors for south-

west Australia for slightly disturbed inshore marine ecosystems

(ANZECC, 2000);...................................................................................123


Quantifying Submarine Groundwater Discharge v and Nutrient Discharge into Cockburn Sound

Table 22: Receiving water quality protection criteria for discharge of

groundwater from aquifer dewatering (WRC, 1999)..............................124

Table 23: Estimates of nutrient discharge into Cockburn Sound via SGD...............126

Table 24: Nitrogen loads for nominated sections of coastline..................................128

Table 25: Nitrogen loads for nominated sections of coastline..................................136

Table 26: Estimates of total nitrogen discharge into Cockburn Sound via

groundwater, and equivalent average values of SGD and nitrogen

concentrations in groundwater ................................................................143

Table 27: Composite estimate of total nitrogen discharge into Cockburn Sound

via groundwater ......................................................................................147

Table 28: Summary of tidal analysis for K1 tidal constituent ..................................170

Table 29: Summary of tidal analysis for O1 tidal constituent ..................................170


vi Coast and Clean Seas Project No. WA9911

LIST OF FIGURES

Figure 1: Study location map........................................................................................3

Figure 2: Regional Geomorphology .............................................................................6

Figure 3: Groundwater Flow Systems of the Superficial Aquifer ................................7

Figure 4: Subcrop Formations beneath the Superficial Aquifer ...................................8

Figure 5: Coastal Surface Geology...............................................................................9

Figure 6: Hydrological cross section (Section A) through the northern ends of

Garden Island and Cockburn Sound .........................................................11

Figure 7: Hydrological cross section (Section B) through the southern end of

Garden Island and James Point.................................................................12

Figure 8: Hydrological cross section (Section C) through the southern ends of

Garden Island and Cockburn Sound .........................................................13

Figure 9: Groundwater bore locations ........................................................................21

Figure 10: Combined thickness of Quaternary and Tertiary sediments (m);..............22

Figure 11: Thickness of Tamala Limestone (m);........................................................23

Figure 12: Thickness of basal clay (m);......................................................................24

Figure 13: Thickness of Quaternary sands (m);..........................................................25

Figure 14: Water table map of the Superficial Aquifer inshore from Cockburn

Sound based on 1998 water level data......................................................29

Figure 15: Tidal fluctuations in groundwater levels inshore from Northern

Harbour from November 2000 to June 2001; sampling interval for

water levels is 1 hour ................................................................................30

Figure 16: Hydrological cross section (Figure 5, Section D) through the

Northern Harbour study site depicting two possible types of

interaction between the coastal saltwedge and saline water in Lake

Coogee; .....................................................................................................35

Figure 17: Northern Harbour (SGD intercomparison) study site ...............................38

Figure 18: Bulk-ground EC measurements by Steiglitz (reproduced from:

Steiglitz, 2001b); ......................................................................................46

Figure 19: R-Transect; bulk-ground EC measurements (reproduced from:

Steiglitz, 2001b)........................................................................................47

Figure 20: E-Transect; bulk-ground EC measurements (reproduced from:

Steiglitz, 2001b)........................................................................................48

Figure 21: B-Transect; bulk-ground EC measurements (reproduced from:

Steiglitz, 2001b)........................................................................................48


Quantifying Submarine Groundwater Discharge vii and Nutrient Discharge into Cockburn Sound

Figure 22: Bulk-ground EC measurements; additional Northern Harbour

transects; ...................................................................................................49

Figure 23: Site map of shore-based experimental design during the SGD

intercomparison; (reproduced from: Krupa et al., 2000) ..........................53

Figure 24: Benthic flux measurements recorded during the SGD

intercomparison; East Transect, station E1...............................................54

Figure 25: Benthic flux measurements recorded during the SGD

intercomparison; East Transect, station E2...............................................54

Figure 26: Example heat pulse data from Northern Harbour indicating seawater

flow into sediment; benthic flux from T2-T3 pair ≈ - 36 cm.d-1,

benthic flux from T1-T3 pair ≈ - 21 cm.d-1 ..............................................55

Figure 27: Example heat pulse data from Northern Harbour indicating SGD;

benthic flux from T4-T5 pair ≈ 20 cm.d-1, benthic flux from T4-T6

pair ≈ 9 cm.d-1 ...........................................................................................55

Figure 28: Benthic flux measurements from Northern Harbour on 8/5/01 from

11:24 AM to 5:16 PM; measurements are every two minutes..................56

Figure 29: Seismic tracks in Northern Harbour and Cockburn Sound

(reproduced from: Steiglitz, 2001b)..........................................................59

Figure 30: Northern Harbour seismic features (reproduced from: Stieglitz,

2001a) .......................................................................................................60

Figure 31: Cockburn Sound seismic features (reproduced from: Stieglitz,

2001a) .......................................................................................................61

Figure 32: Groundwater temperature profiles on 29/2/00 in selected Northern

Harbour monitoring bores; profiles are indicated over the full

depths of bores ..........................................................................................63

Figure 33: Groundwater EC profiles on 29/2/00 in selected Northern Harbour

monitoring bores; profiles are indicated over the screened sections

of bores only .............................................................................................64

Figure 34: Groundwater temperature profiles on 28/11/00 in selected Northern

Harbour monitoring bores; profiles are indicated over the full

depths of bores ..........................................................................................65

Figure 35: Groundwater EC profiles on 28/11/00 in selected Northern Harbour

monitoring bores; profiles are indicated over the screened sections

of bores only .............................................................................................66

Figure 36: Challenger Beach study site ......................................................................68

Figure 37: Bulk-ground EC measurements; Challenger Beach transects; ..................72

Figure 38: BP Refinery study site ...............................................................................75


viii Coast and Clean Seas Project No. WA9911

Figure 39: Bulk-ground EC measurements; BP Refinery transects; ..........................79

Figure 40: Benthic flux measurements from BP Refinery on 12/5/01 from

8:56 AM to 11:10 AM; measurements are every two minutes.................80

Figure 41: Porewater sampling locations ...................................................................89

Figure 42: Porewater ionic chemistry.........................................................................90

Figure 43: Porewater chloride-EC relation.................................................................90

Figure 44: Porewater dilution index ...........................................................................91

Figure 45: Groundwater nitrogen concentrations [mg.l-1] estimated from

submarine porewater analyses ..................................................................92

Figure 46: Groundwater phosphorus concentrations [mg.l-1] estimated from

submarine porewater analyses ..................................................................93

Figure 47: Conceptual models of SGD into Cockburn Sound ...................................98

Figure 48: Cockburn Groundwater Area flow model; finite difference grid,

boundary conditions and hydraulic conductivity zones..........................109

Figure 49: Cockburn Groundwater Area flow model; base of Superficial

Formations ..............................................................................................110

Figure 50: Cockburn Groundwater Area flow model; groundwater abstraction

and artificial recharge locations and rates...............................................111

Figure 51: Cockburn Groundwater Area flow model; groundwater recharge

rates for the low-recharge scenario, recharge is expressed as a

percentage of mean annual rainfall of 0.87 m/yr ....................................112

Figure 52: Cockburn Groundwater Area flow model; groundwater recharge

rates for the high-recharge scenario, recharge is expressed as a

percentage of mean annual rainfall of 0.87 m/yr ....................................113

Figure 53: Cockburn Groundwater Area flow model; calibrated water table

elevations and model computed SGD for the low-recharge and

high-recharge scenarios ..........................................................................114

Figure 54: Abrupt saltwater interface profiles from the solution of Glover

(1959)......................................................................................................116

Figure 55: Last-recorded, total nitrogen concentrations in groundwater [mg.l-1]; ...122

Figure 56: Model simulated SGD for low-recharge scenario...................................129

Figure 57: Model simulated SGD for high-recharge scenario..................................130

Figure 58: Estimated nitrogen load into Cockburn Sound based on model

simulated SGD (low-recharge scenario) and groundwater nutrient

concentrations estimates from porewater sampling;...............................131


Quantifying Submarine Groundwater Discharge ix and Nutrient Discharge into Cockburn Sound


simulated SGD (high-recharge scenario) and groundwater nutrient


Figure 60: Estimated phosphorus load into Cockburn Sound based on model








concentrations measured by Appleyard (1994).......................................137










Figure 66: Estimates of total nitrogen discharge into Cockburn Sound via

groundwater ............................................................................................142

Figure 67: Estimates of total nitrogen discharge into Cockburn Sound via

groundwater with data plotted according to equivalent average

values of SGDs and nitrogen concentrations in groundwater.................143

Figure 68: Tidal data and Power Spectral Density for Fremantle for the period

1/9/2000 to 16/1/2001; sampling interval is 10 minutes.........................160

Figure 69: Cockburn Sound tidal data and ‘repaired’ Cockburn Sound tidal

data for the period 25/11/2000 to 3/1/2001; sample interval is 10

minutes....................................................................................................161

Figure 70: Hydrograph and Power Spectral Density for groundwater

monitoring bore NH1a for the period 25/11/2000 to 27/12/2000;

sampling interval is 10 minutes ..............................................................162


monitoring bore NH2a for the period 25/11/2000 to 27/12/2000;



x Coast and Clean Seas Project No. WA9911


monitoring bore NH2b for the period 25/11/2000 to 27/12/2000;



monitoring bore NH3 for the period 25/11/2000 to 27/12/2000;








Figure 76: Non-dimensional plot of normalised lag verses tidal efficiency for

groundwater monitoring bores NH1 to NH5 ..........................................171

Figure 77: Heat-pulse BFM ......................................................................................176

Figure 78: Example heat-pulse data from laboratory calibration; flow rate =

2.8 mm.s-1 ...............................................................................................177

Figure 79: Calibration curve for the heat-pulse BFM using inner thermistor

pairs T2-T3 and T4-T5 ...........................................................................177

Figure 80: Calibration curve for the heat-pulse BFM using outer thermistor

pairs T1-T3 and T4-T6 ...........................................................................178

Figure 81: Tide-pumping model; amplitudes of tidal groundwater fluxes across

the coastal boundary calculated for typical values of tidal range and

aquifer properties ....................................................................................184

Figure 82: Wave-pumping model; amplitudes of wave-induced groundwater

fluxes across the seabed calculated for typical values of wave height

and aquifer properties .............................................................................185


Quantifying Submarine Groundwater Discharge xi and Nutrient Discharge into Cockburn Sound

EXECUTIVE SUMMARY

The Study

During the last ten years, an apparent divergence between estimates of nutrient

inputs into Cockburn Sound and trends in marine water quality has been observed.

Chlorophyll-a levels in Cockburn Sound have only improved slightly during this time

despite an estimated 70% reduction in the total anthropogenic nitrogen load into the

sound.

D.A Lord & Associates (2001) estimated that approximately 70% of the

anthropogenic nitrogen load into Cockburn Sound was contributed by submarine

groundwater discharge (SGD); however, the estimates of groundwater inputs were

subject to large errors. To assess management options for improving the water

quality in Cockburn Sound, improved estimates of SGD rates and associated nutrient

loads are required.

This study involved an integrated approach, combining a range of measurements

and techniques, to estimate groundwater nutrient loadings and identify major

nutrient inputs into Cockburn Sound. The investigation was conducted by CSIRO

Land and Water as part of the Coast & Cleans Seas Project WA9911, which involved

two components:

1. Mapping and quantifying submarine groundwater discharge and nutrient

discharge into Cockburn Sound, and

2. At one “hot spot”, demonstrating innovative clean up of a groundwater

nutrient plume using a biologically active, in situ treatment wall.

This report is a full account of the activities that were undertaken to accomplish

component one of the study. Component two activities are reported separately by

Patterson et al. (see references).

Hydrogeology of Cockburn Sound

The Perth Basin contains up to 12,000 meters of marine and continental sediments.

The stratigraphic sequence to a depth of around 2,000 metres below the present land

surface elevation contains Jurassic and Cretaceous age sediments that are overlain

by a relatively thin veneer of late Tertiary to Quaternary age sediments. These


xii Coast and Clean Seas Project No. WA9911

upper-most sediments are known, collectively, as the superficial formations. The

present study is focused on estimation of SGD from these formations.

The saturated extent of the superficial formations forms a mostly unconfined aquifer

system called the Superficial Aquifer. Recharge to the aquifer occurs by percolation

of rainfall through well-drained, sandy soils. There is some leakage of groundwater

to the confined aquifer system beneath but most unutilised groundwater discharges

eventually to the Indian Ocean, Swan-Canning River estuary system and other

regional surface drains. Diffuse replenishment and lateral drainage of shallow

groundwater results in a number of distinct flow systems that can be recognised as

mounds in the shallow groundwater table.

The Cockburn Sound study area lies mostly on the western edge of the Jandakot

Mound and partly within the Safety Bay Mound. The superficial formations in this

area have a saturated thickness of around 25 to 30 metres and unconformably

overlie an erosional surface of Cretaceous sediments that act as a low-permeability

base to the Superficial Aquifer. Groundwater in the Superficial Aquifer flows

generally in a westerly direction and discharges to the near shore marine

environment along the coastline of Cockburn Sound. There are virtually no surface

drains because the surface sands are permeable enough to prevent significant

surface runoff.

The coastal strip of aquifer is characterised by very high transmissivity and very

small hydraulic gradients due to secondary porosity in the Tamala Limestone. In

general, groundwater in the upper sand aquifer moves downward into the

underlying, high-permeability limestone, which acts as a regional drainage layer

and conducts groundwater toward the coast. Locally, groundwater in the Tamala

Limestone can be confined by a clay layer at the base of the overlying sands, which

restricts drainage of the sands at these locations.

Site Investigations

Field investigations were concentrated at three locations along the Cockburn Sound

coast:

Northern Harbour

Challenger Beach

BP Refinery


Quantifying Submarine Groundwater Discharge xiii and Nutrient Discharge into Cockburn Sound

Methodologies for SGD estimation included physical, geochemical, isotopic and

geophysical approaches.

Results from four type of benthic flux (seepage) meter revealed that benthic flux

across the seabed was highly variable and strongly influenced by sea level

fluctuations at time scales ranging from seconds to months. Results indicated that

terrestrial SGD across the seabed, and tide-forced and wave-forced benthic fluxes,

were of a similar order of magnitude and could not be effectively distinguished from

one another. To attain an estimate of the terrestrial component of SGD would

require a long time series of benthic flux measurements that span the broad

frequency spectrum of wave and tide phenomena. Due to transient effects of sea

level change and the patchy distribution of SGD, it was concluded that benthic flux

meters were not suited to obtaining regional scale estimates of groundwater

discharge.

Measurements of electrical conductivity (EC) of submarine porewater revealed

different patterns of groundwater discharge at different locations along the

coastline. A narrow shoreline zone of SGD—as indicated by fresher porewater—

was typical at most locations. Fresh porewater was also detected in patches further

from the shore (the maximum distance from shore of sampling was approximately

56 m). Porewater composition typically varied from that of pure seawater to

relatively fresh groundwater over distances across the seabed of only several meters.

It seemed likely that groundwater was also discharging further offshore but beyond

the limits of the EC transects.

Continuous monitoring of the seawater radon inventory in Northern Harbour

revealed a quasi-inverse relation between diurnal tides and groundwater discharge

rates estimated by this method. This relation was consistent with diurnal

groundwater discharge rates measured by the benthic flux meters.

Seismic surveys in Northern Harbour and Cockburn Sound revealed a number of

stationary water-column features that were consistent with fluid migration from the

seabed. It was beyond the scope of the investigation to determine whether the

observed density anomalies in the water column were due to fresh groundwater

discharge, gas migration from seabed sediment, or another cause. Nevertheless, the

results were highly suggestive of focused SGD, possible through offshore solution

features in the coastal limestone. Although seismic techniques cannot be used to


xiv Coast and Clean Seas Project No. WA9911

calculate discharge rates, they have been used successfully at other locations in

Australia to map the occurrence of focused SGD.

Submarine Porewater Survey

The above investigations were complimented by an onshore (groundwater) and

offshore (porewater) survey of nutrient concentrations along the Cockburn Sound

foreshore between Woodman Point and Palm Beach in Rockingham.

Approximately 20 groundwater samples were collected near Northern Harbour and

80 porewater samples were collected between Woodman Point and Palm Beach

during the period September 2001 to June 2002. Porewater and Groundwater

samples indicated N concentrations up to 130 mg.l-1. Porewater samples were

obtained from depths of approximately 0.3-0.5 m below the seabed and within 10

metres of the shoreline using an EC probe to locate the freshest porewater and a

spear probe to extract the sample. All samples were analysed for nutrient

concentrations, chloride concentration, electrical conductivity and pH.

Ionic balances on 18 samples indicated that the dominant porewater chemistry was

that of seawater diluted by various portions of fresh groundwater. The portion of

fresh groundwater contributing to a porewater sample was estimated based on the

chloride concentration of the sample. The approximate nutrient concentration in

groundwater was then determined by assuming all nutrients in the porewater

mixture were derived from the groundwater component.

High nutrient concentrations in submarine porewater detected along the shoreline

south of James Point were consistent with results from previous groundwater

investigations, which revealed high levels of nutrient contamination in this area.

Total nitrogen concentrations greater than 100 mg.l-1 were measured in porewater

less than 0.5-m below the seabed and greater than 50-m offshore. Total nitrogen

concentrations in groundwater were measured at 200 mg.l-1.

High nitrogen concentrations were also detected along a 40-m section of the

shoreline north of James Point and adjacent the HIsmelt facilities. Total nitrogen

concentrations in groundwater of up to 740 mg.l-1 were estimated based on

submarine porewater concentrations of up to 220 mg.l-1. The source of

contamination is unknown; however, there is evidence from groundwater monitoring

of significantly elevated nitrate levels in the inshore aquifer.


Quantifying Submarine Groundwater Discharge xv and Nutrient Discharge into Cockburn Sound

Slightly elevated nutrient concentrations were detected along a section of shoreline

north of Challenger Beach, corresponding approximately to the location of the

Naval Base caravan park. Nutrient enriched groundwater might be the result of

leaching from septic tanks, although the source of contamination has not been

investigated. The average total nitrogen concentration of groundwater in this area

was approximately 8 mg.l-1.

At remaining locations along the shoreline, the average total nitrogen concentration

of groundwater was approximately 4 mg.l-1. Contrary to previous groundwater

investigations, nutrient enriched groundwater was not detected in the Northern

Harbour porewater samples even though high nutrient levels are observed in

groundwater. This provided further evidence that the Northern Harbour plumes

probably discharge further offshore and beyond the shoreline strip of seabed from

which the porewater samples were collected.

Submarine Groundwater Discharge

Several conceptual models of groundwater discharge into Cockburn Sound are

feasible based on the principle that groundwater in the Superficial Aquifer flows

generally toward ocean along pathways of least resistance. In a homogeneous

aquifer, flow is mainly toward the shoreline because it represents the shortest travel

distance to the nearest point of discharge. In an inhomogeneous aquifer,

groundwater may discharge further from the shore if either a low-conductivity

sediment layer restricts vertical flow toward the shoreline, or high-conductivity

zones, such as fractures and solution features, focus SGD to other locations. Both of

these possibilities occur in Cockburn Sound.

Absence of offshore hydrogeological data at a scale that is relevant to SGD is the

main limitation to identifying which conceptual model of groundwater discharge is

most likely at particular locations along the coast. In particular, the presence and

extent of clay aquitard between the Tamala Limestone and overlying sands is poorly

described across most of the study area. In addition, the scale and structure of

secondary porosity in the coastal limestone is unknown.

Previous estimates and implied values of SGD rates based on groundwater flownet

analyses have varied between 0.7 and 8 m3.d-1.m-1 (i.e. by an approximate order of

magnitude). Similar values of SGD were derived by different investigators but the

associated calculations of groundwater velocity were often based on significantly

different values of hydraulic conductivity and hydraulic gradient.


xvi Coast and Clean Seas Project No. WA9911

A groundwater flow model was applied to quantify the aquifer water balance inshore

from Cockburn Sound and estimate regional values of SGD for each 100-m section

of the coast. Modelled discharges varied spatially along the length of the coastline

due to the assigned variation in hydraulic connection between the aquifer and ocean

in the model, distributed groundwater recharge, spatial variability in aquifer

transmissivity and pumping, and focusing of groundwater flow toward coastal

embayments.

Total average SGD into Cockburn Sound between Woodman Point and the southern

boundary of the model was estimated to be 2.8-5.5 m3.d-1.m-1. The estimated range

in average SGD corresponds to low and high estimates of groundwater recharge.

An assumption of high groundwater recharge implies a larger SGD rate, and visa

versa.

Differences assigned to in the hydraulic connection between the aquifer and ocean in

the model had the largest affect on the spatial distribution of SGD. The modelled

distributions of SGD are not corroborated against independent data—since none

exists—and are considered to have large uncertainty. The modelling provided

regional-scale estimates of SGD rates, which cannot take account of the detailed

hydrogeology and associated flow paths that control SGD at the local scale.

Nutrient Discharge

Nutrients in the Superficial Aquifer enter Cockburn Sound by SGD of terrestrial

groundwater and aquifer ‘flushing’ by seawater. The mobile, un-utilised fractions of

nutrients that enter the inshore aquifer are eventually discharged to the marine

environment by advection with regional groundwater flow. It is probable that

seawater movement into and out of coastal sediments also plays a role in

transporting nutrients to the submarine environment. Seawater entering the aquifer

during periods of rising sea level can contact and mix with nutrient enriched

groundwater and then drain back to the ocean during a subsequent period of falling

sea level. This exchange mechanism may be important in locations where the

aquifer beneath a contaminant source is regularly flushed by saltwater flows.

Nitrogen is the nutrient of most concern and has been monitored more intensive than

phosphorus. To date, significant point source emissions of phosphorus into

Cockburn Sound have not been detected and ambient discharges are relatively small.

Previous estimates of groundwater nitrogen loads into Cockburn Sound have varied


Quantifying Submarine Groundwater Discharge xvii and Nutrient Discharge into Cockburn Sound

in the range 180 to 450 tonnes.yr-1 during the period 1978 to 2000. The most recent

estimate by D.A. Lord (2000) was 220 tonnes.yr-1.

Uncertainty in estimates of nutrient loads is non-quantifiable and stems from

uncertainty in the SGD rates and spatial interpolation of nutrient concentrations in

the aquifer. In particular, it is unclear how SGD rates and nutrient concentrations

measured in groundwater bores should be combined to estimate nutrient loads. In

previous estimates, it has been assumed that nutrients in the upper sand aquifer

travel laterally through the sand and discharge at the coast. This conceptual model

ignores vertical draining of groundwater into the underlying, high-permeability

limestone and is highly unlikely except where the limestone is confined by a clay

aquitard that restricts vertical flows.

To improve our understanding of groundwater nutrient transport into Cockburn

Sound, annual SGD nutrient loads computed in this study and previous studies need

be compared with estimates of mass loadings at the source locations (e.g. based on

loss of product or leaching rates). It is unknown whether these estimates have been

made. Since porewater sampling affords a reasonably high level of confidence in

estimates of nutrient concentrations at the location of SGD, inconsistencies between

nutrient mass loadings at source and discharge locations are likely the results of

errors in the estimated SGD rates.

Spatial variability in groundwater recharge and flow, control the transport

pathways and travel times for nutrients to move from source areas inshore from

Cockburn Sound to the ocean. The presence of preferred flow paths mean that some

areas of the aquifer are likely to contain less active—slower flowing—areas of

groundwater compared with other parts of aquifer, which contain the preferential

flow paths. Nutrients that are mobilized into relatively inactive areas of the flow

system are likely to take longer to reach Cockburn Sound. In addition, because the

flushing time is relatively longer, these areas of aquifer are also likely to contain

nutrients for longer periods following contamination events.

Estimates of nutrient mass fluxes into Cockburn Sound were calculated for the

following combinations of SGD rates and nutrient concentrations:


xviii Coast and Clean Seas Project No. WA9911

SGD Rates Nutrient

Concentrations

Nitrogen Load

[tonnes.yr-1]

Phosphorus Load

[tonnes.yr-1]

Model - low recharge Porewater 144 2.8 Model - high recharge Porewater 305 5.6 Model - low recharge Appleyard (1994) 317 1.3 Model - high recharge Appleyard (1994) 585 2.5

A composite estimate of 234±88 tonnes.yr-1 total nitrogen load into Cockburn Sound

was calculated based on the following assumption:

• The model-derived values of SGD were judged the most realistic because

they incorporated the inshore groundwater balance. Groundwater discharge

rates used by Appleyard (1994) thus appear to be comparatively small.

• Groundwater nutrient concentrations are likely to have changed significantly

since 1994. The submarine porewater survey provides the only current and

contemporaneous set of groundwater nutrient concentrations along the

Cockburn Sound coast. This data set includes previously undetected areas of

nutrient-enriched groundwater that discharge from north of James Point. In

addition, porewater samples were collected at the location of groundwater

discharge. More samples were collected along the coastline compared to

previous surveys of groundwater bores, and the density of sampling was

increased at locations where significant nutrient contamination was detected.

• Because the Northern Harbour plumes were not detected by the porewater

survey, a previous estimate of nitrogen mass flux into Northern Harbour

(PPK 2000, Table 3.2, Scheme 4) was adopted. The estimate allows for a

51% reduction in nitrogen loading to the harbour due to plume recovery

operations.

• Other factors of importance in determining effective management strategies

for reducing environmental impacts include estimation of the time of arrival

and discharge of on-shore groundwater contaminants into the near-shore

environment, improvement of the management of on-shore groundwater

resources, and better understanding of seawater intrusion. The latter factor

is regarded as a coupled but inverse process to SGD. However, the central

question of establishing a causal link between SGD-derived nitrogen input


Quantifying Submarine Groundwater Discharge xix and Nutrient Discharge into Cockburn Sound

and the extent of, trigger levels and conditions for occurrence of algal

blooms in the near shore environment remains unanswered even though a

high level of confidence can now be claimed in terms of the net SGD flux and

nitrogen loads into Cockburn Sound.

A potential inadequacy of the porewater survey was the possibility that porewater

samples collected within 10 m of the shoreline were not representative of

groundwater discharging further from the shore. This uncertainty cannot be

resolved without a more detailed investigation of local-scale SGD processes.


xx Coast and Clean Seas Project No. WA9911

ACKNOWLEDGEMENTS

Funding Agencies


National Heritage Trust Coast & Clean Seas Program (Project No. WA9911)

Centre for Groundwater Studies

Kwinana Industries Council

Department of Industry and Technology (former Dept. of Commerce and Trade)

In-kind Contributions

Assistance received from the following individuals and organisations is gratefully

acknowledged:

• The Water and Rivers Commission of Western Australia; hydrological and

hydrogeological data supplied from the state water resource database

• Department of Transport Maritime Division; tide data for Fremantle

• PPK Environment and Infrastructure; groundwater level and nutrient data for

bores in the Jervoise Bay area.

• Steve Krupa* (South Florida Water Management District); seepage

measurements (“Krupaseep”) and environmental data from the Northern Harbour

SGD intercomparison

• Makoto Taniguchi* (Nara University of Education, Japan); electrical conductivity

and temperature profiles for groundwater bores in the Jervoise Bay area

• Christopher Smith*, Daniel O’Rourke* and Ronald Paulsen* (Cornell

Cooperative Extension, New York); seepage measurements (ultrasonic) from the

Northern Harbour SGD intercomparison

• Simon Nield (Nield Consulting, Western Australia); documentation and

computer modelling files for the Cockburn Groundwater Area groundwater flow

model

• Thomas Stieglitz* (James Cook University, Queensland); data and reports on

porewater electrical conductivity and seismic surveys from the Northern Harbour


Quantifying Submarine Groundwater Discharge xxi and Nutrient Discharge into Cockburn Sound

SGD intercomparison, and design details and parts for construction of porewater

EC probe*

• The Scientific Committee on Oceanic Research (SCOR) and the Land-Ocean

Interactions in the Coastal Zone (LOICZ) project of IGBP, with support of the

International Oceanographic Commission (IOC) for financial support of travel

and equipment costs for the SGD intercomparison at Cockburn Sound

• Bill Burnett and Mike Lambert (Florida State University, USA); seepage

measurements (continuous radon monitoring) from the Northern Harbour SGD

intercomparison

• Claus Otto and Lorraine Bates (CSIRO Land and Water); bore completion details

for groundwater monitoring bores on Garden Island

• Gerald Watson and Robert Woodbury (CSIRO Land and Water); technical and

logistical support

Steering Committee Members

Chris Barber (Centre for Groundwater Studies) - Chairman

Mat Vanderklift (Ministry for Planning) - WA Coast and Clean Seas Coordinator

John Braid – former WA Coast and Clean Seas Coordinator

Greg Davis (CSIRO Land and Water)

Steve Morris (Department of Industry and Technology)

Mark Nener (Water Corporation of Western Australia)

Tony Smith (CSIRO Land and Water)

Bradley Patterson (CSIRO Land and Water)

Steve Appleyard (Water and Rivers Commission)

Phil Manning (Water and Rivers Commission)

Phil Hine (Department of Environmental Protection)

Rod Lukatelich (BP Oil) - Kwinana Industries Council representative

* Supported by the SCOR-LOICZ-IOC SGD Intercomparison Programme


Quantifying Submarine Groundwater Discharge 1 and Nutrient Discharge into Cockburn Sound

1 INTRODUCTION

1.1 BACKGROUND

The waters of Cockburn Sound are the most intensively used marine system in

Western Australia (Cockburn Sound Management Council, 2001). Decline in the

health of seagrass in Cockburn Sound first became apparent in the 1950s and by the

mid-1990s approximately 80% of the original meadows had been lost (Department

of Environmental Protection, 1995). A major study in the late 1970s (Department of

Conservation and Environment, 1979) attributed seagrass death to eutrophication and

smothering of plants by epiphytic algae, which were thriving in the nutrient-enriched

waters. Waste disposal by major industries, ocean discharge of sewerage effluent,

and inshore contamination of groundwater that is known to discharge via submarine

groundwater discharge (SGD) into Cockburn Sound were identified as the main

factors contributing to nutrient enrichment. Management responses to reduce

nutrient inputs appeared initially to have been successful and measurable

improvements in marine water quality had occurred by the early-1980s. However,

by the 1990s, chlorophyll-a concentrations had re-deteriorated and were close to the

worst levels experienced in the 1970s (Department of Environmental Protection,

1995; D.A Lord & Associates, 2001). In addition, episodic marine algal blooms

(diatoms) have been observed.

During the last ten years, there has been an apparent divergence between estimates of

nutrient inputs into Cockburn Sound and trends in marine water quality.

Chlorophyll-a levels have only improved slightly during this period despite an

estimated 70% reduction in the total anthropogenic nitrogen load into the sound.

This has raised concerns regarding the accuracy of estimates of nutrient loads,

particularly the quantity of nutrients that enter Cockburn Sound via submarine

groundwater discharge (SGD). Recently, D.A Lord & Associates (2001) estimated

that approximately 70% (219 tonnes.yr-1) of the anthropogenic nitrogen load into

Cockburn Sound was contributed by SGD.

To assess management options for improving water quality in Cockburn Sound, the

accurate determination SGD rates and associated nutrient loads is essential. It

remains a difficult strategic problem to link estimates of SGD nitrogen inputs into


2 Coast and Clean Seas Project No. WA9911

Cockburn Sound to the frequency and magnitude of algal blooms and their

subsequent impact on the marine environment of Cockburn Sound.

1.2 THIS STUDY

The objective of this study was to use an integrated approach involving a range of

measurements and techniques to estimate groundwater nutrient loadings and identify

major inputs into Cockburn Sound.

The investigation was conducted by CSIRO Land and Water as part of the Coast &

Cleans Seas Project WA9911, entitled: Quantifying Submarine Groundwater

Discharge and Demonstrating Innovative Clean-Up to Protect Cockburn Sound from

Nutrients.

This joint project involved two components:

1. Mapping and quantifying submarine groundwater discharge and nutrient

discharge into Cockburn Sound, and

2. At one “hot spot”, demonstrating innovative clean up of a nutrient plume by

installing a biologically active, in situ treatment wall across the direction of

groundwater flow.

A full account of the work undertaken to accomplish the first component is set out in

this report. Details of the insitu remediation trial are reported separately by Patterson

et al. (2002).

1.3 REPORT FORMAT

This report is divided into five sections. Section 1 (this section) provides a brief

background to the study, describes the study motivations, and sets out its objectives.

Section 1 presents an overview of the regional hydrogeology and reviews existing

hydrological data that is relevant to the assessment of SGD. Section 1 details field

investigations and results. Sections 0 and 5 present assessments of SGD and nutrient

discharge into Cockburn Sound, respectively. Conclusions and major finding are

summarised in the Executive Summary at the front of the report.



��

��

��

��

�

�

��

��

��

��

��

��

��

��

� � ��

! �

"��

��

�

� � � � ��

� ��

#$%%%%

#$%%%%

#&%%%%

#&%%%%

'%%%%%

'%%%%%

(')%%%% (')%%%%

('#%%%% ('#%%%%

(''%%%% (''%%%%

('*%%%% ('*%%%%

('(%%%% ('(%%%%

��

��

Figure 1: Study location map



2 HYDROGEOLOGY

2.1 REGIONAL SETTING

The Perth Basin lies to the west of the Darling Fault (Figure 2) and contains up to

12,000 meters of marine and continental sediments. The stratigraphic sequence to a

depth of around 2,000 metres below the present land surface elevation contains

Jurassic (210-140 Ma) and Cretaceous (140-65 Ma) age sediments that are overlain

by a relatively thin veneer of late Tertiary to Quaternary age sediment. The Tertiary

and Quaternary deposits, known as the superficial formations, vary in thickness up to

a maximum of approximately 110 meters collectively.

The saturated extent of the superficial formations forms a mostly unconfined aquifer

system that is called the Superficial Aquifer. Recharge to the aquifer occurs through

direct vertical infiltration of rainfall, whereas groundwater drains primarily by lateral

flow through the aquifer. There is some leakage of groundwater to the confined

aquifer system beneath but most unutilised groundwater discharges eventually to the

Indian Ocean, Swan-Canning River estuary system and other regional surface drains.

The result is a number of distinct regional groundwater mounds and flow systems

that are located between the regional discharge boundaries (Figure 3). SGD from the

Superficial Aquifer is the topic of this report.

The Cockburn Sound study area lies mostly on the western edge of the Jandakot

Mound and partly within the Safety Bay Mound. Over most of the study area, the

superficial formations have a saturated thickness of around 25 to 30 metres and

unconformably overlie an erosional surface of Cretaceous sediments that act as a

low-permeability base to the Superficial Aquifer (Figure 4). In the southern part of

the study area, the superficial formations are underlain by Rockingham Sand, which

occupies an erosional channel in the Cretaceous sediments. This is known to be

more than 100 metres deep in places.

Exchange of groundwater between the Superficial Aquifer and Cretaceous-Jurassic

artesian aquifer system varies regionally. Leakage between aquifers is only crudely

defined at local scales.



��

��

��

��

�

�

��

��

��

��

��

��

��

��

� � � � ��

�

��

�

� � � ��

�� !!"#

��$��%��&��'��(��)��

��&��*��&��$��+��%,��

-��.��-��

./�/0�

��

��

��

��

��

��

! "�� ! "��

! �� ! ��

! �� ! ��

! #�� ! #��

! !�� ! !��

Figure 2: Regional Geomorphology



��

��

��

��

�

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� � � � ��

�

��

�

� � � � ��

�� !��"�� #$$%&'��"�� #$$%&

�� !�� (��

��

�� )�(��!��"�&

*+ +,�

��

��

��

��

��

��

! "�� ! "��

! �� ! ��

! �� ! ��

! #�� ! #��

! !�� ! !��

Figure 3: Groundwater Flow Systems of the Superficial Aquifer



��

��

�� ! � ��"��! �

�#$#%&

��

��

��

��

�

�

��

��'

��(

)*)

+,*),*)

+-)*)

+.)*)

+-,*)

-)*)

-,*)

.)*)

+.,*)

+/)*)+/,*)

,*)

+,*)

+/)*)��

��

��

��

��

� � � � ��

�

��

�

. ) . 0 ��

&��1��&��-22,�#��&��-22,�

��

��

��

��

��

��

! "�� ! "��

! �� ! ��

! �� ! ��

! #�� ! #��

! !�� ! !��

Figure 4: Subcrop Formations beneath the Superficial Aquifer



��

��

��

��

�

�

��

��

��

��

��

��

��

��

� � � � ��

�

��

�

� � � ��

�� !�"

��#��$%��#��$"&��%��

��$��&��%��"��$��"�'��(�$�(��

��

��

��

��

��

��

��

! "�� ! "��

! �� ! ��

! �� ! ��

! #�� ! #��

! !�� ! !��

Figure 5: Coastal Surface Geology



2.2 SUBCROP FORMATIONS

The superficial formations mostly overlie Cretaceous sediments, which include the

Kardinya Shale Member of the Osborne Formation and the Pinjar and Waneroo

Members of the Leederville Formation. A different situation exists in the southern

part of Cockburn Sound where the Rockingham Sand underlies the superficial

formations. The interpreted regional extent of subcrop units (Davidson 1995) is

based on drilling data from the mainland and offshore on Garden Island.

Hydrogeological cross sections through Garden Island and Cockburn Sound are

presented in Figure 6 to Figure 8. The following hydrogeological descriptions of the

subcrop units are based mainly on the synthesis of Davidson (1995).

2.2.1 KARDINYA SHALE MEMBER

The Kardinya Shale Member consists of moderately to tightly consolidated,

interbedded siltstones and shales, which are typically dark green to black, and

includes thin interbeds of mostly fine-grained sandstone. It has relatively low

permeability and is considered to be an aquitard the separates the Superficial Aquifer

from the deeper Leederville Aquifer. Leakage through the Kardinya Shale Member

is considered negligible compared to aquifer flows.

2.2.2 PINJAR AND WANEROO MEMBERS

The Pinjar and Waneroo Members are upper units within the Leederville Aquifer and

consist of discontinuous, interbedded sandstones, siltstones and shale of marine and

non-marine origin. The subcrop of these two units beneath the superficial formations

represents an area of direct contact between the Superficial and Leederville Aquifers.

The intercomparison study area is a region of upward hydraulic gradient where there

is potential for the Leederville Aquifer to discharge to the Superficial Aquifer.

2.2.3 ROCKINGHAM SAND

The Rockingham Sand occupies what is thought to be an erosional channel incised

into the Cretaceous sediments and which probably extends from Rockingham to

offshore beneath the southern end of Garden Island. The unit consists of slightly

silty, medium-grained to coarse-grained sand of shallow marine origin. The

maximum thickness of Rockingham Sand is approximately 110 metres at the

southern end of Cockburn Sound in the Rockingham area.



Coc

kburn

Sound

0m

-40

-80

-120

-160

Superf

icia

lA

quifer

Aquitard

Leederv

ille

Aquifer

Lake

Coo

gee

Jerv

oise

Bay

Par

melia

Ban

kB

asin

Unit

Par

melia

Ban

k

Gard

en

Isla

nd

Ho

loce

ne

sed

iment

Kar

din

yaShal

eM

em

ber

(Osb

orn

eFo

rmat

ion)

Henle

ySan

dst

one

Mem

ber

(Osb

orn

eFo

rmat

ion)

Pin

jar

Mem

ber

(Leed

erv

ille

Fo

rmat

ion)

Wan

nero

oM

em

ber

(Leed

erv

ille

Fo

rmat

ion)

Tam

ala

Lim

est

one

Sp

ear

wo

od

San

d

Kar

din

yaShal

eM

em

ber

(Osb

orn

eFo

rmat

ion)

?

?

?

?

??

SO

UR

CE

:

This

cro

ssse

ctio

nis

bas

ed

on

dat

afr

om

Dav

idso

n(1

99

5)

and

1:5

0,0

00

Envi

rom

enta

l

Geo

logy

Map

Sheets

(GSW

A1

98

6,1

98

5)

Fre

man

tle

and

Ro

ckin

gham

Saf

ety

Bay

San

d

Figure 6: Hydrological cross section (Section A) through the northern ends of Garden Island and Cockburn Sound



Kar

din

yaShal

eM

em

ber

(Osb

orn

eFo

rmat

ion)

Aq

uit

ard

Sup

erf

icia

l

Aq

uifer

Leed

erv

ille

Aq

uifer

Jam

es

Poin

t

Par

melia

Ban

kB

asin

Unit

Par

melia

Ban

k

Gard

en

Isla

nd

Ho

loce

ne

sed

iment

Kar

din

yaShal

eM

em

ber

(Osb

orn

eFo

rmat

ion)

Pin

jar

Mem

ber

(Leed

erv

ille

Fo

rmat

ion)

Wan

nero

oM

em

ber

(Leed

erv

ille

Fo

rmat

ion)

Tam

ala

Lim

est

one

?

?

?

?

??

0m

-40

-80

-12

0

-16

0

?

?

??

Ro

ckin

gham

San

d

Kin

gsP

ark

Fo

rmat

ion

(aq

uit

ard

)

Coc

kburn

Sound

Sp

ear

wo

od

San

d

SO

UR

CE

:

This

cro

ssse

ctio

nis

bas

ed

on

dat

afr

om

Dav

idso

n(1

99

5)

and

1:5

0,0

00

Envi

rom

enta

l

Geo

logy

Map

Sheets

(GSW

A1

98

6,1

98

5)

Fre

man

tle

and

Ro

ckin

gham

Saf

ety

Bay

San

d

Saf

ety

Bay

San

d

Figure 7: Hydrological cross section (Section B) through the southern end of Garden Island and James Point



0m

-40

-80

-12

0

-16

0

Sup

erf

icia

l

Aq

uifer

Leed

erv

ille

Aq

uifer

Roc

kingh

am

Par

melia

Ban

kB

asin

Unit

Par

melia

Ban

k

Gard

en

Isla

nd

Ho

loce

ne

sed

iment

Kar

din

yaShal

eM

em

ber

(Osb

orn

eFo

rmat

ion)

Wan

nero

oM

em

ber

(Leed

erv

ille

Fo

rmat

ion)

Tam

ala

Lim

est

one

?

???

Ro

ckin

gham

San

d

Kin

gsP

ark

Fo

rmat

ion

(aq

uit

ard

)

Coc

kburn

Sound

Saf

ety

Bay

San

d

Ro

ckin

gham

San

d

Tam

ala

Lim

est

one

??

Lake

Coo

loon

gup

SO

UR

CE

:

This

cro

ssse

ctio

nis

bas

ed

on

dat

afr

om

Dav

idso

n(1

99

5)

and

1:5

0,0

00

Envi

rom

enta

l

Geo

logy

Map

Sheets

(GSW

A1

98

6,1

98

5)

Fre

man

tle

and

Ro

ckin

gham

?

?P

inja

rM

em

ber

(Leed

erv

ille

Fo

rmat

ion)

Figure 8: Hydrological cross section (Section C) through the southern ends of Garden Island and Cockburn Sound



2.3 SUPERFICIAL FORMATIONS

Individual stratigraphic units that comprise the superficial formations in the

Cockburn Sound area are—in order of deposition—the Tamala Limestone,

Cooloongup Sand, Becher Sand and Safety Bay Sand. These are known collectively

as the Kwinana Group.

2.3.1 TAMALA LIMESTONE

The Tamala Limestone unit is a calcareous eolianite that unconformably overlies

Cretaceous sediments and Rockingham Sand. It contains various proportions of

quartz sand, fine-grained to medium-grained shell fragments and minor clay lenses.

The limestone typically exhibits secondary porosity in the form of numerous solution

channels and cavities. The average base elevation of Tamala Limestone in the

intercomparison study area is around -25 to -30 metres Australian Height Datum

[mAHD]; see Figure 4.

Garden Island is an offshore outcrop of Tamala Limestone and is part of a former

‘drowned’ dune ridge called the Garden Island Ridge (Seale et al. 1988) which runs

northwest through Point Peron, Garden Island, Carnac Island and Rottnest Island. A

roughly parallel sand and limestone ridge, called the Spearwood Ridge, is located

onshore from the coast along the contact between the Tamala Limestone and Safety

Bay Sand. The area between these ridges is known as the Warnbro-Cockburn

Depression, in which Tamala Limestone also outcrops as submarine reef.

Passmore (1970, Figure 37) depicts a section through the Tamala Limestone and

overlying sediments from the southern end of Garden Island to the mainland across

Cockburn Sound. The section is along a similar line to Section C in this report and

is based on drilling logs from Fremantle Port Authority Line 1 No. 7 (FPA7), Line 10

No. A (FPA10A) and No. F1 (FPAF1). The base of the Tamala Limestone occurs at

approximately –25 to –30 mAHD at these locations, which is consistent with drilling

logs from the mainland.

2.3.2 COOLOONGUP SAND

The Cooloongup Sand unit consists of fine-grained to coarse-grained feldspathic

quartz sand of grey and yellow-brown colour, with variable amounts of shell material

(up to 25%). It unconformably overlies the Tamala Limestone.



2.3.3 BECHER SAND

Becher Sand originated in the near-shore marine environment and consists of grey,

fine-grained to medium-grained quartz and skeletal sand that is mostly structureless

and bioturbated. It unconformably overlies the Tamala Limestone and

unconformably underlies the Safety Bay Sand. Becher Sand extends along the

coastal margin and is typically 10 to 15 metres thick. The base of the unit can locally

contain a layer of silty calcareous clay that is rich in shell fragments.

2.3.4 SAFETY BAY SAND

Safety Bay Sand unconformably overlies Tamala Limestone and Becher Sand and

consists of cream, unlithified, calcareous fine-grained to medium-grained quartz sand

and shell fragments. Traces of fine-grained, black heavy minerals are also present.

Safety Bay Sand is clearly visible along the coastal margin as white, aeolian sand

dunes, which extend offshore into Cockburn Sound as submarine bank units.

2.4 GROUNDWATER OCCURRENCE AND FLOW

2.4.1 ARTESIAN AQUIFER SYSTEM

Artesian aquifers of the Perth Basin comprise a multi-layered aquifer system that is

recharged by slow downward leakage of groundwater through overlying formations.

This system extends offshore beneath the seabed and probably discharges to the

ocean at least several kilometers from the coast—possibly through offshore faults

(Davidson, 1995). Nevertheless, the precise mechanism(s) for artesian groundwater

discharge and locations of discharge are unknown.

It is relevant to note that offshore in the central and southern parts of Cockburn

Sound, the upper members of the Leederville Aquifer—the Pinjar and Waneroo

Members—subcrop directly beneath the Tamala Limestone, which is only 10 to 15

metres thick in this area. The average piezometric head of the Leederville Aquifer in

the same area is around 5 metres above the water table elevation of the Superficial

Aquifer (Davidson, 1995). This indicates that there is potential for artesian

groundwater discharge to occur in some parts of Cockburn Sound; however, it is not

investigated in this study.



2.4.2 ROCKINGHAM AQUIFER

The Rockingham Aquifer is defined in association with the Rockingham Sand and is

locally confined by discontinuous clay lenses at the base of the superficial

formations. Confinement of groundwater in the Rockingham Aquifer is indicated by

propagation of ocean tidal fluctuations more than 150 metres inland from the

shoreline (Passmore, 1970). This suggests either very high aquifer transmissivity—

allowing high rates of tidal groundwater flow—or propagation of the tidal signal

through elastic expansion and compression of the aquifer under confined or semi-

confined conditions. Based on the sediment type, the latter explanation is more

likely.

The ocean saltwedge can penetrate deeper into the Rockingham Aquifer because the

sediments are thicker and deeper than in the Superficial Aquifer. The bottom part of

the aquifer contains seawater to an elevation of around –65 mAHD, while the top 40

metres contain groundwater of salinity less than 1,000 mg.l-1 (Davidson, 1995). Un-

utilised groundwater from the Safety Bay Mound discharges to the ocean over the

top of the saltwedge.

The Rockingham Aquifer is assumed to have similar hydraulic conductivity to sands

in the superficial formations.

2.4.3 SUPERFICIAL AQUIFER

Groundwater in the Superficial Aquifer flows generally in a westerly direction and

discharges to the near shore marine environment along the coastline at Cockburn

Sound and Safety Bay (Figure 3). The aquifer is recharged by infiltration and

percolation of rainfall. There are virtually no surface drains because the surface

sands are permeable enough to prevent significant surface runoff.

Table 1 is a compilation of estimates of aquifer hydraulic properties for the

Superficial Formations inshore from Cockburn Sound. Note that the coastal strip is

characterised by very high transmissivity, due to secondary porosity in the Tamala

Limestone, and has very small horizontal hydraulic gradients. Several kilometers

inland from the coast is a relatively narrow band of lower permeability sediments

that run roughly parallel to the coastline along the contact between the Bassendean

Dunes and Tamala Limestone (see Figure 14). A dramatic change in hydraulic

gradient is the main evidence for the existence of this flow barrier. The East Beeliar

Wetlands, a north-south chain of lakes and wetlands located approximately 5



kilometres from the coast, are a surface expression of higher groundwater levels that

are elevated behind this barrier. Water levels in these lakes are up to 18 metres

above mean sea level.

Horizontal hydraulic gradients in the coastal strip are as small as 0.0002—equivalent

to a 1 metre change in water level over a 5 kilometre horizontal distance—but are

larger than 0.01 across some parts of the flow barrier. Thus, hydraulic gradients

within this region of the aquifer vary by a factor of around 50. Nield (1999)

attributed these steep gradients to the presence of clay in the Tamala Limestone, as

indicated in hydrogeological logs from a dense network of monitoring bores in the

vicinity of Aloca’s residue disposal areas. The fact that horizontal hydraulic

gradients vary by a factor of approximately 50 indicates that there is a similar order

of magnitude variation in the Superficial Aquifer transmissivity and hydraulic

conductivity. Nield (1999) estimated that the hydraulic conductivity of the

Superficial Aquifer in the Cockburn Groundwater Area varies spatially in the range 6

to 1,660 m.d-1, for the assumption of low rainfall recharge, and 8 to 3,000 m.d-1 for

the assumption of high rainfall recharge.

Along the coastal margin of the superficial formations, groundwater in the Tamala

Limestone is locally confined where the shelly clay layer at the base of the Becher

Sand is present. This layer, commonly referred to as ‘basal clay’ or ‘shell bed’ in

drilling logs, is in the order of 1-metre thick. Vertical hydraulic head differences

between groundwaters in the Tamala Limestone and Safety Bay Sand, plus

propagation of tidal fluctuations hundreds of metres into the Tamala Limestone,

provide evidence that the limestone is confined or semi-confined in some locations.

Tidal fluctuations are rapidly attenuated inshore from the coast in the Safety Bay

Sand and Becher Sand units because groundwater is essentially unconfined. Davis et

al. (1994) observed diurnal tidal fluctuations in piezometric head of up to 0.4 metres

in the Tamala Limestone within the BP Refinery, whereas water level fluctuations in

the overlying sands were negligible at the same locations.



Table 1: Estimated hydraulic properties of the Superficial Formations inshore from Cockburn Sound

Reference/Source Property/Value Location/Method

Bodard (1991) T = 190 – 235 m2/d k = 10 – 30 m/d S = 0.3 [1] (unconfined) B = 10 m

Safety Bay Sand

Bodard (1991) T = 1,700 – 2,600 m2/d k = 100 – 250 m/d n = 30% S = 0.3 ± 1.5 (unconfined) S = 0.02 (confined) B = 10 m

Tamala Limestone

Walker (1994) k = 100 – 250 m/d (aquifer unspecified) BP Refinery, slug testing of 20 bores

Davidson (1995) k = 6 – 50 m/d Cockburn Sound area, superficial aquifer

Nield (1999) T = 40,000 m2/d (superficial) Alcoa Kwinana Refinery; pump test

Nield (1999) k = 400 – 1,660 m/d (low recharge scenario) k = 800 – 3,000 m/d (high recharge scenario)

Cockburn Sound coastal strip; Cockburn Groundwater Area model calibration, superficial aquifer

PPK (2000) T = 13,000 m2/d (JBTB1 early-time data) T = 9,400 m2/d (JBTB1 late-time data) T = 19,900 m2/d (JBMB9S recovery) T = 23,800 m2/d (JBMB9S early-time data) S = 2.0E-06 (as above) T = 54,200 m2/d (JBMB9D early-time data) S = 1.4E-01(as above) T = 9,000 m2/d (JBTB2 early-time data) T = 28,800 m2/d (WPM5C early-time data) S = 9.0E-02 (as above) T = 39,700 m2/d (JBMB1early-time data) S = 1.0E-01(as above) T = 25,000 m2/d (adopted mean value) S = 1.0E-01(adopted mean value)

Inshore from Northern Harbour; constant rate pump tests on bores JBTB1 and JBTB2 (observation bores JBMB9S, JBMB9D, WPM5C, JBMB1), superficial aquifer

Nield in PPK (2000) k = 900 m/d (superficial) Northern Harbour; groundwater model calibration

Smith & Hick (2001)

k = 53 - 174 m/d (Safety Bay Sand) Northern Harbour, tidal method

n: porosity k: saturated hydraulic conductivity i: hydraulic gradient B: aquifer saturated thickness T = kB: aquifer transmissivity S: aquifer storage coefficient



2.5 GROUNDWATER BORES

A relatively large number of groundwater bores have been installed into the

Superficial Aquifer adjacent to Cockburn Sound (Figure 9), mainly for water supply

and groundwater monitoring purposes. Water and Rivers Commission (WRC) holds

information about these bores in the State Water Resources Information System

(SWRIS/WIN) and AQWABase. Some Kwinana industries also collect groundwater

data for abstraction and environmental monitoring purposes.

In this study, bore information was gathered from the following sources:

• Water and Rivers Commission - SWRIS/WIN and AQWABase databases.

• Bodard, J.M., Assessment of the BP refinery site, Kwinana, Western Australia:

Volume I Hydrological framework and site work recommendations, BHP

Research, Melbourne Laboratories, 1991.

• BP Refinery groundwater database.

• Davis, G.B.; C. D. Johnston, and B. M. Patterson, Deep drilling data, multiport

installation specifications, groundwater chemistry and soil gas composition for

the first bioremediation field trial at the Kwinana refinery site, Report to BP

(Kwinana), Report No. 94/34, September 1994.

• Rust PPK Environment and Infrastructure, Kwinana Works Groundwater

Monitoring Manual, prepared for Wesfarmers CSBP Ltd., May 1996,

98H065A:PR2:0639:RevA, 1996.

• PPK Environment and Infrastructure, Jervoise Bay Project Groundwater

Recovery Plan, report prepared for Department of Commerce and Trade, March

2000, 98L084A:PR2:5426, 2000.

• CSIRO Land and Water – Garden Island monitoring bores.

Three hundred and fifty bores were registered in the SWRIS/WIN database within the

search window 374000 to 387000 mE and 6425000 to 6445000 mN (AMG zone 50).

One hundred of these were registered for WRC and the rest on behalf of other

agencies. Seven hundred and sixty-four bores were recorded in AQWABase for the

same search window. All tabular information was collated and organized within a

spreadsheet, then imported into ArcView for spatial analysis and display.



Shallow groundwater bores in Figure 9 are those bores that do not fully penetrate the

Superficial Formations. These are denoted as “logged” if a record of stratigraphy

was made during drilling, otherwise they are denoted as “unlogged”. “Fully

penetrating” bores are those bores drilled fully through the Superficial Formations

and intersecting Cretaceous sediment. Bores that partially penetrate the Superficial

Formations are referred to as “shallow bores”. Thicknesses of stratigraphic units

were determined from drilling logs if this information was available.

The main purpose of reviewing drilling information was to refine the description of

regional stratigraphy (see Section 2.3) and develop a more detailed conceptual

hydrogeological model at a scale relevant to the investigation of SGD. Nevertheless,

it was found that the amount and quality of information available was insufficient to

significantly refine the regional model at local scales. Detailed groundwater

investigations have been carried out in some areas, for example, adjacent to Northern

Harbour and beneath the BP Refinery, but this level of information does not extend

across the study area. Only a relatively small number of the bores intersect

Cretaceous sediment (Figure 9 and Figure 10) and total thicknesses of the Superficial

Formations are known only at these locations.

Shallow bores range in depth from only a few metres to almost the base of the

Superficial Formations. Many of these are unlogged and do not provide stratigraphic

information. Interpretation and interpolation of stratigraphy based on the available

drilling information is made more difficult by the mixed quality of the bore logs.

Dependent on the drilling method purpose, these vary from detailed descriptions of

stratigraphy and sediment types to coarse descriptions of texture.

A summary of the stratigraphic information derived from the drilling logs is

presented in Figure 10 to Figure 13. These depict total thicknesses of Quaternary-

Tertiary sediments (Figure 10), Tamala Limestone (Figure 11), basal clay (Figure 12)

and Quaternary sands consisting of Safety Bay Sand and Becher Sand (Figure 13). A

greater than symbol (>) is used to indicate that a particular unit was intersected

during drilling but was not fully penetrated. For example, “>7” indicates that there is

at least 7 metres of the particular sediment type but the total thickness is unknown at

the drilled location. The symbol “A” is used to indicate that a unit is apparently

absent and “U” indicates no data (i.e. unknown).



��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

� ��

�

� � � ��

�

��

��

��

�

��

��

�� !��

�"��#�!��$��

�"��#�!��$��

��%��&��$'��&��$�(��'��

��$��(��'��$��#�� $ ��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 9: Groundwater bore locations



��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

�� !"�� #�$%�&

�� !"��'��"�(�"��)

* ��#��(��)&

+��* ��

��)�#+��* ��&

��)�#�� &

�,��-�)�-��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 10: Combined thickness of Quaternary and Tertiary sediments (m); greater than (>) symbols indicate partial thicknesses determined from shallow bore logs, total thicknesses are unknown at these locations



��

��

��

��

��

�

��

�

��

��

��

��

��

��

��

��

��

��

��

��

�

��

�

��

��

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

�� !��"#��$�%&�'

�� !��"#��(��#�)�#��*

+��$��)��*'

,��+��

��*�$,��+��'

��*�$�� '

�-��.�*�.��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 11: Thickness of Tamala Limestone (m); greater than (>) symbols indicate partial thicknesses determined from shallow bore logs, total thicknesses are unknown at these locations; “A” indicates the unit is absent; “U” indicates no data or unknown



��

��

��

��

��

��

�

��

�

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

�� !"�� #�$%�&

�� !"��'��"�(�"��)

* ��#��(��)&

+��* ��

��)�#+��* ��&

��)�#�� &

�,��-�)�-��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 12: Thickness of basal clay (m); greater than (>) symbols indicate partial thicknesses determined from shallow bore logs, total thicknesses are unknown at these locations; “A” indicates the unit is absent; “U” indicates no data or unknown



��

��

��

��

��

�

�

��

��

��

�

��

�

��

��

� ��

��

�

��

�

�

��

�

�

��

�

��

�

�

�

�

�

�

��

�

��

�

��

��

� �

� ��

�

�

�

��

�

�

�

�

��

��

��

��

��

��

��

��

��

��

��

��

�

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� �

�

��

�� !�� "��#$��%�&'�(

� !�� "��#$��)��$�*�$�� +

,�� %��*�� +(

-��,��

�� +�%-��,�� (

�� +�%��!��(

�.��/�+�/��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 13: Thickness of Quaternary sands (m); greater than (>) symbols indicate partial thicknesses determined from shallow bore logs, total thicknesses are unknown at these locations; “A” indicates the unit is absent; “U” indicates no data or unknown



2.6 GROUNDWATER LEVELS

2.6.1 REGIONAL GROUNDWATER TABLE

Figure 14 is a water table map for the Superficial Aquifer inshore from Cockburn

Sound and presents 1998 water level data compiled by Nield (1999). Water levels

are close to their minimum seasonal values, with maximum values being

approximately 0.4 metres higher. At locations of multiple depth bores, water levels

from the lower part of the Superficial Formations were used.

Extremely small hydraulic gradients occur along the coastal margin of the aquifer.

The 1-m water table elevation is generally a minimum of 2 km inshore from the

coast but is typically up to 4 km inshore. Hydraulic gradients within the coastal strip

of the Superficial Aquifer are in the order of .0002 (.02%) to .0005 (.05%), which

suggests that extremely high-permeability sediments occur in this area. Estimates of

hydraulic conductivity of Tamala Limestone vary up to values greater than

3,000 m.d-1 (see Table 1 and Sections 2.3.1 and 2.4.3).

2.6.2 INFLUENCE OF SEA LEVEL VARIATION

Because the coastal margin of the Superficial Formations contains highly permeable

sediments, shallow groundwater levels are strongly influenced by changes in sea

level at daily (tidal), seasonal and interannual time scales. In general, longer time-

scale sea level changes propagate further into the aquifer from the coast and affect a

larger part of the groundwater system. The ocean saltwedge—which in the study

area is more like a saltwater ‘tongue’—is known to extend up to 2 km inshore. An

extended rise or fall of sea level causes a corresponding rise or fall in both the

elevation of the saltwater interface and shallow groundwater levels.

Nield (PPK, 2000) compared monthly average sea levels to groundwater levels in the

Woodman Point monitoring bore WPM5, located approximately 200 m inshore from

the coast at Northern Harbour. The data depicts a strong correspondence between

sea level and groundwater level at monthly, annual and interannual time scales

during the twelve-year period from 1987 to 1999.

Walker (1994) investigated aquifer tidal propagation beneath the BP Refinery using

five pairs of monitoring bores located at distances of 10 m, 750 m and 1,280 m from

the coast. Diurnal tidal fluctuations in groundwater levels of approximately 2 cm

were measured at the site furthest from the coast. This is consistent with either high



permeability sediments or semi-confinement of Tamala Limestone by overlying

basal clay.

Nield (pers. comm.) found detectable tidal fluctuations in groundwater levels up to

1 km from the coast beneath Alcoa’s alumina refinery, which is located directly

south of Challenger Beach (see Figure 1).

More recently, Smith and Hick (2001) analysed aquifer tidal propagation at Northern

Harbour to obtain estimates of aquifer diffusivity—the ratio of transmissivity to

aquifer storage coefficient (see Appendix A and Table 1). The dominant diurnal

frequencies (K1 and O1 tidal constituents) were found to propagate more than 150 m

into the aquifer during each 24-hour tidal cycle. Since the basal clay is not present at

this location, these results indicate the presence of high-permeability coastal

sediments.

2.6.3 INFLUENCE OF HIGH-PERMEABILITY LIMESTONE

It is clear from the above information that the Tamala Limestone has a large tidal

efficiency due to high permeability associated with secondary porosity. Tidal

efficiency is further increased in areas where the limestone is semi-confined by

overlying clay. Efficient tidal propagation through the limestone effectively extends

the coastal sea-level boundary inshore and underneath the coastal sands. This effect

is clearly illustrated by water level data collected in several piezometers inshore from

the beach at Northern Harbour.

Figure 15 depicts groundwater levels recorded in two sets of nested piezometers that

were installed at Northern Harbour prior to an SGD intercomparison experiment (see

Section 3.1). Drilling and piezometer completion details are reproduced in

Section 3.1.1. The piezometer pairs were installed in an interdunal depression at

distances of approximately 180 m (NH1) and 95 m (NH2) from the shoreline. One

deep piezometer, approximately 17 m deep and slotted over the bottom 6 m, and one

shallow piezometer, approximately 5 metres deep and slotted over the bottom 3 m,

was installed at each location. The bottoms of the deep piezometers were located

just above the top of the limestone.

In late January 2001, a temporary trench was excavated adjacent to the piezometer

locations as part of works being undertaken for the Northern Harbour development.

Dewatering of the trench caused a draw down of shallow water table of

approximately 0.7 m near to piezometer nest NH2. Though the dewatering discharge



was initially directed away from the piezometer locations, it was subsequently re-

directed into the interdunal depression that contained the piezometers. This resulted

in local surface flooding and a ‘draw up’ of the shallow water table in this area.

A relevant observation is that, despite deviations in shallow water levels of

approximately 1.2 m at NH2b and 0.8 m at NH1b, the piezometric heads in both of

the deep piezometers were unaffected and continued to reflect sea level.

At this location, it is apparent that the basal clay is absent and the limestone behaves

like a prescribed head boundary condition at the base of the Quaternary sands. Sea

level fluctuations are propagated laterally though the limestone and then upward into

the overlying sediment. This data suggest that, where coastal sands overly high-

permeability limestone, groundwater movement through the sands is likely to be

predominantly vertical with minor lateral flow. In other words, groundwater in the

sand drains downward into the underlying limestone. This hypothesis is supported

by the fact that hydraulic gradients in the sands are much flatter than would be

expected if flow were predominantly horizontal since the hydraulic conductivity of

sand is significantly less than the conductivity of limestone.



��

� � � � ��

�

��

� ��

�

�

�

��

�

�

�

��

�

�

�

�

��

��

� �

��

�

�

�

��

�

�

�

�

�

��

�

�

�

�

�

�

�

�

�

�

�

��

�

�

�

�

��

�

�

��

�

�

�

��

��

��

��

�

��

�

�

�

��

�

�

��

��

�

�

�

�

�

�

��

��

��

� ��

�

�� !� "��

��

��

��

��

��

#

�$

�

�

%

�$��

��#

�%

��

#&' %&''&�

�&$

�&�

�&'

�&��&�

�&�

'&'

�&$

�&�

�&�'&%�&%

�&�

'&�

�&��&�

�&%

�&�

�&�

�&� �&��&�

�&��&$

�&�

�&�

�&�

�&%

�&�

�&��&'

�&�

%&�

�&��&�

�&'

�&�

�&$�&�

�&%�&�

�&�

�&$

�&%

�&%

�&$

�&$�&� �&�

�&��&%�&%�&%�&��&�

�&�

��&$��&'��&�

�%&�

��&� ��&#

��

�'&�

��

��&#

�$&'

��&%��&%

��&�

��&�

��&' ��&�

��&�

��&�

��&�

��&�

�#&�

�#&�

��&'��&� ��&�

��&��%&�

��&��&�

��&#�%&�

��&�

��&�

�� (�

��

�� )�&��*+��(��,

��, ��(!-�.��

�!��/��,�

��

��

��

��

��

��

��

��

��

��

��

��

��

�� !� �"�� 0� ��1��"�� '��*+��

23�3��

Figure 14: Water table map of the Superficial Aquifer inshore from Cockburn Sound based on 1998 water level data (source: Nield, 1999)



Figure 15: Tidal fluctuations in groundwater levels inshore from Northern Harbour from November 2000 to June 2001; sampling interval for water levels is 1 hour



2.7 SEAWATER INTRUSION

The saltwedge within the Superficial Formations adjacent to Cockburn Sound is

known to extend up to 2 km inshore. High permeability of Tamala Limestone results

in a long, flat tongue of saltwater rather than a classically shaped curved saltwedge.

This is commonly observed as layer of saline water at the base of the Superficial

Formations. Where the Tamala Limestone is confined above by basal clay that

extends to the coast, a second, much smaller saltwedge is present in the sand units

overlying the clay. Only a few bores in the Cockburn Sound study area intersect the

saltwedge and enable direct observation of saltwater intrusion. Unless a bore is

installed specifically for this purpose, it is normally undesirable to drill into or below

the saltwater-freshwater interface. As a result, most groundwater bores are screened

above the interface.

Knowing the position and shape of the saltwater-freshwater interface can be useful

for inferring information about SGD. For example, if groundwater discharges

through offshore geological structures such as limestone reefs—rather than directly

along the shoreline—then the position of the saltwedge and groundwater levels along

the shoreline should indicate this. Other aspects of the hydrogeology such as the

groundwater discharge velocity, aquifer permeability and shoreline geometry also

affect the position and shape of the saltwedge. Unfortunately, the potential for

conducting these types of analyses is hindered in this study due to the extremely flat

hydraulic gradients and the long, flat shape formed by intruded seawater.

Where the depth of the saltwater-freshwater interface adjacent to Cockburn Sound is

known, the interface position is reasonably well predicted by the Ghyben-Herzberg

approximation. This is a simple hydrostatic relation based on the density difference

between freshwater and seawater. It predicts that, in a freshwater aquifer, the depth

to an abrupt saltwater-freshwater interface will be approximately 40 times the height

of the groundwater table above mean sea level. For example, if the water table

elevation at a particular location is 0.5 m above mean sea level, the Ghyben-

Herzberg approximation predicts that the depth to the saltwater-freshwater interface

below mean sea level will be approximately (0.5 × 40 m) 20 m. If the aquifer is less

than 20-m thick at that location, then the freshwater head is sufficiently large to

prevent saltwater intrusion.

Adjacent to Cockburn Sound the base of the Superficial Formations is around -20 to

-25 m AHD. According to the Ghyben-Herzberg approximation, the inshore extent



of saltwater intrusion should correspond broadly to the 0.6 m AHD elevation contour

of the groundwater table. It is clear from Figure 14 that this can be a considerable

distance inshore. More generally, Table 2 can be used to estimate the location of the

saltwedge toe based on the saturated thickness of aquifer and water table elevation.

Tabulated values are derived from the Ghyben-Herzberg approximation, with values

in the second column being equal to values in the first column divided by 41.

Table 2: Relation between aquifer thickness and water table elevation at the position of the saltwedge toe based on the Ghyben-Herzberg approximation.

Aquifer Saturated Thickness

[m]

Water Table Elevation at

Saltwedge Toe [m AMSL]

15 .36 20 .48 25 .61 30 .73 35 .85 40 .98 45 1.1 50 1.2

Monitoring bores that are known to intercept the saltwedge in the study area are the

Mayor Road multiport bores (MR4M to MR6M) and the Cockburn Saltwater

Interface monitoring bores (CSI1/97, CSI2/97 and CSI3/97). The Mayor Road bores

are located directly east and south of Lake Coogee adjacent the northern end of

Cockburn Sound (Figure 17). The Cockburn Saltwater Interface bores are located

closer to the coast near Challenger Beach (Figure 36). Table 3 clearly shows that the

aquifer contains saltwater at the bottom ports of the Mayor Road monitoring bores,

which are located between 25.5 m and 27.5 m below ground level. The top of casing

elevations are less than 4 m AHD and therefore the saltwater interface in this area is

estimated to be around -21 to -24 m AHD.

It is worthwhile to note that Lake Coogee is an expression of the groundwater table

but is saline. Salinities in the lake have varied seasonally in the range 10,000 to

45,000 mg.l-1 of total dissolved solids (Hirschberg, 1990). The type of interaction

that occurs between the coastal saltwedge and saline water in Lake Coogee has not

been clearly established. Figure 16 depicts two possibilities. In case “a”, Lake

Coogee is convectively unstable due to the density contrast between saline lake water



and relatively fresher groundwater. This results in interaction between the lake

‘saltwedge’ and coastal saltwedge. The flow arrows indicate that the saltwedges are

in fact convection cells in which saltwater circulates from the sea (or lake) into the

aquifer and back out. In case “b”, the lake is convectively stable. A low-

permeability lakebed lining prevents natural convection beneath the lake. The

coastal saltwedge passes directly underneath and does not interact with it. In both

cases, Lake Coogee is a discharge lake with respect to the regional groundwater flow

and therefore accumulates solutes derived from aerosol fallout, which is high close to

the coast.

Table 3: Electrical conductivity of groundwater samples from saltwater monitoring bores

EC @ 25 deg. C [µS.cm-1]

Bore ID

Top of Casing

Elevation [mAHD]

Depth to Inlet [m] Min Max Mean Readings

MR4M 1.97 3.5 2,100 7,260 4,007 10 6.5 1,870 1,910 1,890 10 13.0 1,440 1,920 1,803 10 19.0 1,920 3,110 2,144 10 25.5 29,700 32,700 31,411 9 MR5M 3.78 5.0 669 1,210 889 9 10.0 728 1,730 1,411 11 15.0 1,340 1,840 1,448 10 21.0 1,850 17,000 3,919 10 26.0 17,000 20,600 19,600 10 MR6M 3.14 4.0 1,910 3,500 2,543 9 9.0 1,280 1,470 1,385 10 15.0 1,220 1,520 1,357 10 21.0 2,670 3,200 2,855 10 26.0 25,400 27,800 26,650 10 MR7M 3.57 5.0 919 1,630 1,377 10 10.5 1,700 2,030 1,953 10 16.0 2,570 2,900 2,831 10 21.5 4,340 5,310 4,822 10 27.5 30,000 32,100 31,070 10

EC - uncompensated [µS.cm-1]

Bore ID

Top of Casing

Elevation [mAHD]

Depth to Inlet [m] Min Max Mean Readings

CSI1/97 11.14 30.1-32.1 44,700 66,000 50,220 5 CSI2/97 11.2 18.9-20.9 7,730 10,030 8,732 5 CSI3/97 19.68 25.3-27.3 1,400 1,600 1,538 5

Source: WRC SWRIS/WIN database



Cockburn Saltwater Interface monitoring bores CSI1/97 and CSI2/97 are located

together approximately 150 m from the coast. Seawater intrusion is detectable at

about -9 m AHD in CSI2/97, which appears to sample a mixture of seawater and

terrestrial groundwater. At approximately -19 m AHD in CSI1/97 the aquifer

appears to contain seawater only. Monitoring bore CSI3/97 is located further

inshore, approximately 350 m from the coast. The aquifer is apparently unaffected

by saltwater intrusion at -6 m AHD at this location.

Gerbaz (1999) monitored groundwater levels and the saltwater interface position

beneath the BP Refinery in eleven monitoring bores. The saltwedge was intercepted

in nine bores and the average depth below water table to the saltwater interface was

between 17 m and 19 m; however, the distances between the coast and the

monitoring bores were not reported. The extent of saltwater intrusion beneath the

refinery was probably affected by groundwater pumping from the refinery production

bores and was observed to be greater in the middle of the refinery area than at the

northern and southern boundaries.

Passmore (1970) describes saltwater intrusion into the Rockingham Aquifer at the

southern end of Cockburn Sound.



MR

5M

MR

4M

NH2

NH1

Nor

ther

nH

arbou

rLa

keC

ooge

e

Tam

ala

Lim

est

one

Kar

din

yaSh

ale

Mem

ber

(Osb

orn

eFo

rmat

ion)

Ho

rizo

nta

lsc

ale

1:1

00

00

Vert

ical

scal

e1

:20

00

Vert

ical

exag

gera

tio

nx5

Sand

unit

s

NH3NH4NH5

Sand

unit

s

MR

5M

MR

4M

NH2

NH1

Nor

ther

nH

arbou

rLa

keC

ooge

e

Tam

ala

Lim

est

one

Kar

din

yaSh

ale

Mem

ber

(Osb

orn

eFo

rmat

ion)

Ho

rizo

nta

lsc

ale

1:1

00

00

Vert

ical

scal

e1

:20

00

Vert

ical

exag

gera

tio

nx5

Sand

unit

s

NH3NH4NH5

Sand

unit

s

(a)

(b)

Figure 16: Hydrological cross section (Figure 5, Section D) through the Northern Harbour study site depicting two possible types of interaction between the coastal saltwedge and saline water in Lake Coogee; (a) Convectively unstable lake, (b) Convectively stable lake



3 FIELD INVESTIGATIONS

Field-based activities were concentrated at three locations along the coast of

Cockburn Sound (Figure 1):

Northern Harbour

Challenger Beach

BP Refinery

In November and December 2000, Northern Harbour was the site of an international

scientific SGD intercomparison experiment (SCOR-LOICZ-IOC) that contributed

valuable information about groundwater-discharge processes that operate in

Cockburn Sound. In addition to the site being of interest because of known SGD

occurrence, this area was of particular interest because two groundwater-nutrient

plumes discharge into Northern Harbour (PPK 1999).

Challenger beach is located approximately mid way between Northern Harbour and

BP Refinery. There is limited industrial development inshore from this location and

Tamala Limestone is fully exposed along the shoreline. Groundwater discharge is

thought to be greater in this area because the limestone aquifer is direct contact with

the ocean.

BP Refinery is located along a stretch of foreshore where nutrient contamination of

groundwater is known to occur at high levels. A beach site adjacent to the Small

Boat Harbour at BP Refinery was selected for the in situ remediation trial that was

conducted as a counterpart to this investigation (see Section 1.2).

A major field component of this study was a nutrient survey of submarine porewater

conducted along Cockburn Sound foreshore between Woodman Point and Palm

Beach in Rockingham. The results of the survey are reported in Section 3.4 and

provided the basis of estimates of nutrient mass flux in Section 5.

3.1 NORTHERN HARBOUR

A map of the Northern Harbour and hydrogeological cross-section through the coast

are depicted in Figure 17 and Figure 16, respectively.



�

��

��

��

��

��

��

��

��

��

� ��

��

��

��

��!

"#��

"�"

��

��

��

��$

�%!&

'��

��

�"��

��()*�

�+�,-

�'��

�..�

$��

&/�

�0��

0��

��

�$ ��

��1&

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� �

��

��

��

��

� � � ��

� ��

� ��

� ��

� �

��

��

��

�2��

��

��

�2��

��

��

��

��

�2��

��

��

��

�

��

��

3�

3�

3� 3�

��

��

3�

3�

� 3�

��

��3�

3�

��

3� ��

��

3�

�

3�

��"�

4/�4/

�

%)�%

%)5%

%)2%

%)�%

��

��

�

4/��

4/��

4/�5

4/��

4/��

4/�3

6'%'�

� 6'%'� 6'%

'��

6'%'�

�

6'%'�

� 6'%'�

�

6'%'�

6'%'�

6'%'7

�6'�

'�

6'%'2

6'%'�

3

(.%

�'

�'��

6'%'�

5

�'��

�'8�'

7

�'�

�'�

�'��

�'2

�'� �'

5

(.%

�9

(.%

�9 (.%

3'

�� %

�4:��.

�*4�

�4��

��

!;.�

�)��

*�4�

/��!

��

6'�'�

��

��

��

��

��

��

� ��

��

��

��

� ��

��

��

��

��

��

��

� �

��

��

��

��

��

��

��

�

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 17: Northern Harbour (SGD intercomparison) study site



3.1.1 HYDROGEOLOGY

The bottom of the Superficial Formations in the Woodman Point area is intersected

at an elevation of around -30 mAHD (see Section 1). Thus, the saturated thickness

of the Superficial Aquifer is approximately 30 metres. The Kardinya Shale Member

of the Osborne Formation subcrops beneath the Superficial Formations and form a

low-permeability base. Regional water table gradients are small and indicate that the

aquifer has very high transmissivity in this area. Typical water table slopes are in the

order of 0.0003 (.03%).

Two piezometer nests (NH1 and NH2) were installed at Northern Harbour to

monitor inshore groundwater levels during the SGD intercomparison trial. Their

locations are indicated in Figure 17 and Figure 16. At each location, there is one

deep piezometer, approximately 17 metres deep and slotted over the bottom 6

metres, and one shallow piezometer, approximately 5 metres deep and slotted over

the bottom 3 metres. Bore completion details are presented in Table 4.

Table 4: Completion details of intercomparison monitoring bores

Bore ID

Latitude [Degrees, decimal minutes]

Longitude [Degrees, decimal minutes]

Depth [mBGL]

Screen Length

[m]

Ground Surface RL [mAHD]

Top of Pipe RL [mAHD]

NH1a (Deep) -32 08.356 115 45.863 16.44 6.0 1.414 1.721 NH1b (Shallow) -32 08.358 115 45.863 5.44 3.0 1.444 1.944 NH2a (Deep) -32 08.399 115 45.840 18.76 6.0 1.046 1.454 NH2b (Shallow) -32 08.397 115 45.840 5.30 3.0 1.009 1.412

Drilling and piezometer installations were carried out by CLW on 15th November

2000 using an auger drill and solid auger. Because the water table was close to the

surface and most drilling was through fully saturated and unconsolidated sediment,

this method did not allow accurate logging of the stratigraphy. It was possible to

determine the depth to limestone based on drilling resistance and drill ‘chatter’ but

the predominantly sandy sequence above the limestone could only be described

generally. Some interpretation of stratigraphy was possible during retrieval of the

auger sections by examining sediments adhered to the auger blades.

The first attempt to install the deep piezometer at location NH2 failed during drilling,

when the auger plug became stuck and the top of the hole ‘slumped’ due to the



drilling spoil presumably entering cavities in the limestone. A new deep hole

(NH2a) was drilled within 2 m of the abandoned hole.

Limestone was encountered at approximately 14 mBGL (elevation 13 mAHD) in the

abandoned hole at NH2 and at approximately 19 mBGL (elevation 18 mAHD) in

NH2a. The depth to limestone at NH1a is approximately 18 mBGL (elevation

17 mAHD).

In general, the sand sequence above the limestone became coarser and muddier with

increasing depth, varying from Safety Bay Sand at the surface to coarse and very

course, dark grey muddy sand above the limestone. The muddy sand contained

rounded sand grains up to 2 mm diameter, shell fragments and sparse plant fibre.

The bottom of NH2a contained yellow, coarse and very coarse muddy sand, with

rounded sand grains up to 3mm diameter.

The interpreted stratigraphic sequence at the site is Tamala Limestone, ?Cooloongup

Sand, Becher Sand and Safety Bay Sand.

3.1.2 GROUNDWATER LEVELS

Groundwater levels in all four piezometers were monitored using capacitive water

level probes (Dataflow Systems) and model 392 data recorders (Dataflow Systems).

Water levels were logged initially at 10-minute intervals during the period of the

SGD intercomparison experiment but the sampling frequency was subsequently

decreased to once per hour, which was adequate to detect and analyse tidal

fluctuations. The collected data is described and analysed in Section 2.6.3 and

Appendix A.

3.1.3 EC SURVEYS OF SUBMARINE POREWATER

The electrical conductivity (EC) of sediment porewater was measured and mapped in

Northern Harbour during the SGD intercomparison. (Stieglitz, 2001b). The

technique employed a conductivity probe (Stieglitz et al., 2000) that was pushed into

the sediment to obtain values of bulk-ground electrical conductivity, as a measure of

the combined electrical conductivity of the sediment and its porewater.

Measurements correspond to a spherical domain, which was approximately 4 cm in

diameter for the instrument used.



The EC probe is constructed as a Wenner array of four ring-electrodes, with the two

inner current electrodes generating an electrical field in the sediment and the two

outer voltage electrodes measuring the voltage potential. In use, both current and

voltage are measured with conventional multimeters and the ratio of current to

induced voltage is proportional to the bulk-ground EC.

The bulk-ground EC of sediment containing pure seawater is less than the EC of

seawater alone because non-conducting sediment grains affect in situ measurements

made with the EC probe. Thus, the sediment porosity also affects bulk-ground EC.

Nevertheless, the salinity of the porewater can be estimated—if the sediment

physical properties are assumed relatively uniform—by calibrating measurements

against bulk-ground EC for sediment that is known to contain pure seawater.

EC transect locations and results from the SGD intercomparison (Stieglitz, 2001b)

are depicted in Figure 18 to Figure 21. Survey times are given in Table 5. Note that

Figure 18 is a map of the small section of beach inside of Northern Harbour where

shore-based activities during the intercomparison were focused. The two inshore

stations on E-transect were groundwater-monitoring sites NH1 and NH2; see Figure

17. Note, also, that a bulk-ground EC value of around 15 to 16 mS.cm-1 indicates

that the sediment contained mostly seawater. The EC of seawater is approximately

55 mS.cm-1.

Table 5: Bulk-ground EC surveys by Steiglitz (2001b); measurement times and ocean water levels

R-Transect

Date Time [WST]

Water level [mAHD]

Origin of depth [mAHD]

04/12/00 1330 0.6 0.8 04/12/00 1700 0.8 0.8 05/12/00 1730 0.8 0.8

B-Transect

04/12/00 1330 0.6 0.6

E-Transect

04/12/00 1330 0.6 0.6

Figure 19 and Figure 20 depict bulk-ground EC readings along R-transect and E-

transect, which are both perpendicular to the shore. Both transects extend

approximately 8 m offshore and to a sediment depth of around 1 m below the seabed.



A vertical sampling interval of 0.1 m was used. The data indicate that, at the times

of these surveys, there were shoreline zones of relatively fresh SGD approximately

5-6 m wide. Figure 21 depicts bulk-ground EC along B-transect, which is parallel to

the shoreline. This profile indicates that SGD was higher at the southern end of the

beach, near to the location of R-transect.

Since the SGD Intercomparison experiment, additional porewater EC surveys in

Northern harbour have been conducted at the locations, dates and time indicated in

Table 6 and Figure 17. Results, which are depicted in Figure 22, show EC profiles

that were obtained at a uniform depth of approximately 0.5-m below the seabed.

Transect lengths were limited by seabed slope and depth of seawater since

measurements could only be taken in up to chest-deep water. An attempt was made

to extend transects further offshore by taking measurements from a boat but this

proved impracticable due to the difficulty of maintaining a stationary position over

measurement locations.

Table 6: Bulk-ground EC surveys by CLW; measurement locations and times

Northern Harbour

Location (MGA94) Name

Easting [m] Northing [m] Date

Time (WST)

Tide Stage

NHT1 383390 6443299 27/8/01 11:30 Low NHT2 383414 6443265 27/8/01 11:40 Low NHT3 383357 6443314 27/8/01 11:50 Low NHT4 383302 6443323 27/8/01 12:00 Low NHT5 383365 6443317 15/11/01 13:30 - NHT6 383417 6443282 15/11/01 14:00 -

Figure 22 indicates a zone of shoreline SGD approximately 5-m wide at transect

locations NHT5 and NHT6. There is also evidence of fresher porewater and SGD

further offshore at these locations. It is unlikely that all groundwater from the

superficial formations discharges to such a narrow zone along shoreline, which at

some locations appears to be only several metres wide. Seismic data collected

during the SGD intercomparison provides additional evidence that groundwater also

discharges into Northern Harbour further offshore (see Section 3.1.8).



3.1.4 GROUNDWATER CHEMISTRY AND ISOTOPIC COMPOSITION.

Figure 17 shows the locations of groundwater chemistry and isotopic compositions

from selected bores in the proximity of Northern Harbour.

Table 7 is a summary of the major ion chemistry data and stable isotope

compositions of the sampled groundwaters.

Table 7 shows and confirms the elevated nitrogen (nitrate and ammonium)

concentrations of up to 130 and 21 mg.L-1, respectively, that have previously been

reported at Cockburn Sound. Enriched stable isotope compositions and its

correlation with chloride show the influence of mixing of seawater intrusion with the

shallow groundwater discharging toward the coast.

3.1.5 POREWATER CHEMISTRY

Blue bars in Figure 22 indicate locations along porewater EC transects where

porewater samples were collected. The samples were obtained from approximately

0.5 m below the seabed using a spear probe and all samples were analysed for major

ions, nutrients, EC and pH. Analyses were performed by the Chemistry Centre of

Western Australia and results are reported in Table 8.

These results form part of a more extensive survey of porewater nutrients along the

Cockburn Sound foreshore and are discussed in more detail in Section 3.4.

Table 7: Groundwater chemistry and Isotopes

Bore Date Time Water Level Field Field Field Field Field Field Field Field Anion DeuteriumSampled Depth to TOC EC Salinity TDS Temp. pH Eh DO DO Alkalinity CO3 Ca Cl EC HCO3 K Mg N_NH3 N_NO3 Na P_SR SO4_S Si Balance pH sample/smow

m µS/cm mg/L C mV % mg/L mg/L mg/L mg/L mg/L mS/m mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L % (delta)JBMB10 30/11/2000 15:14 12.1 1,405 0.5 553 21.9 7.1 -7 2 1 300 <2 112 200 137 366 16 18 10 2.8 144 <0.01 77 4.2 1.3 7.6 -10.4JBMB13 30/11/2000 15:40 10.2 1,664 0.7 654 21.7 7.1 -9 2 0.2 300 <2 131 260 161 366 17 31 3.3 8.1 175 <0.01 108 3.8 3 7.8 -14.1JBMB3 30/11/2000 16:16 4.9 795 0.1 311 21.1 7.3 -18 67 5.9 213 <2 94 86 77 259 5 14 <0.01 5 56 0.01 37 2.7 3.4 7.9 -16.7JBMB16 30/11/2000 16:37 12.6 3,060 1.4 1212 22.1 6.8 7 2 0.2 425 <2 163 640 296 519 35 48 23 1.4 391 <0.01 155 3.6 -0.1 7.7 -14.5JBMB14 1/12/2000 - - 1,659 0.7 648 21.2 7.2 -11 53 4.7 250 <2 127 300 160 305 11 24 <0.01 8.4 183 <0.01 90 3 1.7 7.9 -13.6JBMB4 1/12/2000 11:03 4.3 2,790 1.3 1096 21.7 7.0 -2 2 0.3 300 <2 161 590 268 366 24 43 7.3 11 355 0.01 148 3.2 1.8 7.6 -14.1JBMB7 1/12/2000 10:20 5.3 4,880 2.5 1928 22.3 6.8 7 2 0.2 425 <2 198 1,200 472 519 47 77 21 0.03 693 <0.01 234 2.8 0.3 7.6 -12.2

JBMB9D 1/12/2000 10:43 4.1 50,100 32.6 - 24.1 7.3 -17 5 0.4 263 <2 426 18,000 5,200 320 377 1,210 11 <0.01 9740 0.02 2,480 2.1 -0.8 7.4 4.1NH1A 30/11/2000 - 1.3 2,680 1.2 1052 21.1 7.5 -30 2 0.2 245 <2 96 560 282 299 27 62 <0.01 5.2 371 <0.01 204 3.5 2.8 7.9 -13.0NH1B 30/11/2000 - 1.5 872 0.2 338 20 7.7 -42 1 0.1 263 <2 44 79 86 320 8 45 0.33 <0.01 77 0.01 50 3.5 4.9 8.0 -17.7NH2A 29/11/2000 - 1.0 3,530 1.7 1380 21.1 7.4 -23 5 0.4 238 <2 127 840 339 290 26 72 <0.01 6.8 515 <0.01 218 3.9 2.7 7.8 -13.3NH2B 29/11/2000 16:50 1.0 2,180 0.9 843 19.5 7.9 -52 3 0.2 263 <2 47 390 205 320 20 84 <0.01 <0.01 286 0.01 194 2.5 4.5 7.9 -12.3NH3 30/11/2000 - - 1,048 0.3 403 19.3 7.8 -45 4 0.3 263 <2 54 140 101 320 7 41 <0.01 <0.01 107 <0.01 43 3.3 3.6 8.0 -8.5NH4 30/11/2000 - - 1,377 0.5 541 21.7 7.8 -44 2 0.2 313 <2 60 210 134 381 9 47 <0.01 <0.01 171 0.01 86 3.8 1.6 8.0 -8.9NH5 30/11/2000 - - 1,657 0.7 652 21.3 7.7 -40 3 0.2 313 <2 75 280 161 381 11 50 <0.01 0.1 207 0.01 115 4.2 1.6 8.0 -8.2

Test (A.L.) 1/12/2000 11:12 15.2 4,690 2.5 1900 26.8 6.5 27 2 0.2 1390 <2 338 710 490 1,690 80 63 130 0.51 441 3.4 33 4.4 -5.6 7.3 -10.6WPM5C 30/11/2000 14:45 11.0 1,767 0.7 694 21.7 7.0 -5 6 0.6 288 <2 153 290 173 351 7 34 0.02 21 176 0.01 49 4.3 5 7.5 -14.2

CC1 (200m) 1/12/2000 - - - - - - - - - - 138 <2 23 720 256 168 20 36 0.16 0.01 473 0.01 106 10 -0.4 7.6 -23.8CC2 (400m) 1/12/2000 - - - - - - - - - - 188 <2 18 670 250 229 16 14 <0.01 0.09 517 0.01 58 7.4 2 7.8 -23.5

NH6 6/12/2000 10:44 1.3 52,100 33.5 22.1 7.9 -53 2 0.2 - - - - - - - - - - - - - - - - -NH7 6/12/2000 10:43 1.0 52,100 33.4 22.2 7.9 -53 2 0.2 - - - - - - - - - - - - - - - - -

Table 8: Northern Harbour porewater chemistry

CCWA ID Name Date CO3 Ca Cl EC @ 25 °C

[mS.m-1] HCO3 K Mg N_NH3 N_NO2 N_NO3 N_total Na P_SR P_total S_SO4

Ionic Balance

[%] pH

01E0594/012 NHT5-3m 15-Nov-01 <2 163 5600 1680 262 123 399 0.52 <0.02 0.01 0.59 3070 0.03 0.03 618 1 8 01E0594/013 NHT6-1m 15-Nov-01 <2 178 1800 665 281 47 133 0.11 <0.02 <0.01 0.19 1030 0.01 0.02 361 2.3 7.4 01E0594/014 NHT6-5m 15-Nov-01 <2 174 1600 562 281 44 125 0.19 <0.02 <0.01 0.26 972 0.01 0.08 352 4.3 7.4

Notes: 1. All concentration are mg.l-1 2. N_NH3 – nitrogen ammonia fraction by FIA (filtered); N_NO2 – nitrogen nitrite fraction by FIA (filtered); N_NO3 – nitrogen nitrate fraction by FIA (filtered); N_total

– total nitrogen, persulphate by FIA (unfiltered); P_SR – phosphorus soluble reactive by FIA (unfiltered); P_total – total phosphorus, persulphate by FIA (infiltered) 3. FIA – Flow Injector Analysis



Reproduced from: Steiglitz (2001)

NH1

NH2

Figure 18: Bulk-ground EC measurements by Steiglitz (reproduced from: Steiglitz, 2001b); transect locations in Northern Harbour during the SCOR SGD intercomparison experiment



R-Transect.

0 1 2 3 4 5 6 7 8

-1.5

-1.0

-0.5

0.0

Dep

th (

m)

0m AHD

beach surface

samplingpoints

0 1 2 3 4 5 6 7 8

-1.5

-1.0

-0.5

0.0

Dep

th (

m)

water level

04/12/00 ca. 1330 WST

0 1 2 3 4 5 6 7 8

-1.5

-1.0

-0.5

0.0

Dep

th (

m)

04/12/00 ca. 1700 WST

water level

0 1 2 3 4 5 6 7 8

Distance along transect (m)

-1.5

-1.0

-0.5

0.0

Dep

th (

m)

05/12/00 ca. 1730 WST

approximate position ofNY Mafia's seepage meter

water level

0.0 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 13.5 15.0

Bulk Ground Conductivity (mS/cm)

Figure 19: R-Transect; bulk-ground EC measurements (reproduced from: Steiglitz, 2001b)



E-Transect.


0.0 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 13.5 15.0

E1(seepage

meter)

0 1 2 3 4 5 6 7 8 9 10

-1.5

-1.0

-0.5

0.0

Dep

th (

m) beach surface

samplingpoints

0m AHD

0 1 2 3 4 5 6 7 8 9 10

Distance along transect (m)

-1.5

-1.0

-0.5

0.0

Dep

th (

m)

06/12/00 ca. 1045 WST

water level

Figure 20: E-Transect; bulk-ground EC measurements (reproduced from: Steiglitz, 2001b)

B-Transect.


0.0 1.5 3.0 4.5 6.0 7.5 9.0 10.5 12.0 13.5 15.0

00 10 20 30 40 50 60 70 80 90

-1.0

-0.5

0.0

Dep

th (

m)

0m AHD

beach surface (waterline)

samplingpoints

R-Transect East Transect West Transect

00 10 20 30 40 50 60 70 80 90

Distance along beach (m)

-1.0

-0.5

0.0

Dep

th (

m)

04/12/00 ca. 1730 WST

R-Transect East Transect West Transect

Figure 21: B-Transect; bulk-ground EC measurements (reproduced from: Steiglitz, 2001b)



NHT2 - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 Distance from Shoreline [m]

Bulk-ground Conductivity

[mS/cm]

NHT1 - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 Distance from Shoreline [m]


[mS/cm]

NHT3 - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 Distance from Shoreline [m]


[mS/cm]

NHT4 - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 Distance from Shoreline [m]


[mS/cm]

NHT5 (middle of beach) - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Distance from Shoreline [m]


[mS/cm]

NHT6 (east end of beach) - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Distance from Shoreline [m]


[mS/cm]

Figure 22: Bulk-ground EC measurements; additional Northern Harbour transects; blue bars indicate locations at which porewater samples were collected



3.1.6 BENTHIC FLUX MEASUREMENTS

Several different designs of benthic flux meter (BFM)—or seepage meter—were

deployed in Northern Harbour during the SGD intercomparison and on subsequent

occasions. Basic knowledge of the instrument designs is required to make

meaningful interpretations of results. Therefore, measurement principles and

operation procedures for each type of meter are described briefly in Appendix B.

Benthic flux measurements from Northern Harbour are reported below.

Results from SGD Intercomparison

Instrument locations in Northern Harbour during the SGD intercomparison are

depicted in Figure 23. Benthic flux measurements were made over a ten-day period

from 27th November to 6th December, during which the daily mean sea level fell by

around 0.4 m. This was superimposed by diurnal tidal fluctuations that varied in

amplitude from approximately 0.1 m to 0.35 m, with the amplitude of tidal variation

generally decreasing during the measurement period. Further detail about local tides

is contained in Appendix A.

It is now apparent that most of the measurement stations were located beyond the

narrow shoreline zone of SGD that was identified subsequently during profiling of

porewater EC (see Section 3.1.3). While this did not invalidate the experimental

design of the intercomparison—which aimed to compare the performance of

different meters—it did mean that the measured fluxes were probably seawater

exchanges between the sea and aquifer. This notion was supported by the fact that

discharging water collected by Lee-type meters (Appendix B) was saltwater with

seawater EC. This also highlights the advantage of carrying out a beach-probe EC

survey to provide guidance on the deployment of seepage meters (Section 3.1.3).

A summary of benthic flux measurements from the SGD intercomparison is

presented in Table 9. Operation of Lee-type meters was labor intensive and

measurements could only be carried out during the day. The ultrasonic BFM was

deployed continuously at station E2, and for a short period at E1. Time-averaged

values of benthic flux were automatically calculated and recorded every 15 minutes.

This meant that variation in benthic flux at time scales smaller than 15 minutes was

not investigated.

The best measurement sequences were obtained at stations E1 and E2 on East

Transect (Figure 24 and Figure 25); both were located relatively close to shore but



beyond the shoreline zone of SGD. Each plot depicts sea level in Northern Harbour

(denoted as “Tide”) and benthic flux measurements obtained using Lee-type (Lee,

1977) and ultrasonic (Paulsen et al., 1997) meters. Lee-type benthic flux

measurements varied between -3 cm.d-1 and 33 cm.d-1 at station E1, and generally

increased in magnitude during the monitoring period. This increase was apparently

in response to falling sea level and the associated increase in the local hydraulic

gradient toward the sea. Negative values of benthic flux during 27th November

indicate that sea water was ‘recharging’ the aquifer when the sea level was at its

maximum level during the measurement sequence. There is also evidence that

benthic flux responded dynamically to the diurnal tide, decreasing on a rising sea

level and increasing on a falling sea level. There was less variation in benthic flux

measured by the Lee-type meter at station E2 (5 to 14 cm.d-1) though the trend of

increasing benthic flux in response to falling sea level was detectible.

Table 9: Benthic flux measurements; SCOR SGD intercomparison

Meter Type Range of recorded fluxes

[cm.d-1] Averaging Period

Lee-type (drum & bag) -3 to 32 1-3 hours Krupaseep (heat pulse) 1 to 3 - CLW heat-pulse BFM -60 to 60 0.2-10 seconds Ultrasonic -12 to 61 15 minutes

In general, it is apparent from the results that benthic flux is variable and strongly

influenced by sea level. There is a quasi-inverse relation between sea level and

measured benthic flux, which varied from -12 cm.d-1 to 27 cm.d-1 at station E1 and

-12 cm.d-1 to 61 cm.d-1 at station E2.

Heat-pulse BFM Measurements

Subsequent to the SGD intercomparison, benthic flux measurements were made in

Northern Harbour using a heat-pulse BFM developed by CSIRO Land and Water

(Appendix B). On these occasions, care was taken to deploy the instrument within

the narrow shoreline zone of SGD. The results indicated that water fluxes across the

sediment-water interface were highly variable in response to sea level fluctuations at

time scales ranging from seconds to days. In addition, the tide-driven and wave-

driven fluxes were a similar order of magnitude to the predicted magnitude of

terrestrial SGD. Thus, the task of separating SGD measurements into terrestrial-

groundwater and marine-water components is difficult. This would require very long



time series of benthic flux measurements that span the broad frequency spectrum of

wave and tide phenomena.

Figure 26 and Figure 27 are examples of the complex heat pulse data obtained under

field conditions in Northern Harbour, in comparison with the results obtained under

laboratory conditions (Appendix B). Water depths, which were detected by a

pressure transducer on the BFM, are also plotted. They reveal subtle details about

water level variation during the 66-second measurement period. Figure 26 indicates

that, after the heater fired at 20 seconds, the water level at the measurement location

remained relatively constant for the next 30-40 seconds. Small-scale fluctuations in

water level at time scales of 1-2 seconds during this period caused only minor

perturbations in benthic flux. At the time of measurement, the direction of benthic

flux was into the sediment. Figure 27 is more typical of the data obtained from

Northern Harbour and reveals the influence on benthic flux of ocean swells, which

have typical periods of around 10 to 15 seconds in Cockburn Sound. At the time the

heater fired, the direction of benthic flux was out of the sediment at a rate of around

10-20 cm.d-1 (0.001-0.002 mm.s-1). Subsequently, several reversals in flow direction

occurred that resulted in multiple temperature peaks at each of the thermistor

locations. Determining the arrival time of the temperature front at thermistors

immediately after the heater fired was the preferred method for computing the

direction and magnitude of benthic flux.

Time-series of benthic flux measurements collected in Northern Harbour at two-

minute intervals (Figure 28) indicate that the direction of benthic flux can be into or

out of the sediment dependent on the water level above the instrument location at the

time of measurement. The effect of ocean swells on benthic flux appears to be

significant even under relatively calm conditions, such as those inside of Northern

Harbour. Benthic flux rates of between –60 and 60 cm.d-1 were measured.

Analytical models of tide and wave affects on benthic flux, which are presented in

Appendix C, also corroborate these results.



Figure 23: Site map of shore-based experimental design during the SGD intercomparison; (reproduced from: Krupa et al., 2000)



E1 - East Transect

-20.0

-10.0

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

26/11/00 27/11/00 28/11/00 29/11/00 30/11/00 01/12/00 02/12/00 03/12/00 04/12/00 05/12/00 06/12/00 07/12/00

Time

Ben

thic

Flu

x (c

m/d

ay)

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Tid

e (m

AH

D)

Ultrasonic Lee-type Tide

Figure 24: Benthic flux measurements recorded during the SGD intercomparison; East Transect, station E1

E2 - East Transect

-20.0

-10.0

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

26/11/00 27/11/00 28/11/00 29/11/00 30/11/00 01/12/00 02/12/00 03/12/00 04/12/00 05/12/00 06/12/00 07/12/00

Time

Ben

thic

Flu

x (c

m/d

ay)

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

Tid

e (m

AH

D)

Lee-type Ultrasonic Tide

Figure 25: Benthic flux measurements recorded during the SGD intercomparison; East Transect, station E2



BFM Data - Northern Harbour8/5/01, 11:46 AM

2100

2120

2140

2160

2180

2200

2220

2240

0 10 20 30 40 50 60 70

Time [s]

Tem

p [

cou

nts

]

0

50

100

150

200

250

300

Wat

er D

epth

[co

un

ts]

T1 T2 T3 T4 T5 T6 Depth

Figure 26: Example heat pulse data from Northern Harbour indicating seawater flow into sediment; benthic flux from T2-T3 pair ≈ - 36 cm.d-1, benthic flux from T1-T3 pair ≈ - 21 cm.d-1

BFM Data - Northern Harbour8/5/01, 11:52 AM

2110

2120

2130

2140

2150

2160

2170

2180

2190

2200

0 10 20 30 40 50 60 70

Time [s]

Tem

p [

cou

nts

]

0

50

100

150

200

250

300

Wat

er D

epth

[co

un

ts]

T1 T2 T3 T4 T5 T6 Depth

Figure 27: Example heat pulse data from Northern Harbour indicating SGD; benthic flux from T4-T5 pair ≈ 20 cm.d-1, benthic flux from T4-T6 pair ≈ 9 cm.d-1



Figure 28: Benthic flux measurements from Northern Harbour on 8/5/01 from 11:24 AM to 5:16 PM; measurements are every two minutes



3.1.7 CONTINUOUS RADON MONITORING IN NORTHERN HARBOUR

During the SGD intercomparison, researchers from the Department of Oceanography

at Florida State University conducted continuous radon monitoring of the water

column off a small pontoon anchored approximately 25 m offshore in Northern

Harbour. They observed a diurnal signal in radon concentration that was highest at

low tide indicating higher groundwater discharge during that period. Based on

modelling of the inventory of radon in the water column they calculated an average

discharge of fresh groundwater of 10 ± 7 cm.d-1 in the harbour. This result was

reported at the SCOR/LOICZ Working Group 112 meeting held in Catania (Sicily)

following the intercomparison. At the time of writing this report a detailed report on

the continuous radon monitoring work was not available.

Radon (Rn222) is a soluble radionuclide gas that is naturally occurring in

groundwater. It is formed from the decay of Radium, which comes from the decay

of Uranium. Groundwater becomes enriched with radon through contact with

sediments that contain traces of uranium and radium. Because it has a relative short

half-life (3.82 days), radon is ‘lost’ relatively quickly following groundwater

discharge to the surface, where the decayed radon production is not supported

through contact with minerals comprising the aquifer matrix. Because the decay rate

of radon is well known, and if the radon inventory in the water column is monitored

and its concentration in groundwater is known, it is possible to estimate the rate of

groundwater discharge into the water column. Importantly, this method provides an

independent estimate of SGD that does not involve aquifer properties or any aspects

of the inshore groundwater balance.

3.1.8 SEISMIC SURVEY

A seismic survey in a small area of Northern Harbour and one seismic line outside of

the harbour in Cockburn Sound were conducted during the SGD intercomparison

(Figure 29). Thomas Stieglitz (Stieglitz, 2001a) from the Marine Geophysical

Laboratory of James Cook University conducted this work using a 3.5 KHz ORE

Systems seismic profiler mounted of the boat Tartan II.

Note that this technique cannot be used to calculate seepage rates but has been used

successfully at other locations in Australia to map the occurrence of focused SGD

(e.g. Wonky Holes off the coast of Queensland).



The survey in Cockburn Sound revealed a number of stationary ‘features’ that were

consistent with fluid of different density to seawater migrating from the seabed.

These features are indicated on the sections of seismic chart depicted in Figure 30

and Figure 31. Numbered features might be indicative fresh groundwater discharge

from solution features and fractures in the Tamala Limestone; however, it was

beyond the scope of this survey to determine whether the observed density anomalies

were due to fresh groundwater discharge, gas migration from seabed sediment or

possibly some other reason. Most features were recorded more than once (e.g.

features 1, 2, 3, 4 and 6) indicating they might be permanent. Because limited data

was gathered, the number and positions of recorded features may not be

representative of the seabed in Northern Harbour and Cockburn Sound.

Generally, this type of seismic feature is recorded where there is a density difference

that affects the reflection of sound; e.g. large schools of fish, a cloud of suspended

sediment or fluid with a different density compared to the surrounding seawater.

Usually, recordings of fish shoals on seismic charts are not as straight and symmetric

as the seismic features recorded in Northern Harbour and Cockburn Sound. In

addition, fish shoals are normally mobile in shallow waters, which prevents repeated

recordings.



115.755 115.760 115.765

-32.155

-32.150

-32.145

-32.140

NORTHERN HARBOUR Seismic track 01/12/00

La

titu

de

So

uth

Longitude East

115.755 115.760 115.765

-32.155

-32.150

-32.145

-32.140


La

titu

de

So

uth

Longitude East

115.755 115.760 115.765

-32.155

-32.150

-32.145

-32.140


La

titu

de

So

uth

Longitude East

115.755 115.760 115.765

-32.155

-32.150

-32.145

-32.140


L

atit

ud

e S

ou

th

Longitude East

Figure 29: Seismic tracks in Northern Harbour and Cockburn Sound (reproduced from: Steiglitz, 2001b)



Figure 30: Northern Harbour seismic features (reproduced from: Stieglitz, 2001a)



Figure 31: Cockburn Sound seismic features (reproduced from: Stieglitz, 2001a)



3.1.9 GROUNDWATER TEMPERATURE AND EC PROFILING

Temperature and EC profiling of selected groundwater monitoring bores in the

Northern Harbour area was carried out on two occasions as part of the SGD

intercomparison. The results of this work, which are reproduced in Figure 32 to

Figure 34, provide useful information about the position of the saltwater-freshwater

interface in the area. Note that profiles were obtained using a 0.5-m vertical

sampling interval.

EC values are indicated over screened sections of bores only but the temperature

profiles are indicated over the full depths of bores, based on the assumption that

water inside a bore casing is approximately in thermal equilibrium with groundwater

directly outside the bore. This is supported by the fact that abrupt temperature

changes between screened and unscreened section of bores were not observed.

The data indicate that groundwater EC is around 2 mS.cm-1 near the water table and

begins to increase with depth from around -10 mAHD. However, the saltwater-

freshwater interface was not detected in the measured bores. Groundwater EC was

generally less than 6 mS.cm-1 in all bores at screened sections. One profile indicated

an abrupt increase in EC inside bore JBMB7 on 28/11/00; at approximately

-14 mAHD, EC increased from less than 5 mS.cm-1 to greater than 13 mS.cm-1. The

accuracy of this data is questionable because the profile measured in nearby

monitoring bore JBMB9D does not indicate the same increase in salinity despite

being several metres deeper than JBMB7.

The general conclusion is that the saltwater-freshwater interface is below -18 mAHD

in the Northern Harbour area. Relatively small increases in salinity start from a

depth of around -10 mAHD and are apparently associated with relatively minor

mixing and dispersion above the saltwater-freshwater interface.



Northern Harbour Groundwater Bores(Date: 29/2/00)

-25

-20

-15

-10

-5

0

5

16 18 20 22 24 26 28 30

Water Temperature (ºC)

Ele

vati

on

(m

AH

D)

JBMB1

JBMB2

JBMB3

JBMB4

JBMB5

JBMB7

JBMB8

JBMB9D

JBMB12

JBMB13

JBMB15

OB10

OB11

OB12

JBTB1

JBTB2

Figure 32: Groundwater temperature profiles on 29/2/00 in selected Northern Harbour monitoring bores; profiles are indicated over the full depths of bores




-25

-20

-15

-10

-5

0

5

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000

EC (µS/cm)

Ele

vati

on

(m

AH

D)

JBMB1

JBMB2

JBMB3

JBMB4

JBMB5

JBMB7

JBMB8

JBMB9D

JBMB12

JBMB13

JBMB15

OB10

OB11

OB12

JBTB1

JBTB2

Figure 33: Groundwater EC profiles on 29/2/00 in selected Northern Harbour monitoring bores; profiles are indicated over the screened sections of bores only




-25

-20

-15

-10

-5

0

5

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Water Temperature (ºC)

Ele

vati

on

(m

AH

D)

JBMB3

JBMB4

JBMB7

JBMB9D

JBMB10

JBMB13

JBMB14

JBMB16

WPM5C

NH1A

NH2A

NH3

NH4

NH5

Figure 34: Groundwater temperature profiles on 28/11/00 in selected Northern Harbour monitoring bores; profiles are indicated over the full depths of bores




-25

-20

-15

-10

-5

0

5

0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000

EC (µS/cm)

Ele

vati

on

(m

AH

D)

JBMB3

JBMB4

JBMB7

JBMB9D

JBMB10

JBMB13

JBMB14

JBMB16

WPM5C

NH1A

NH2A

NH3

NH4

NH5

Figure 35: Groundwater EC profiles on 28/11/00 in selected Northern Harbour monitoring bores; profiles are indicated over the screened sections of bores only



3.2 CHALLENGER BEACH

Challenger Beach (Figure 36) is located at the northern tip of the Cape Peron Safety

Bay Sand (Figure 5). At its northern end, the beach pinches out and Tamala

Limestone is exposed as a small coastal cliff that runs northward toward Woodman

Point. This strip of coastline is thought to have relatively high SGD because high-

permeability limestone is in direct contact with the sea. It is also in the middle of the

embayment between Woodman Point and James Point where SGD is probably

focused. In general, coastal embayments focus SGD toward their centers because the

coastline extends further into the coastal aquifer at this location and groundwater

generally flows toward the nearest point on the coast.

3.2.1 HYDROGEOLOGY

The base of the Superficial Formations in the area of Challenger Beach is at an

elevation of around -25 mAHD. Saturated thickness of the Superficial Aquifer is

approximately 25 m and regional water table gradients are around 0.0003 (.03%)

toward the coast. There is evidence of basal clay at the base of the coastal sand units

(Figure 12) but there is insufficient data to interpolate the spatial distribution or

thickness of clay. The Kardinya Shale Member of the Osborne Formation subcrops

beneath the Superficial Formations.



��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� ��

�

��

��

�

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

!"�

�#��

��

$��%

&'(��

&)�&��#�"��

#��

��(��

*��+

%,��

��#��

��(�&

��-

.��!

/0!�

��

�1��

�1��1�

2(#�)

��

�3�##��

��&��

�(&��4

��

��

��##�

5�4�

��*#�

��6,

��

��##�

5�4�

��*�&

#��6,

��

7��5

��%

��

&��

� �

��

��

(��&��

��*��

),

��

��

��

��

��

� �

��

� �

��

��

��

��

��

��

��

��

��

��

��

��

��

� �

��

� �

��

� �

��

� �

��

Figure 36: Challenger Beach study site



3.2.2 EC SURVEYS OF SUBMARINE POREWATER

Locations where porewater EC surveys were conducted are indicated in Figure 36

and the dates and times of surveys are listed in The variable distribution of porewater

salinity and occurrence of relatively fresher porewater at significant distances

offshore, suggest that a considerable portion groundwater discharge may take place

via preferred pathways associated with solution features and fractures in the

limestone.

Table 10. Results are depicted in Figure 37. See Section 3.1.3 for additional

information about this technique.

Based on transect CBT7 it is evident that fresh groundwater discharges a significant

distance offshore relative to aquifer thickness. Relatively fresh porewater was

observed at 33-m from the shoreline and it seems probable that groundwater may

have been discharging further offshore. Transects could only be measured in up to

1.5-m deep water, which limited the length of transect at each location. Note that

distinct shoreline zones of SGD did not occur at each transect—for example,

compared transects CBT3 and CBT6.

The variable distribution of porewater salinity and occurrence of relatively fresher

porewater at significant distances offshore, suggest that a considerable portion

groundwater discharge may take place via preferred pathways associated with

solution features and fractures in the limestone.

Table 10: Bulk-ground EC surveys; measurement locations and times

Challenger Beach

Location (MGA94) Name

Easting [m] Northing [m] Date Time (WST) Tide Stage

CBT1 384395 6438453 27/8/01 13:00 Rising CBT3 384357 6438730 27/8/01 13:20 Rising CBT4 384385 6438578 27/8/01 13:30 Rising CBT5 384389 6438410 27/8/01 13:40 Rising CBT6 384397 6438197 27/8/01 13:55 Rising

CBT7 (Boat Ramp) - - 15/11/01 08:30 Falling





porewater samples were collected. Samples were obtained from approximately


ions, nutrients, EC and pH. Results are reported in Table 11 and discussed in more

detail in Section 3.4.

Table 11: Challenger Beach porewater chemistry



Ionic Balance

[%] pH

01E0594/015 CBT7-8m 15-Nov-01 <2 151 5400 1540 262 130 360 0.04 <0.02 3.3 3.5 3240 0.03 0.07 805 2.5 7.9 01E0594/016 CBT7-19m 15-Nov-01 <2 144 7300 2240 268 176 496 0.03 0.29 1.7 1.8 4540 0.09 0.13 1080 3.7 8 01E0594/017 CBT7-25m 15-Nov-01 <2 61 4000 1320 293 96 246 0.03 0.38 2.2 2.3 2590 0.07 0.16 589 2.7 8.2 01E0594/018 CBT7-33m 15-Nov-01 <2 137 3800 1220 293 91 233 0.04 0.38 2.3 2.5 2430 0.11 0.15 564 3.6 8.2

Notes: 1. All concentration are mg.l-1 2. N_NH3 – nitrogen ammonia fraction by FIA (filtered); N_NO2 – nitrogen nitrite fraction by FIA (filtered); N_NO3 – nitrogen nitrate fraction by FIA (filtered); N_total –

total nitrogen, persulphate by FIA (unfiltered); P_SR – phosphorus soluble reactive by FIA (unfiltered); P_total – total phosphorus, persulphate by FIA (infiltered) 3. FIA – Flow Injector Analysis


CBT1 (boat ramp) – 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 Distance from Shoreline [m]


[mS/cm]

CBT3 - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 8 Distance from Shoreline [m]


[mS/cm]

CBT4 - 0.5m

0 2 4 6 8

10 12 14 16 18



[mS/cm]

CBT5 - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 Distance from Shoreline [m]


[mS/cm]

CBT6 - 0.5m

0 2 4 6 8

10 12 14 16 18



[mS/cm]

CBT7 (boat ramp) - 0.5m

0 2 4 6 8

10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Distance from Shoreline [m]


[mS/cm]

Figure 37: Bulk-ground EC measurements; Challenger Beach transects; blue bars indicate locations at which porewater samples were collected



3.3 BP REFINERY

Figure 38 shows the Small Boat Harbour at BP Refinery, located directly south of

James Point (Figure 1) near to the boundary between BP Refinery and CSBP. A

regional-scale hydrogeological cross section through this part of the coast is depicted

in Figure 7.

3.3.1 HYDROGEOLOGY

Numerous bores have been drilled within the BP Refinery and Bodard (1991a,

1999b) conducted a detailed review of the hydrogeology utilising this drilling

information. Following is a summary of Bodard’s analysis of lithological data from

approximately 45 boreholes:

• The Leederville Formation subcrops beneath the Superficial Formations and is

generally intersected at around 30-35 m below ground surface.

• The stratigraphic sequence of ?Rockingham Sand, Tamala Limestone, basal clay,

Becher Sand and Safety Bay Sand can be mapped across the refinery site above

the Leederville Formation.

• Rockingham Sand is typically 5-7 m thick and is first intersected between 23 m

and 25 m below ground surface.

• Tamala Limestone overlies Rockingham Sand. It varies in thickness generally

between 7 m and 11 m mainly due to relief on the top surface of the limestone.

There is a distinctive knoll in this relief that is around 10 m higher than the

surrounding limestone. The knoll is approximately 800 m wide at its base and

centered just inshore from the beach on the southwest tip of James Point.

Groundwater seepage has been observed on some occasions from a small outcrop

of the limestone knoll, located on the shoreline immediately west of the refinery.

• Basal clay is consistently found at the base of the Becher Sand. The clay typical

is 1-m thick and up to 2-m thick in places. It appears to infill the lower parts of

depressions in the upper surface of Tamala Limestone and pinches out around the

limestone knoll; it is absent where the top of the Tamala Limestone is above

approximately -8 m AHD. On the other hand, the offshore distribution of basal

clay is unknown.

• Becher Sand is typically 8 m thick and first intersected at around 5-6 m below

ground surface.


• Safety Bay Sand is typically 5 m thick.

Hydraulic head in the sand aquifer is up to 0.5 m higher than in the Tamala

Limestone due to the basal clay acting as an aquitard. This means that water levels

in the sand aquifer are relatively more influenced by vertical recharge. Water levels

greater than 1 m AHD and significantly above the regional aquifer water table

(Figure 14) are observed in the sand aquifer in the refinery area. This suggests that

shallow ‘mini-mounds’ develop above the basal clay in response to local

groundwater recharge and that shallow groundwater flow may reflect these gradients

rather than the regional hydraulic gradient.

Tidal propagation in the limestone is more efficient than in other areas due to the

clay layer acting as a confining ‘lid’. Tidal fluctuations are propagated by elastic

expansion and compression of the aquifer and aquifer diffusivity—the ratio of

transmissivity to aquifer storage coefficient—is increased (see Appendix A for

additional information).



�

��

��

��

��

��

��

��

��

��

� ��

��

��

��

��!

"#��

"�"

��

��

��

��$

�%!&

'��

��

�"��

��()*�

�+�,-

�'��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� ��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

� ��

��

��

'.��

�/��

'.�0 '.

��/

��'.

�)1�"

��

��'�

�2�

�3��

�

�

�

�

4��

�"

��"

��3��

��

��

5�3�

��$��

��6&

��

��

5�3�

��$�"

��

6&��

.��5

��!

7��

"��

� �

��

��

��

"��

$��&

��

��

��

�

�

��

��

��

��

��

��

Figure 38: BP Refinery study site


3.3.2 EC SURVEYS SUBMARINE POREWATER

Locations of EC surveys, dates and times of surveys and results are presented in

Figure 38, Table 12 and Figure 40, respectively. See Section 3.1.3 for additional

information about this technique.

A shoreline zone of SGD was observed inside the Small Boat Harbour (transects

BPT1 and BPT2). In addition, there was evidence of focused SGD up to 56 m

offshore—the maximum length of transect. Porewater EC’s varied from pure

seawater to almost freshwater values within distances as small as several metres.

This pattern might reflect shoreline SGD from the sand aquifer and offshore SGD

from the limestone via preferred pathways.

The pattern of SGD south of the harbour (transects BPT3-BPT5) was less distinct.

Table 12: Bulk-ground EC surveys; measurement locations and times

BP Refinery

Name Location (MGA94) Date Time (WST) Tide Stage

Easting [m] Northing [m]

BPT1 (Small Boat Harbour)

382988 6432706 12/11/01 12:40 Low

BPT2 (Small Boat Harbour)

382988 6432706 13/11/01 08:50 High

BPT2 382966 6432638 13/11/01 10:35 Falling BPT3 383000 6432568 13/11/01 10:55 Falling BPT5 383000 6432568 15/11/01 10:50 Falling



porewater samples were collected. Samples were obtained from approximately


ions, nutrients, EC and pH. All analyses were performed by the Chemistry Centre of

Western Australia. Results are reported in Table 13 and discussed in more detail in

Section 3.4. Note that there is extensive inshore contamination of groundwater by

nutrients in this area and this is reflected by high ammonia concentrations in the

porewater samples.

Table 13: BP Refinery porewater chemistry



Ionic Balance

[%] pH

01E0594/001 Seawater 12-Nov-01 <2 410 19000 5170 159 432 1290 5.4 0.08 1.2 6.6 10600 <0.01 <0.01 2790 0.9 8.3 01E0594/002 BPT1-5m 12-Nov-01 <2 118 1000 552 378 29 122 100 <0.02 <0.01 100 664 <0.01 0.08 714 -4.8 7.6 01E0594/004 BPT2-1m 13-Nov-01 <2 9 1800 657 214 4 165 2.3 0.92 12 15 1150 0.01 0.06 453 -1 8.3 01E0594/005 BPT2-12m 13-Nov-01 <2 113 1300 585 384 42 137 100 <0.02 <0.01 100 830 0.02 0.04 776 -3.7 7.6 01E0594/006 BPT2-19m 13-Nov-01 <2 206 4400 1560 421 115 356 85 <0.02 <0.01 85 2720 0.06 0.07 1140 1.5 7.7 01E0594/007 BPT2-31m 13-Nov-01 <2 171 4800 1720 409 142 437 64 <0.02 <0.01 64 3130 0.05 0.07 1390 3.9 7.8 01E0594/008 BPT2-45m 13-Nov-01 <2 127 960 512 403 31 127 100 <0.02 <0.01 100 666 0.01 0.05 713 -2.2 7.6 01E0594/009 BPT2-56m 13-Nov-01 <2 134 1100 514 390 34 127 100 <0.02 <0.01 100 693 0.01 0.05 748 -5.2 7.5 01E0594/010 BPT5-4m 15-Nov-01 <2 140 4400 1460 183 124 293 230 <0.02 110 340 2330 0.01 0.09 592 -4.3 8.2 01E0594/011 BPT5-10m 15-Nov-01 <2 243 9700 2810 214 260 668 100 <0.02 29 130 5570 0.06 0.08 1780 0 7.8

Notes: 1. All concentration are mg.l-1 2. N_NH3 – nitrogen ammonia fraction by FIA (filtered); N_NO2 – nitrogen nitrite fraction by FIA (filtered); N_NO3 – nitrogen nitrate fraction by FIA (filtered); N_total

– total nitrogen, persulphate by FIA (unfiltered); P_SR – phosphorus soluble reactive by FIA (unfiltered); P_total – total phosphorus, persulphate by FIA (infiltered) 3. FIA – Flow Injector Analysis



3.3.4 BENTHIC FLUX MEASUREMENTS

CLW’s heat-pulse BFM was deployed inside the Small Boat Harbour on only one

occasion. The corresponding time series of benthic flux measurements, which were

recorded at two-minute intervals, are depicted in Figure 40.

Measurements were made in shallow water within the shoreline zone of SGD that

was indicated by the porewater EC surveys—see BPT1 and BPT2 in Figure 39.

Similar to results obtained in Northern Harbour, both the direction and magnitude of

benthic flux were found to fluctuate in response to relatively small-scale variation in

sea level. There is clear evidence of a quasi-inverse relation between tide stage and

benthic flux but this is appears to be superimposed by fluctuations in benthic flux

due to waves and ocean swells.



BPT2 (Small Boat Harbour) - 0.5m

0 2 4 6 8 10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 Distance from Shoreline [m]


[mS/cm]

BPT3 (south of Small Boat Harbour) - 0.5m

0 2 4 6 8 10 12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 Distance from Shoreline [m]


[mS/cm]

BPT1 (Small Boat Harbour) - 0.5m

0 2 4 6 8 10 12 14 16

18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Distance from Shoreline [m]


[mS/cm]


0 2 4 6 8 10 12

14 16 18

0 1 2 3 4 5 6 7 8 9 10 Distance from Shoreline [m]


[mS/cm]


0 2 4 6 8 10

12 14 16 18

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Distance from Shoreline [m]


[mS/cm]

Figure 39: Bulk-ground EC measurements; BP Refinery transects; Blue bars indicate locations where water samples were collected



Figure 40: Benthic flux measurements from BP Refinery on 12/5/01 from 8:56 AM to 11:10 AM; measurements are every two minutes



3.4 COCKBURN SOUND SUBMARINE POREWATER SURVEY

A survey of submarine porewater was undertaken to assess the nutrient status of

groundwater discharging into Cockburn Sound at the location of discharge. Between

September 2001 and June 2002, approximately 80 samples were collected along the

foreshore of Cockburn Sound between Woodman Point and Palm Beach in

Rockingham. Sampling locations are indicated in Figure 41.

Results of sample are tabulated in Table 14. See Table 8, Table 11 and Table 13 for

transect results from Northern Harbour, Challenger Beach and BP Refinery,

respectively.

3.4.1 IONIC CHEMISTRY

Based on ionic balances carried out on 18 samples, the dominant porewater

chemistry is that of seawater diluted by various portions of fresh groundwater.

Major ion ratios are depicted in Figure 42.

It was concluded that the portion of fresh groundwater contributing to a porewater

sample could be estimated based on the chloride concentration of the sample, as

compared to the chloride concentration of undiluted seawater.

The relation between porewater EC and chloride concentration is depicted in Figure

43.

3.4.2 CALCULATIONS OF GROUNDWATER NUTRIENT CONCENTRATIONS

Nutrient concentrations in groundwater were estimated from the measured nutrient

concentrations in porewater, allowing for dilution of groundwater by seawater. The

following relation was used

pg NDIN ×= (1)

sp

sg

CiCl

ClClDI

−−

= (2)

where:Ng = nutrient concentration in groundwater; Np = nutrient concentration in

porewater mixture; DI = dilution index; Clg = chloride concentration in groundwater

(500 mg.l-1); Cls = chloride concentration in seawater (19,000 mg.l-1); Clp = chloride



concentration in porewater mixture. The relation assumes that all nutrients in the

porewater mixture are derived from the groundwater component. Values of the

dilution index for the collected porewater samples are listed in Table 14.

To collect the porewater samples, an EC probe (Section 3.1.3) was used to locate the

freshest porewater at each sample location. Porewater salinity varied considerably at

each site and typically varied over spatial scales of several meters. Figure 44

indicates how the value of the dilution index increases as the chloride concentration

(i.e. seawater component) of the porewater mixture increases. The median value of

the dilution index for all porewater samples was approximately 2, which is

equivalent to a porewater mixture containing approximately 50% seawater and 50%

groundwater.

Figure 45 and Figure 46 depict total nitrogen and total phosphorus concentrations in

groundwater that were estimated from porewater nutrient concentrations using (1)

and (2) above.

3.4.3 SYNTHESIS OF RESULTS

High nutrient concentrations in submarine porewater—up to 130 mg.l-1 in samples

P11 to P25—were detected along the section of coastline south of James Point. The

corresponding estimates of total nitrogen concentrations in groundwater were up to

200 mg.l-1 and are consistent with results from previous groundwater investigations

(see Section 5.2), which reveal high levels of nutrient contamination in this area. A

seawater sample collected inside of the Small Boat Harbour at BP Refinery had a

total nitrogen concentration of greater than 6 mg.l-1, which is in the order of 30 times

greater than the ‘background’ level in Cockburn Sound (DEP, 1996). Nutrient

enrichment in the harbour is assumed to result from groundwater discharge and

limited flushing by seawater. Nevertheless, there no history of algae blooms

apparent in the harbour.

Extremely high porewater nitrogen concentrations were detected along a short 40-m

section of shoreline north of James Point and adjacent the HIsmelt facilities (samples

P33 and P58 to P60). Total nitrogen concentrations in groundwater of up to

740 mg.l-1 were estimated based on submarine porewater concentrations of up to

220 mg.l-1. There is also evidence of significantly elevated nitrate levels in the

aquifer inshore from this area of coast. Nitrogen concentrations of up to 60 mg.l-1

N_NO3 have been measured in groundwater bores that are monitored by HIsmelt



(Corporate Environmental Consultancy, 2002); however, neither ammonia nor total

nitrogen has been monitored. The source of nutrients for this contamination is

unknown.

Elevated nutrient concentrations were also detected in samples P32 and P39, which

had estimated total nitrogen concentrations in groundwater of 31 and 10 mg.l-1,

respectively.

Slightly elevated concentrations were detected along a section of shoreline north of

Challenger Beach, which corresponds approximately to the location of the Naval

Base caravan park (samples B4 to B6, and P40 to P42). The average of estimated

total nitrogen concentrations in groundwater was approximately 8 mg.l-1 in this area,

compared to 4 mg.l-1 in other areas outside of the above zones of contamination.

Elevated nutrient concentrations in groundwater might be the result of leaching from

septic tanks; however, the source of these nutrients has not been established.

Contrary to previous groundwater investigations, nutrient enriched groundwater was

not detected in the Northern Harbour porewater samples (P1 and P2). However,

elevated nitrogen (ammonium concentrations in the in the range up to 20 mg.L-1 and

one occurrence of 130 mg.L-1) concentrations do occur in groundwater adjacent to

Northern Harbour (Table 7). This result suggests that the Northern Harbour

groundwater probably also discharge further offshore and beyond the immediate

shoreline strip of seabed beneath which the porewater samples were collected (see

Sections 3.1.3 and 3.1.8 for further information about groundwater discharge into

Northern Harbour). As a general principle, the further inshore that a contaminant

enters groundwater, the deeper in the aquifer it is likely to be by the time it reaches

the coast because groundwater flow toward the coast is driven slowly downward by

recharge to the water table. Waste injection by Weston Bioproducts was into saline

groundwater at the base of the superficial aquifer. Remediation operations to recover

the Northern Harbour plumes (PPK, 2000) might also have reduced SGD rates and

nutrient discharge into Northern Harbour at the time of porewater sampling.

Along Rockingham foreshore (sample locations RP1 to RP10) total nitrogen

concentrations in porewater samples ranged from < 0.5 to 5 mg.l-1. The

corresponding estimates of total nitrogen concentration in groundwater were between

1 and 26 mg.l-1. The highest estimates were obtained at sample locations RP2 and

RP9; however, these elevated levels may be artifacts of high-salinity porewater. The

chloride concentrations of both samples were greater than 15,000 mg.l-1, which



indicates that the porewater mixture was approximately 80% seawater and 20%

groundwater. Dilution indexes were greater than 5 and the measured porewater

concentrations were scaled by this amount to estimate groundwater concentrations.

Based on the eight other porewater samples from this section of coast, the estimated

total nitrogen concentration in groundwater was between 1 and 8 mg.l-1. In

comparison, Appleyard (1994) found that groundwater in bores 125601 to 125609

(Figure 55, Section 5.2)—located in the same area—had total nitrogen

concentrations between 1 and 4 mg.l-1.

Table 14: Porewater chemistry

Name Easting (AMG84) Northing (AMG84) Date Cl DI EC @ 25 °C [mS.m-1] N_NH3 N_NO2 N_NO3 N_total P_total pH

P1 383143 6443680 11-Feb-02 13000 3.08 3770 0.17 <0.02 <0.01 0.50 0.07 7.6 Seawater 383143 6443680 11-Feb-02 19000 - 4850 <0.01 <0.02 <0.01 0.15 <0.01 8.1 P2 383506 6443456 11-Feb-02 8200 1.71 2420 1.00 <0.02 <0.01 1.70 0.08 7.5 P3 384076 6440315 11-Feb-02 15000 4.63 3740 0.37 <0.02 <0.01 0.76 0.06 7.6 B1 384086 6440191 21-Feb-02 11000 2.31 2930 0.17 <0.02 <0.01 0.41 0.07 7.8 B2 384112 6440083 21-Feb-02 9900 2.03 2750 0.11 <0.02 <0.01 0.40 0.06 7.5 B3 384172 6439840 21-Feb-02 5100 1.33 1830 0.09 0.02 0.23 0.58 0.09 7.5 P4 384233 6439666 11-Feb-02 5500 1.37 1630 <0.01 <0.02 0.51 1.50 0.04 7.4 B4 384243 6439556 21-Feb-02 14000 3.70 3610 0.41 <0.02 0.06 1.40 0.14 7.6 B5 384308 6439381 21-Feb-02 6900 1.91 1990 0.01 <0.02 5.70 5.70 0.03 7.4 B6 384343 6439276 21-Feb-02 9300 1.53 2770 0.01 <0.02 2.80 3.20 0.02 7.5 P41 384408 6439137 14-Feb-02 13000 3.08 3290 <0.01 <0.02 2.40 3.70 0.03 7.8 P40 384512 6438849 14-Feb-02 11000 2.31 2860 <0.01 0.05 1.20 1.90 0.06 7.7 P42 384549 6438642 14-Feb-02 17000 9.25 4110 <0.01 <0.02 0.67 1.10 0.04 7.6 P43 384556 6438398 14-Feb-02 3300 1.18 1100 0.02 <0.02 <0.01 0.14 0.08 8.1 P48 384549 6438219 14-Feb-02 7600 1.62 2210 0.26 <0.02 <0.01 0.56 0.07 7.8 P47 384554 6437967 14-Feb-02 11000 2.31 2970 0.04 <0.02 <0.01 0.19 0.05 7.7 P46 384514 6437550 14-Feb-02 20000 - 4630 <0.01 <0.02 0.14 0.42 0.02 7.7 P45 384471 6437235 14-Feb-02 21000 - 4670 0.45 <0.02 <0.01 0.74 0.01 7.6 P44 384438 6437102 14-Feb-02 14000 3.70 3630 0.04 <0.02 <0.01 0.24 0.08 7.7 P39 384338 6436713 14-Feb-02 10000 2.06 2770 <0.01 <0.02 1.80 5.10 0.02 7.8

Notes: 1. All concentration are mg.l-1 2. N_NH3 – nitrogen ammonia fraction by FIA (filtered); N_NO2 – nitrogen nitrite fraction by FIA (filtered); N_NO3 – nitrogen nitrate fraction by FIA (filtered); N_total – total nitrogen, persulphate

by FIA (unfiltered); P_SR – phosphorus soluble reactive by FIA (unfiltered); P_total – total phosphorus, persulphate by FIA (unfiltered) 3. FIA – Flow Injector Analysis

Table 14: Porewater chemistry…continued

Name Easting

(AMG84) Northing (AMG84)

Date Cl DI EC @ 25 °C [mS.m-1]

N_NH3 N_NO2 N_NO3 N_total P_total pH

P38 384236 6436442 14-Feb-02 4400 1.27 1410 <0.01 <0.02 0.22 0.51 0.06 7.8 P37 384133 6436212 14-Feb-02 14000 3.70 3470 <0.01 <0.02 0.79 1.40 0.02 7.7 P36 383988 6435911 14-Feb-02 9900 2.03 2720 <0.01 <0.02 0.07 0.26 0.05 7.7 P35 383844 6435689 14-Feb-02 10000 2.06 2810 0.08 <0.02 <0.01 0.20 0.08 7.7 P34 383740 6435500 14-Feb-02 12000 2.64 3150 0.10 <0.02 <0.01 0.25 0.06 7.7 P63 383608 6435348 27-Mar-02 14000 3.70 3470 0.26 <0.02 <0.01 1.00 0.14 7.7 P62 383597 6435326 27-Mar-02 16000 6.17 3610 0.06 <0.02 <0.01 0.24 0.05 7.7 P61 383586 6435305 27-Mar-02 16000 6.17 3740 0.46 <0.02 <0.01 1.00 0.02 7.5 P60 383578 6435288 27-Mar-02 14000 3.70 3690 31 <0.02 <0.01 31.00 0.07 7.7 P33 383579 6435272 14-Feb-02 8800 1.81 2620 130.00 <0.02 <0.01 200.00 0.07 7.7 P59 383569 6435268 27-Mar-02 13000 3.08 3530 220 <0.02 <0.01 240.00 0.03 7.4 P58 383557 6435248 27-Mar-02 15000 2.31 3450 4.1 <0.02 <0.01 4.50 0.01 7.4 P57 383548 6435229 27-Mar-02 10000 2.06 2720 0.28 <0.02 <0.01 0.41 0.11 7.7 P56 383535 6435207 27-Mar-02 16000 6.17 4180 0.09 <0.02 <0.01 0.31 0.09 7.6 P55 383522 6435183 27-Mar-02 16000 6.17 3960 0.75 <0.02 <0.01 1.30 0.08 7.6 P54 383511 6435162 27-Mar-02 15000 4.63 3900 0.48 <0.02 <0.01 0.80 0.1 7.6 P53 383500 6435141 27-Mar-02 15000 4.63 3840 0.57 <0.02 <0.01 0.83 0.11 7.6 P52 383483 6435122 27-Mar-02 17000 9.25 3670 0.04 <0.02 <0.01 0.21 0.17 7.7 P49 383468 6435103 27-Mar-02 18000 18.50 4360 0.06 0.07 1.6 2.90 0.07 7.5 P50 383468 6435103 27-Mar-02 16000 6.17 3880 0.05 <0.02 <0.01 0.28 0.01 7.5 P32 383472 6435100 14-Feb-02 13000 3.08 3430 4.10 <0.02 <0.01 10.00 0.04 7.6 P31 383400 6434972 14-Feb-02 7900 1.67 2350 0.07 <0.02 0.09 0.43 0.08 7.9

Notes: 1. All concentration are mg.l-1 2. N_NH3 – nitrogen ammonia fraction by FIA (filtered); N_NO2 – nitrogen nitrite fraction by FIA (filtered); N_NO3 – nitrogen nitrate fraction by FIA (filtered);

N_total – total nitrogen, persulphate by FIA (unfiltered); P_SR – phosphorus soluble reactive by FIA (unfiltered); P_total – total phosphorus, persulphate by FIA (unfiltered)

3. FIA – Flow Injector Analysis


Name Easting




P30 383150 6434659 14-Feb-02 12000 2.64 3490 0.12 <0.02 0.08 0.56 0.11 7.7 P29 383033 6434504 14-Feb-02 12000 2.64 3160 0.14 <0.02 0.21 0.55 0.09 7.6 P28 382880 6434181 14-Feb-02 9500 1.95 2690 0.27 <0.02 <0.01 0.63 0.06 7.6 P26 382793 6433973 14-Feb-02 8800 1.81 2460 0.87 <0.02 <0.01 1.20 0.06 7.6 P27 382727 6433778 14-Feb-02 21000 - 4780 0.25 <0.02 <0.01 0.67 0.13 7.7 P25 383004 6433464 13-Feb-02 3200 1.17 1070 1.60 <0.02 <0.01 2.50 0.12 7.5 P24 383068 6433230 13-Feb-02 1900 1.08 705 6.30 <0.02 <0.01 11.00 0.09 7.4 P23 383069 6433023 13-Feb-02 2200 1.10 707 59.00 <0.02 <0.01 86.00 0.03 7.3 P22 383128 6432873 13-Feb-02 1100 1.03 558 48.00 <0.02 <0.01 100.00 0.08 7.4 P18 383135 6432729 13-Feb-02 4400 1.27 1430 76.00 <0.02 0.01 120.00 0.14 7.8 P20 383154 6432688 13-Feb-02 7100 1.55 2200 34.00 <0.02 20.00 96.00 0.03 8.1 P17 383157 6432674 13-Feb-02 6800 1.52 2070 85.00 <0.02 7.50 130.00 0.05 7.7 Seawater 383157 6432674 13-Feb-02 19000 - 4620 1.10 <0.02 0.22 2.00 0.01 8.0 P21 383165 6432650 13-Feb-02 7000 1.54 2160 61.00 <0.02 11.00 120.00 0.07 7.7 P16 383172 6432604 13-Feb-02 4200 1.25 1420 29.00 <0.02 <0.01 120.00 0.13 7.4 P15 383177 6432540 13-Feb-02 2100 1.09 876 63.00 <0.02 <0.01 79.00 0.65 7.3 P14 383110 6432080 13-Feb-02 3800 1.22 1240 24.00 <0.02 <0.01 36.00 0.10 7.8 P13 383063 6431875 13-Feb-02 5300 1.35 1620 6.20 <0.02 0.08 9.30 0.13 7.8 P12 383018 6431724 13-Feb-02 10000 2.06 2860 6.60 <0.02 0.34 12.00 0.09 7.7 P11 382980 6431628 13-Feb-02 14000 3.70 3630 0.81 <0.02 <0.01 1.40 0.10 7.6 P10 382937 6431538 13-Feb-02 7400 1.59 2140 0.71 <0.02 <0.01 1.40 0.13 7.6 P9 382563 6430768 13-Feb-02 9600 1.97 2680 0.48 <0.02 <0.01 1.20 0.07 7.6





Name Easting




P8 382326 6430487 13-Feb-02 9500 1.95 2690 0.05 <0.02 <0.01 0.20 0.07 7.6 P7 382061 6430184 11-Feb-02 11000 2.31 3180 0.18 <0.02 <0.01 0.38 0.16 7.7 P6 381884 6429943 11-Feb-02 3700 1.21 1200 0.16 <0.02 <0.01 0.38 0.12 8.0 P5 381579 6429513 11-Feb-02 6900 1.53 2040 0.24 <0.02 <0.01 0.48 0.04 7.7 RP11 381186 6429081 11-Jun-02 8,600 1.78 2450 2.1 <0.02 <0.01 2.3 0.07 7.8 RP10 381004 6428916 11-Jun-02 8,600 2.31 3120 0.01 <0.02 0.15 0.33 0.07 7.8 RP9 380811 6428758 11-Jun-02 11,000 5.97 3360 2.3 <0.02 <0.01 2.3 0.08 7.7 RP8 380625 6428604 11-Jun-02 15,900 5.14 3310 1.0 <0.02 <0.01 1.1 0.06 7.8 RP7 380453 6428421 11-Jun-02 15,400 2.31 2870 3.3 <0.02 <0.01 3.5 0.08 7.7 RP6 380248 6428289 11-Jun-02 11,000 1.67 2220 0.15 <0.02 <0.01 0.34 0.06 7.5 RP4 380026 6428172 11-Jun-02 7,900 1.83 2450 0.58 <0.02 <0.01 0.69 0.08 7.9 RP3 379803 6428059 11-Jun-02 8,900 1.08 687 0.46 <0.02 <0.01 0.62 0.07 7.9 RP1 379305 6428043 11-Jun-02 1,900 2.31 3300 0.27 <0.02 <0.01 0.47 0.09 7.8 RP2 379554 6428022 11-Jun-02 11,000 5.14 2880 4.9 <0.02 <0.01 5.0 0.34 7.3






��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�� !��"�#$��

�%&

��'��&��(&

��%��)��'��

!��*��$*��+��'��)��

!��*��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�

��

��

��

��

&��,��)�� $��%��'-��%��'�(��-��'��(��'��,��)�')��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 41: Porewater sampling locations



Major Ion Ratios (November 2001 sampling)

-1,000

0

1,000

2,000

3,000

4,000

5,000

6,000

7,000

8,000

9,000

10,000

11,000

12,000

0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 20,000 Chloride Concentration [mg.l -1 ]

Ionic Concentration

[mg.l -1 ]

Sodium Magnesium Potassium Sulphate Calcium Bicarbonate

Figure 42: Porewater ionic chemistry

Porewater Chloride-EC Relation

0

1,000

2,000

3,000

4,000

5,000

6,000

0 5,000 10,000 15,000 20,000 Chloride Concentration [mg.l -1 ]

Electrical Conductivity

@ 25°C [mS.m -1 ]

Nov 2001 sampling Feb 2002 sampling Mar 2002 sampling Jun 2002 sampling

Figure 43: Porewater chloride-EC relation



Porewater Dilution Index

0

5

10

15

20

25

30

35

40

0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 18,000 Cl [mg.l -1 ]

DI

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Percent Seawater

Porewater sample Dillution index Percent seawater

Figure 44: Porewater dilution index



��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�� ! !� "��

�#��$!$ �%��&��'��( ��)#*�&+

��'��,��(��,��(��

-�!��

!�.�!��

!��.� �� .�!��/�!��

&��'��( �� 0��)�� 0��)��+��)��+�,��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 45: Groundwater nitrogen concentrations [mg.l-1] estimated from submarine porewater analyses



��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

�� !��"�#$��

�%&

��'��&��(&

��%��)��'��

��

!��*��

!��*��$*��+��'��)��

��

��

��

��

��

��

��

��

��

��

��

(��',��&*�)*��-��.�/��'�$��(��&��,��&*�)*��

��0��

��1�0��

��0�2��

��$��%��'3��$��%�(��3��'��(��'��,��)�')��

��

Figure 46: Groundwater phosphorus concentrations [mg.l-1] estimated from submarine porewater analyses



4 QUANTIFYING SGD INTO COCKBURN SOUND

Submarine groundwater discharge can be quantified in two ways:

a). By inference from inshore water balance considerations

b). By direct and indirect field measurements of SGD.

Each approach can be employed independently of the other, with the advantage that

results can be crosschecked to reduce uncertainty in estimates. Field measurements

of SGD should be physically realistic and feasible when considered in context of the

inshore hydrogeology and aquifer water balance.

4.1 CONCEPTUAL MODELS OF SGD

Groundwater in shallow, coastal aquifers flows generally toward the ocean along

pathways of least resistance. If an aquifer is homogeneous, flow is mainly toward

the shoreline because it represents the shortest travel distance for groundwater to

discharge from the aquifer. In an inhomogeneous aquifer, groundwater will also

discharge offshore from the coast if either (a) low-conductivity sediment restricts

vertical flow toward the shoreline, or (b) high-conductivity zones, such as fractures

and solution features, focus SGD to other locations. Both of these possibilities occur

in Cockburn Sound.

A number of conceptual models of groundwater discharge into Cockburn Sound are

feasible and six possibilities are depicted in Figure 47. Other patterns of

groundwater discharge are also possible. In all cases, groundwater could discharge

further offshore and beyond the indicated zones of seawater intrusion due to

preferred flow through low-resistance pathways in the limestone. Potentially,

groundwater could discharge many hundreds of metres offshore by this mechanism.

Note, also, that there is potential for artesian groundwater from the Leederville

Aquifer to discharge into Cockburn Sound through overlying limestone. This

possibility is not depicted in Figure 47.

Absence of offshore hydrogeological data is the main limitation to identifying which

type of SGD is most likely to occur at different locations along Cockburn Sound



foreshore. In particular, the extent of basal clay is poorly described across most of

the study area and there is virtually no offshore hydrogeological data.

In Type I SGD, coastal sands overlie Tamala Limestone and the basal clay unit

between them is absent. Because there is no significant impediment to vertical flow

through the sands, groundwater discharges upward toward the sea and SGD is

focused along a relative narrow strip of the shoreline. The average hydraulic head

along the seabed is uniform and therefore hydraulic gradients in the aquifer are

steepest toward the closest point on the seabed, which is the shoreline.

In Type II, the shoreline zone of SGD is narrower because the coastal sands are

absent and high permeability limestone is in direct contact with the sea. This

situation is thought to occur between Northern Harbour and Challenger Beach where

limestone is exposed along the shoreline as a coastal cliff. It is emphasised, once

more, that groundwater probably discharges further offshore through secondary

porosity in the limestone.

Type III depicts the situation in which basal clay is present between the sand and

limestone but does not extend offshore. Groundwater flow toward the coast is ‘split’

vertically either side of the clay, though groundwater in the limestone can move

upward and discharge to the shoreline once it has passed beneath the clay. Note that

the dominant groundwater motion in the coastal sands inshore of the clay is vertical.

This is because it is relatively ‘easier’ for groundwater in the sands to drain vertically

downward into high-permeability limestone, travel laterally beneath the clay and

then discharge vertically upward to the ocean.

Basal clay extends a significant distance offshore in Type IV, which prevents

groundwater in the limestone discharging vertically at the shoreline. Two zones of

SGD occur—one along shoreline and the other further offshore—and saltwedges

form above and below the clay. The offshore zone of SGD may be poorly defined

dependent on the competency of the clay as an aquitard and the way it terminates

offshore (e.g. as a gradually thinning layer with poorly defined boundary or abruptly

with a relatively well-defined boundary). Inshore, water levels in the sand and

limestone are effectively ‘de-coupled’ by the aquitard, with tidal fluctuations

propagating more efficiently into high-permeability and semi-confined limestone.

Local, shallow groundwater mounds may develop in the sands above the clay in

response to local recharge.



Type V depicts the case where the basal clay extends offshore and pinches out at the

base of a submarine limestone ridge or similar elevation feature in the top surface of

the limestone. The offshore zone of SGD may be more distinct—compared to

Type IV—because the limestone is in direct contact with the sea at the location of

groundwater discharge.

In Type VI, basal clay extends a sufficient distance offshore that groundwater in the

limestone is unable to flow beyond it. Groundwater ‘leaks’ up through the aquitard

and overlying sand to emerge as diffuse SGD.

Figure 47 presents a two-dimensional, vertical impression of SGD; however, the

coastal hydrogeology also affects horizontal patterns of groundwater flow. For

example, there is evidence of local mounding of shallow groundwater above basal

clay in the sand aquifer beneath BP Refinery (see Section 3.3.1). As a general

principle, it is expected that complex hydrogeology will be reflected by complex

patterns of shoreline and offshore SGD. To completely resolve these flow systems

would require systematic field investigations across large spatial and temporal scales.

Such detailed measurements may be impracticable to make in many situations.



Cockburn Sound

Tamala

Limestone

Cretaceous Sediment

Quaternary

Sand

Shoreline

SGD

Saltwater

Freshwater

Freshwater

Saltwater

Cockburn Sound

Tamala

Limestone

Cretaceous Sediment

Quaternary

Sand

Shoreline

SGD

SaltwaterFreshwater

Saltwater

Cockburn Sound

Tamala

Limestone

Cretaceous Sediment

Shoreline

SGD

Saltwater

Freshwater

Saltwater

TYPE I

TYPE II

TYPE III

Figure 47: Conceptual models of SGD into Cockburn Sound



Cockburn Sound

Tamala

Limestone

Cretaceous Sediment

Quaternary

Sand

Shoreline

SGD

Saltwater

Offshore

SGD

Freshwater

Freshwater

Saltwater

TYPE IV

Cockburn Sound

Tamala

Limestone

Cretaceous Sediment

Quaternary

Sand

Shoreline

SGD

Saltwater

Offshore

SGD

Freshwater

Freshwater

Saltwater

TYPE V

Saltwater

Cockburn SoundTamala

Limestone

Cretaceous Sediment

Quaternary

Sand

Shoreline

SGD

Saltwater

Diffuse offshore SGD

Freshwater

Freshwater

Saltwater

TYPE VI

Saltwater

Salt-

waterSalt-

water

Figure 47: Conceptual models of SGD into Cockburn Sound…continued



4.2 PREVIOUS ESTIMATES OF SGD

Table 15 summaries previous estimates of SGD into Cockburn Sound. While some

of the tabulated values were reported directly in the referenced literature, others were

implied by the reported values of hydraulic gradient, hydraulic conductivity and

aquifer thickness. In these cases, SGD in Table 15 was calculated by applying

Darcy’s law.

Table 15: Estimates of Submarine Groundwater Discharge into Cockburn Sound and related sites

Reference SGD Rate [m3.d-1.m-1]

Location and Method Assumed Values

Allen (1981) ~ 8 (superficial.) North Perth coastline, flownet

-

Bodard (1991) 2.3 - 5.2 (superficial) 0.1 - 0.6 (sand) 2.2 - 4.6 (limestone)

BP Refinery, flownet Sand: v = 15 - 73, B = 10, n = 0.3; Limestone: v = 270 - 562, B = 10, n = 0.3

WAWA (1993) 3.0 (superficial) Cockburn Groundwater Area, flownet water balance (1992 scenario)

Hamilton Hill: k = 100 - 200; Henderson: k = 100 - 200; Kwinana Beach: k = 500 - 800

Appleyard (1994) 0.7 - 2.5 (superficial) 0.1 - 0.5 (sand) 0.6 - 2.0 (limestone)

Cockburn Sound, flownet k = 10 - 50, i = 0.001 - 0.003

Davidson (1995) 2.5 (superficial) Jandakot groundwater mound, regional flownet

T > 4,000

HGM (1998) 3.0 - 8.0 (superficial) Northern Harbour Jervoise Bay, offshore CTD profiling

-

Nield (1999) 2.3 - 4.4 (superficial) Cockburn Groundwater Area, model calibration on hydraulic head (1999 scenario)

k = 900

PPK (2000) 5.0 (mean) 3.0 - 8.0 (seasonal)

Northern Harbour Jervoise Bay, flownet

T = 25,000, i = 0.0002

PPK (2000) 3.0 - 9.0 (superficial) Northern Harbour, harbour salinity measurements

-

Symbols and Units v: groundwater velocity [m.d-1] n: porosity [1] B: saturated thickness of aquifer unit [m] k: saturated hydraulic conductivity [m.d-1] i: hydraulic gradient [1] T = kB: aquifer transmissivity [m2.d-1]



4.2.1 FLOWNET CALCULATIONS

Bodard (1991) estimated that groundwater pore velocities in the area of the BP

Refinery varied in the range 15 to 73 m.yr-1 in coastal sands and 270 to 562 m.yr-1 in

Tamala Limestone. Based on average saturated thickness of sand equal to 10 m,

average thickness of limestone equal to 10 m, and porosity 0.3 for sand and

limestone, the estimated groundwater velocities equate to SGD of 0.1 to

0.6 m3.d-1.m-1 from the sand and 2.2 to 4.6 m3.d-1.m-1 from the limestone. The sum

of these give total SGD from the Superficial Formations of between 2.3 and

5.2 m3.d-1.m-1.

WAWA (1993) estimated that total SGD into Cockburn Sound in 1992 was

75,342 m3.d-1. A flownet approach was used and a hydraulic conductivity value of

200 m.d-1 was assumed for the Superficial Formations. Distributed uniformly over

25 kilometres of coast, this is equivalent to an average SGD of around 3.0 m3.d-1.m-1.

Appleyard (1994) calculated that SGD into Cockburn Sound was in the range 0.1 to

0.5 m3.d-1.m-1 from superficial sands and 0.6 to 2 m3.d-1.m-1 from the Tamala

Limestone. Hydraulic conductivity values between 10 and 50 m.d-1 were assumed

and applied with gradients between 0.001 and 0.003 to compute SGD.

Based on groundwater flownet calculations, Davidson (1995) estimated that total

SGD along the coastline of the Jandakot mound was approximately 66,450 m3.d-1,

which equates to average SGD along the shoreline of around 2.5 m3.d-1.m-1.

Transmissivity of the Superficial Formations was assumed greater than 4,000 m2.d-1.

More recently, PPK (2000) estimated that average SGD into Northern Harbour was

approximately 5 m3.d-1.m-1 based on average transmissivity 25,000 m2.d-1 and

average hydraulic gradient 0.0002. The estimated seasonal range of SGD in that

study was 3 to 8 m3.d-1.m-1.

Thus, previous estimates of SGD from Superficial Formations, based on flownet

analyses, have varied between 0.7 and 8 m3.d-1.m-1; a factor of approximately ten. A

number of the estimates arrived at a similar final value of SGD; however, they were

based on significantly different values of hydraulic conductivity and hydraulic

gradient, which vary between some estimates by an order of magnitude. These

differences highlight the inherent uncertainty in flownet analyses that are based on

water level data alone.



4.2.2 CTD PROFILING

Halpern Glick Mansel (1998) used results of conductivity-temperature-depth (CTD)

profiling in Jervoise Bay to estimate that SGD into Northern Harbour was

approximately 3-8 m3.d-1.m-1. Also using CTD data, DAL (2000) estimated that

SGD into Northern Harbour was 3-9 m3.d-1.m-1. In both studies, SGD was estimated

to be the quantity of fresh groundwater that was required to dilute seawater in the

harbour to the observed values of salinity.

Because these estimates were derived without reference to aquifer hydrogeology,

they provide important independent checks on the values of SGD calculated from

inshore flownet analyses. Nevertheless, seawater inside Northern Harbour is not

uniformly diluted by groundwater and therefore the above estimates of SGD involve

assumptions about mixing of groundwater and seawater in the harbour and the

harbour-flushing rate. These assumptions have significant influence on the estimated

value of SGD.

4.3 DIRECT MEASUREMENT OF SGD

Direct measurements of SGD by BFM indicated that shoreline SGD in Cockburn

Sound is strongly influenced by sea level dynamics and associated fluxes of seawater

into and out of the seabed. Based on the experimental results described in Sections

3.1.6 and 3.3.4, and the modelling results in Appendix C, it is concluded that the

fresh-groundwater component of total benthic flux measured by BFM cannot be

distinguish from the seawater component.

It is also clear from EC surveys of submarine porewater (Sections 3.1.3, 3.2.2 and

3.3.2) that benthic flux is distributed non-uniformly over the seabed and appears to

be relatively ‘patchy’ in distribution. Even if it were possible to measure the fresh-

groundwater component of benthic flux, an impractical amount of data would need

to be collected to enable point measurements of SGD to be interpolated and

integrated to yield a value of total SGD into Cockburn Sound.

Direct, point measurements of SGD using benthic flux meters are of limited

usefulness to develop aggregated estimates of total SGD. Benthic flux

measurements are strongly influenced by seawater exchange between the seabed

sediments and ocean in response to short time-scale sea level variations such as tidal



fluctuations. In addition, the field measurements indicate that the distribution of

SGD is patchy at spatial scales of less than ten metres. To properly describe the

distribution of SGD across the study area and aggregate volumetric estimates of SGD

would require an impracticable number of measurements. However, significant

insight into the flow and discharge mechanisms of the SGD flux was obtained from

high temporal resolution measurements by benthic flux meters, indicating at the

study site that tidal forcing is the dominant factor controlling the SGD process at

Cockburn Sound. Geochemical and isotope based estimates of SGD on the other

hand provide a more reliable regional scale estimate of the net SGD flux per unit

length of shoreline, and these measurements results are more comparable to

estimates of SGD based on onshore numerical modelling of groundwater fluxes and

groundwater balance. These figures are the most useful to coastal zone

environmental managers with responsibilities for protecting the near-shore marine

environment from land-based contaminant sources.

It is concluded that terrestrial SGD into Cockburn Sound cannot be reliably

estimated by direct measurement techniques employed in this study.

4.4 ESTIMATION OF SGD BY INSHORE WATER BALANCE

The alternative to estimating SGD by direct and indirect measurement techniques is

to infer SGD from the inshore aquifer water balance.

This section describes the application of a groundwater flow model to quantify the

aquifer water balance inshore from Cockburn Sound and estimate SGD into

Cockburn Sound. Estimates of groundwater discharge are obtained from the model

water budget of the modeled domain as the residual between groundwater inflow, net

recharge and pumping. Simulated values of SGD vary spatially along the length of

the coastline due to (a) variation in hydraulic connection between the coastal geology

and ocean, (b) spatial variability in groundwater recharge, aquifer transmissivity and

pumping, and (c) focusing of groundwater flow toward coastal embayments.

4.4.1 GROUNDWATER FLOW MODEL

A pre-existing groundwater flow model (Nield 1999) was used to estimate SGD.

The model was originally constructed in MODFLOW (McDonald and Harbaugh 1988).



In this study it was implemented in Visual MODFLOW v.2.8.2 (Waterloo Hydrologic

2000). A summary of the model structure is presented below:

• The finite difference grid (Figure 48) consists of 26,596 cells, which are each 100

× 100 m.

• Superficial Formations inshore from Cockburn Sound are represented as a single-

layer, unconfined aquifer with impermeable base. A single-layer model was

justified on the basis that Tamala Limestone is highly transmissive and most

groundwater flow in the Superficial Formations is conducted laterally through

secondary porosity in the limestone. It was assumed that groundwater movement

in overlying sands is predominantly downward (vertical) into the limestone in

response to local rainfall recharge. Contours corresponding to bottom elevations

of the Superficial Formations (Figure 49) vary from around –12 m Australian

Height Datum (AHD) at the southeast corner of the model domain to

approximately –30 m AHD in the vicinity of Northern Harbour.

• A line of general head cells (Figure 48) was used to represent the coastal

boundary and conductance values were assigned that varied according to which

geological unit contacts the ocean. Table 16 lists the conductance and external

head values, which were determined during calibration (see Section 4.4.3).

Where Tamala Limestone is exposed along the shoreline, the assigned

conductance value was higher than at locations where the aquifer discharges to

the ocean through overlying, relatively lower-permeability sand. External head

along the coast was set equal to mean sea level of -0.05 m AHD.

• Constant head cells (Figure 48) were specified along the inland boundary. Head

was set equal to 20 m AHD along the northern portion of boundary,

corresponding to the 20-m water table elevation contour. Along the southern

portion of boundary, head was set equal to invert levels along the Peel Main

Drain.

• The north and south boundaries of the model lie along assumed lines of regional

groundwater flow and were specified as no-flow boundaries.

• More than 700 pumping wells are simulated in the model (Figure 50). Unless

recorded pumping data was available—data was available for less than 1% of

wells—the abstraction rates from bores were set equal to 80% of the licensed

groundwater allocation limit. Therefore, the modelled groundwater abstraction

reflects the total allocation allowance rather than total abstraction and is possibly



an overestimate of pumping withdrawals. Thus, SGD estimate is possibly

underestimated

Table 16: Cockburn Groundwater Area model: conductance and external head values for the general head coastal boundary cells

Low-recharge Scenario High-recharge Scenario Conductivity Zone

Geological Unit Conductance

[m.d-1] External Head [m]

Conductance [m.d-1]

External Head [m]

1 Safety Bay Sand 500 -0.05 1000 -0.05 2 Tamala Limestone 1000 -0.05 2000 -0.05

4.4.2 GROUNDWATER RECHARGE

Recharge is the key water balance component that affects the model estimates of

SGD; however, there is significant uncertainty in quantifying recharge. Low

groundwater recharge implies less groundwater flow through the aquifer and low

SGD, while high recharge implies larger SGD. Two different scenarios were

simulated to provide estimates of SGD under assumptions of low and high average

groundwater recharge. Spatial distributions of recharge were assigned according to

land use and vegetation cover, including the categories: bushland, cleared or open

bushland, agriculture and horticulture, and industrial and urban. Each model cell

was assigned a land-use code and each land-use code assigned a groundwater

recharge rate. The resulting recharge distributions for low-recharge and high-

recharge scenarios are depicted in Figure 51 and Figure 52, respectively.

Recharge rates assigned to the different land use categories varied between 0% and

30% of mean annual rainfall in the low-recharge scenario, and between 0% and 40%

in the high-recharge scenario. The equivalent average recharge rates applied over the

entire model domain were 0.13 m.yr-1 (i.e. 14.7% of mean annual rainfall) and

0.24 m.yr-1 (i.e. 27.5% of mean annual rainfall), respectively.

4.4.3 CALIBRATION

Nield (1999) calibrated low-recharge and high-recharge scenarios by trial and error.

Hydraulic conductivity of each zone in Figure 48 was adjusted to obtain suitable



matches between the observed and modelled hydraulic heads. The calibrated values

of hydraulic conductivity are presented in Table 17 and model-simulated water table

contours are depicted in Figure 53. Conductance values of the coastal boundary cells

(Table 16) were also adjusted during calibration.

Table 17: Cockburn Groundwater Area model: calibrated hydraulic conductivity values

Hydraulic Conductivity [m.d-1] Conductivity Zone Low-recharge

Scenario High-recharge

Scenario

1 400 800 2 1660 3000 3 540 1000 4 500 900 5 25 30 6 6 8 7 34 34 8 20 20 9 120 120

10 25 35 11 10 14 12 30 30 13 60 60

4.4.4 RESULTS

SGD from each coastal boundary cell is depicted in Figure 53 for both low-recharge

and high-recharge scenarios. The modelled distributions of SGD along the coast

varied mainly in response to change in hydraulic connection between the aquifer and

ocean. SGD was largest along two sections of coastline, north and south of

Woodman Point, where Tamala Limestone directly contacts the ocean. Smaller

variations in SGD occurred in response to variability in recharge intensity inshore

from the coast, and variation in the density and distribution of pumping bores. In the

low-recharge scenario, SGD from the coastal boundary cells varied spatially between

1.4 and 4.6 m3.d-1.m-1. This increased to between 2.4 and 7.9 m3.d-1.m-1 in the high-

recharge scenario.

The model water budgets for both recharge scenarios are presented in Table 18 and

Table 19. They confirm that SGD and groundwater pumping (out flows) were

balanced by groundwater recharge and flow across the inland boundary (in flows).

Because all terms in the water balance are similar order of magnitude, uncertainty

and errors in estimations of groundwater inflow, recharge and pumping are all



manifest as similar order-of-magnitude error and uncertainty in SGD. Total SGD in

the low-recharge scenario was 48% of total groundwater out flow from the model

and groundwater abstraction from bores was approximately 46%. The remaining 6%

of out flow was minor discharges to the Peel Main Drain. In the high-recharge

scenario, SGD increased to 64% of total out flow and pumping reduced to 32%.

Based on these estimates, none of the water balance terms can be neglected as

insignificant.

Allowing for low and high estimates of recharge, and arbitrary ±20% errors in

pumping and groundwater flow across the inland boundary, mean SGD is estimated

to be in the range 2.5-4.8 ± 0.9 m3.d-1.m-1. Thus, total average SGD could be

between 1.6 and 5.7 m3.d-1.m-1, which is a factor of variation of approximately 3.5.

This order-of-magnitude uncertainty stems from uncertainty in the reliability and

accuracy of the available hydrologic and hydrogeologic data.

Table 18: Cockburn Groundwater Area model: model water balance for low-recharge scenario

Balance Term Flow In [Kl.d-1]

Flow Out [Kl.d-1]

Net Flow [Kl.d-1]

Inland boundary 56,982 -6,855 50,127 Groundwater recharge 65,288 0 65,288 Groundwater Abstraction 0 -58,245 -58,245 Artificial recharge 3,836 0 3,836 SGD 0 -61,007 -61,007

Total 126,106 -126,107 -1

Mean SGD = 61,007/24,000 = 2.5 m3.d-1.m-1 of shoreline

Table 19: Cockburn Groundwater Area model: model water balance for high-recharge scenario

Balance Term Flow In [Kl.d-1]

Flow Out [Kl.d-1]

Net Flow [Kl.d-1]

Inland boundary 55,186 -7,046 48,140 Groundwater recharge 121,750 0 121,750 Groundwater Abstraction 0 -58,245 -58,245 Artificial recharge 3,836 0 3,836 SGD 0 -115,480 -115,480

Total 180,772 -180,771 1

Mean SGD = 115,480/24,000 = 4.8 m3.d-1.m-1 of shoreline



4.4.5 CONCLUSIONS

Based on aquifer water-balance considerations, total average SGD across the

coastline adjacent to Cockburn Sound is estimated to be 2.5-4.8 ± 0.9 m3.d-1.m-1.

Most of this groundwater is believed to discharge to the ocean via preferred

pathways in Tamala Limestone. The estimated range in SGD, 2.5-4.8 m3.d-1.m-1,

corresponds to low and high estimates of average groundwater recharge. The error,

± 0.9 m3.d-1.m-1, is calculated based on arbitrary ±20% errors in pumping and

groundwater inflow across boundaries. The magnitude of this error could be larger.

In the modelling, differences in hydraulic connection between the aquifer and ocean

were observed to have the largest affect on the spatial variability of SGD.

Nevertheless, the modelled distributions of SGD cannot be corroborated against

independent data—since none exists—and are considered to have large uncertainty.

For example, the model can be successfully calibrated using conductance values

along the coast that result in significantly different spatial distributions of SGD,

compared to those depicted in Figure 53. The model provides regional-scale

estimates of SGD that do not take into account the detailed hydrogeology, and

associated flow paths and rates that control SGD at the local scale.

It is concluded that the estimated values of total average SGD are consistent with the

available regional hydrogeological and hydrologic data for the study area.

Nevertheless, results from local-scale measurements are affected by local-scale

processes and should not be constrained by regional-scale estimates derived from the

modelling. Focusing of flow through preferred pathways and tidal effects, for

example, may results in local SGD significantly larger or smaller than the regional-

scale estimates.



1

3 4

9

7

13

13

56

2

2

1

1

4

1011 8

813

13

13

13

12

13

13

13

13

0 2,500 5,0001,250Meters

�

LegendHydraulic conductivity zoneModel gridGeneral head cellDrain cell (set to drain invert levels)Constant head cell (20 mAHD)Inactive cell

COCKBURNSOUND

Figure 48: Cockburn Groundwater Area flow model; finite difference grid, boundary conditions and hydraulic conductivity zones



-25

-24

-23

-22

-21

-20

-19

-28

-29

-30

-17

-15

-14

-13

-12

-17

-20

0 2,500 5,0001,250Meters

�

LegendBasement elevation contour [mAHD]General head cellDrain cell (set to drain invert levels)Constant head cell (20 mAHD)Inactive cell

COCKBURNSOUND

Figure 49: Cockburn Groundwater Area flow model; base of Superficial Formations



537

809

717

745

835

699

3247

1800

4540

1403

4186

1753

3206 1534

1206

16774110

1808

0 2,500 5,0001,250Meters

�

LegendBore abstraction [Kl/day]

0 - 8788 - 395396 - 835836 - 1,808

1,809 - 4,540

Artificial recharge [Kl/day]274411General head cellDrain cell (set to drain invert levels)Constant head cell (20 mAHD)Inactive cell

COCKBURNSOUND

Figure 50: Cockburn Groundwater Area flow model; groundwater abstraction and artificial recharge locations and rates



0 2,500 5,0001,250Meters

�

LegendLow Recharge ModelPercent rainfall recharge

0%10%20%25%30%General head cellDrain cell (set to drain invert levels)Constant head cell (20 mAHD)Inactive cell

COCKBURNSOUND

Figure 51: Cockburn Groundwater Area flow model; groundwater recharge rates for the low-recharge scenario, recharge is expressed as a percentage of mean annual rainfall of 0.87 m/yr



0 2,500 5,0001,250Meters

�

LegendHigh Recharge ModelPercent rainfall recharge

0%22%32%40%General head cellDrain cell (set to drain invert levels)Constant head cell (20 mAHD)Inactive cell

COCKBURNSOUND

Figure 52: Cockburn Groundwater Area flow model; groundwater recharge rates for the high-recharge scenario, recharge is expressed as a percentage of mean annual rainfall of 0.87 m/yr



0

2

3

1

4

5

6

78

9

0.5

1011

1213

14 15

16

17

18 19

20

0 2,000 4,0001,000Meters

�

LegendLow RechargeHigh RechargeGeneral head cellDrainConstant headInactive cell

COCKBURNSOUND

0246810

SGD [Kl/d/m]

Figure 53: Cockburn Groundwater Area flow model; calibrated water table elevations and model computed SGD for the low-recharge and high-recharge scenarios



4.5 ESTIMATION OF SGD BASED ON SALTWATER-FRESHWATER RELATIONS

Figure 54 depicts saltwater interface profiles that were calculated using the abrupt

interface model of Glover (1959). Glover’s solution expresses the vertical depth, z

[L], to the saltwater interface as a function of the horizontal distance, x [L], inshore

from the coast

2

2 2⎟⎠⎞

⎜⎝⎛+=

αα K

Q

K

Qxz (3)

The solution also incorporates groundwater discharge, Q [L2/T], to the seawater

boundary and assumes that the aquifer is isotropic with hydraulic conductivity K and

infinitely thick. The term α is a linear expansion coefficient that relates salinity to

density and has a constant a value of 0.00225 [-] for a freshwater aquifer in contact

with the sea. The saltwedge geometry therefore depends on the value of Q/K, which

is the ratio of SGD to aquifer hydraulic conductivity. By applying Darcy’s Law, Q =

KBi, it can be seen that Q/K = Bi, where B is the saturated thickness of aquifer and i

is the hydraulic gradient.

In the aquifer adjacent to Cockburn Sound the saltwedge extends up to 2,000 m

inshore, indicating that the value of Q/K is relatively small; a value of around .0025

is predicted from Figure 54. For K in the range 500-2,000 m.d-1, Glover’s abrupt

interface solution predicts SGD (Q) in the range 1.25-5 m3.d-1.m-1.



Figure 54: Abrupt saltwater interface profiles from the solution of Glover (1959)



4.6 SUMMARY

Conceptual models of SGD into Cockburn Sound based on the available regional-

scale hydrogeological information indicate that groundwater is likely to discharge

both along the shoreline and offshore from the coast. SGD along the shoreline has

been detected by field measurements (e.g. Section 3.1.3) and it is probable that

groundwater also discharges tens to hundreds of metres offshore via high-

conductivity zones in Tamala Limestone (e.g. Section 3.1.8). Accurate patterns of

SGD at local scales of interest cannot be determined because they depend upon

detailed, local-scale hydrogeology, which is not described across the study area. In

particular, there is virtually no offshore hydrogeological data in Cockburn Sound.

Direct measurements of SGD using benthic flux meters are not useful for

aggregating estimates of total SGD. Benthic flux measurements are strongly

influenced by seawater exchange between the coastal sediments and ocean in

response to sea level variation. In addition, field measurements indicate that the

distribution of SGD is patchy at spatial scales of less than ten metres. To properly

describe the distribution of SGD across the study area and aggregate volumetric

estimates of SGD would require an impracticable number of measurements.

However, BFM’s that measure SGD fluxes over short time scales of seconds to

minutes are very useful for understanding the SGD flux transfer mechanisms at the

sediment-water interface.

Based on groundwater flow modelling and aquifer water-balance considerations,

total average SGD across the coastline adjacent to Cockburn Sound is estimated to

be 2.5-4.8 ± 0.9 m3.d-1.m-1. The range, 2.5-4.8 m3.d-1.m-1, is derived from low and

high estimates of average groundwater recharge; the error, ± 0.9 m3.d-1.m-1, is

calculated based on arbitrary ±20% errors in pumping and groundwater inflow across

boundaries. Calibrated values of hydraulic conductivity from the flow modelling,

and the simulated rates of SGD, imply a saltwater-freshwater interface position that

is consistent with field observations.



5 QUANTIFYING NUTRIENT DISCHARGE INTO COCKBURN

SOUND

5.1 MECHANISMS FOR NUTRIENT DISCHARGE

Nutrients in the inshore aquifer can enter Cockburn Sound via two mechanisms:

• SGD of terrestrial groundwater, and

• Aquifer ‘flushing’ by seawater

The first case represents simple advection by groundwater flow of dissolved, mobile

nutrients that enter the aquifer between the coast and top of the regional flow system.

The un-utilised mobile fraction will eventually discharge into Cockburn Sound along

with the regional groundwater flow. Nutrient transport pathways and travel times

along these pathways—from source to sea—are determined by the spatial and

temporal patterns of groundwater recharge and flow. Nutrients are mobilized by

local recharge and transported toward the coast by regional groundwater flow, which

is dominated by channelised flows through high-conductivity pathways. Nutrients

that are mobilized into more active, or faster flowing parts of the regional flow

system will reach Cockburn Sound more quickly than nutrient that are mobilized into

relative inactive parts of the flow system. Thus, transport times are non-trivial to

calculate since SGD estimates based on the inshore water balance do not account for

preferred flow paths through the Tamala Limestone. In addition, the scale of aquifer

heterogeneity and its relation to the scale of contaminant sources inshore from

Cockburn Sound has not been investigated and is not known.

The second mechanism for groundwater nutrient discharge into Cockburn Sound has

not been investigated. It is probable that seawater movement into and out of coastal

sediments plays a role in transporting nutrients from subsurface to submarine

environments. Seawater entering the aquifer during periods of rising sea level can

contact and mix with nutrient enriched groundwater. Some of this water will drain

back to the ocean during a subsequent period of falling sea level and transport

nutrients with it. The longer the period of sea level variation, the further into the

aquifer this effect might propagate. The effect may be important in locations where

the aquifer beneath a contaminant source is regularly flushed by saltwater flows.



5.2 NITROGEN CONCENTRATION IN GROUNDWATER

Table 20 and Figure 55 present summaries of groundwater nutrient data inshore from

Cockburn Sound. Nitrogen is the main nutrient of concern and has been monitored

more intensely than phosphorus. Measured phosphorus concentrations in

groundwater are typically several orders of magnitude less than nitrogen. Significant

point source emissions of phosphorus in groundwater have not been detected to date.

Figure 55 is a summary of last-recorded, total nitrogen concentrations in

approximately 200 bores across the study area. This information was collated from

several sources, which are listed on Figure 55, and spans the ten-year period from

1991 to 2000. Bores that were sampled by Appleyard (1994) to estimate nutrient

discharge into Cockburn Sound are labeled using the notation from that study.

Because the data are not contemporaneous, they provide a mostly qualitative

impression of past and recent distributions of nitrogen in groundwater. Based on the

calculated SGD rates, groundwater is expected to travel hundreds of metres through

the aquifer each year and, therefore, quantitative analysis of nutrient data that was

collected in different years is not appropriate.

Two prominent and well-known areas of nitrogen contamination of groundwater are

indicated in Figure 55: (a) inshore from Northern Harbour, and (b) south of James

Point. PPK (1999) investigated two nutrient-enriched groundwater plumes

discharging into Northern Harbour that were found to emanate from the sludge

drying beds at Woodman Point Wastewater Treatment Plant, and from a wastewater

injection bore operated under license by Weston Bioproducts. Both operations were

recently decommissioned and remediation options for the plumes were investigated

by PPK (2000). Recovery of the plumes by groundwater pumping commenced in

December 2000 and is currently ongoing.

Nutrient enrichment of groundwater south of James Point is the direct result of

contamination by industry. CSPB produces nitrogen and phosphate fertilizers, and

ammonium sulphate is produced by WMC as a by-product of nickel refining.

Other potential sources of nitrogen in groundwater discharging into Cockburn Sound

include fertilizers applied for market gardening, horticulture and domestic lawns, and

leachate from septic tanks in residential areas.



Table 20: Summary of nutrient concentrations in groundwater in Cockburn Sound area and related sites

Reference Nutrient Concentration Location and Method

Bodard (1991) NO3 = 1.2, 2.5 (sand) NO3 = 10, 20 (limestone)

BP Refinery; representative values Bore screens: 2-14 mBGL sand 16-35 mBGL limestone

Davidson (1995) NO3 > 1 Swan Coastal Plain; groundwater samples from bores in the superficial aquifer

Appleyard (1994) N-NO3 = <0.2 – 160 (sand) N-NH3 = <0.02 - 394 (sand) N-total = 0.9 - 394 (sand) P-total = <0.01 - 0.74 (sand) N-NO3 = <0.2 – 1.6 (limestone) N-NH4 = 0.2 - 383 (limestone) N-total = 1 - 383 (limestone) P-total = 0.02 - 0.47 (limestone)

Cockburn Sound; groundwater samples from 39 bores in the superficial aquifer

PPK (1999) N-total = 8 Inshore from Northern Harbour beneath market garden area

PPK (1999) N-NO3 = 0.5 - 25 (WPMC5) N-NO3 = 26 (JBMB1) N-total = 0.5 – 25 (15 mean)

Woodman Point Waste Water Treatment Plant plume into Northern Harbour, bores WPMC5 and JBMB1

PPK (1999) N-NH3 = 22 - 252 (OB7) N-NH3 = 21 - 62 (JBMB7) N-total = 22 – 279 (128 mean)

Weston Bioproducts plume into Northern Harbour, bores OB7 and JBMB7

PPK (2000) N-NO3 = <0.05 – 64.5 N-NH4 = <0.1 – 210 N-DIN = 0.6 – 210

Inshore from Northern Harbour; groundwater samples from 28 bores

This Report N-NO3 = <0.0 – 21.0 N-NH4 = <0.1 – 130

Inshore from Northern Harbour; groundwater samples from 22 bores (see Table 7.

All concentrations are mg.l-1



�

��

�� !��""#�$$��!�%�&��""'��(�� !�)��*+,��(��(�� !�)��-$��(��(�� !�)��,�-$��(�� !�)��.�/��(�� !�)��,0*��(��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

��

*�,1��2�+&��

,�-$

*��!��$��**.$

*��-�� !��

��

��

��

��

��

��

��

��

��

��

��

.��!/��1��&��-��3��

��4��

��4��4��5��

��&��-��!0��-��!�.��0��!��.��!��/�� ! ��

��

��

��

��

��

��

��

��

��

��

��

��

Figure 55: Last-recorded, total nitrogen concentrations in groundwater [mg.l-1]; data spans the ten-year period from 1991 to 2000; groundwater bores sampled by Appleyard (1994) are labeled using notation from that study



5.3 SIGNIFICANCE OF MEASURED CONCENTRATIONS

It is difficult to assess the significance of groundwater nutrient concentrations in

isolation from an assessment of the marine, near-shore nutrient budget. The

ecological and environmental implication of nutrient-enriched groundwater

discharge into Cockburn Sound depends not only on the nutrient concentrations in

groundwater but also on the rate of groundwater discharge (i.e. the nutrient mass

flux), the relative sizes of other nutrient inputs, the sizes of the extant nutrient

‘pools’ in the receiving environment, and the dynamics of nutrient cycling. Such an

assessment is beyond the scope of this report.

As a guide, Table 21 presents default trigger values of nutrients and pH for slightly

disturbed marine ecosystems in south-west Australia. The values are taken from the

“Australian and New Zealand Guidelines for Fresh and Marine Water Quality”

(ANZECC, 2000). Default trigger values represent default concentrations that

should trigger further investigation to determine whether there is a genuine risk to

the ecosystem. Note that trigger concentrations apply to the receiving waters; they

are not discharge water-quality guidelines.

Table 21: Default trigger values for physical and chemical stressors for south-west Australia for slightly disturbed inshore marine ecosystems (ANZECC, 2000); inshore waters are defined as coastal lagoons (excluding estuaries) and embayments and waters less than 20-m depth

Water Quality Measure Default Trigger Value

Total nitrogen (N-total) 0.23 mg.l-1 Oxides of nitrogen (N-NO3, N-NO2) 0.005 mg.l-1 Ammonium (N-NH4) 0.005 mg.l-1 Total phosphorus (P-total) 0.02 mg.l-1 pH 8.0-8.4

1. This table is adapted from Table 3.3.6 of ANECC (2000) 2. Guideline trigger values represent bioavailable concentrations or

unacceptable levels of contamination that trigger the incorporation of additional information or further investigation to determine whether a real risk to the ecosystem exists.

The Water and Rivers commission’s Water Quality Protection Note for discharge of

groundwater from aquifer dewatering operations (WRC, 1999) provides a table of

water-quality protection criteria for receiving water bodies. The relevant water

quality indicators and criterion are reproduced in Table 22. For the same reasons as

discussed above, the criterion in Table 22 relate to acceptable levels of perturbation



of background conditions in the receiving water body rather than absolute

concentrations in the discharge water.

Table 22: Receiving water quality protection criteria for discharge of groundwater from aquifer dewatering (WRC, 1999)

Indicator Criterion

Nutrients

Water discharges should not add nutrients (e.g. phosphorus, nitrogen) in quantities sufficient to cause excessive or nuisance algal growth in the receiving water. The water discharged should not cause the seasonal background nutrient levels in the receiving water body to increase by more than 10%.

pH Water discharged should not cause the receiving water’s seasonal background pH to vary by more than ± 0.2 units.

1. This table is adapted from Table 2 of WRC (1999)

Perhaps a more useful approach is to compare the measured nutrient concentrations

in submarine porewater to typical nutrient concentrations in shallow groundwater at

other locations on the Swan Coastal Plain. Nitrogen is usually present in

groundwater as nitrate (NO3) and levels in urbanized areas and intensive horticultural

areas can be relatively high due to leachate from septic tanks and nitrogen fertilizers.

Nitrate levels within urbanized and cultivated areas vary depending on the type of

land use but commonly exceed 20 mg.l-1 (Davidson, 1995). Elsewhere, nitrate

concentrations in shallow groundwater are typically less than 1 mg.l-1.

Background ammonia concentrations in groundwater are normally small (less than

nitrate) because the ammonium is oxidized to nitrate before reaching the water table.

In contrast, nutrient contaminated porewater samples from Cockburn Sound contain

high concentrations of ammonia (10’s-100’s of milligrams per litre) and relatively

small nitrate concentrations. This is indicative of groundwater that has been directly

polluted by ammonia-based contaminants.

As a general guide for the metropolitan area, a total nitrogen concentration in

groundwater greater than 5 mg.l-1 is indicative of nutrient enrichment due to human

activities. Depending on the rate of discharge, and the characteristics of the receiving

water body, discharge of nutrient enriched groundwater may, or may not, have the

potential to impact on the aquatic ecosystem.



5.4 PREVIOUS ESTIMATES OF NUTRIENT DISCHARGE

Estimates of nitrogen discharge into Cockburn Sound by SGD during the period

1978 to 2000 have varied approximately in the range 490 to 1240 kg.d-1 (180 to

450 tonnes.yr-1). Previous estimates are listed in Table 23. Most estimates prior to

2000 were in the order of 900 kg.d-1 (330 tonnes.yr-1) and the latest estimate by D.A.

Lord (2000) was approximately 600 kg.d-1 (220 tonnes.yr-1).

Appleyard (1994) estimated that the phosphorus load into Cockburn Sound at that

time was approximately 6 kg.d-1 (2 tonnes.yr-1), which is several orders of magnitude

less than for nitrogen.

Despite the considerable scope for errors in the above estimates, uncertainty is not

usually addressed. Both the calculation of SGD rates (e.g. Section 4.2) and spatial

interpolation of groundwater nutrient data can introduce large errors in estimated

values of nutrient mass flux. For example, a simple doubling of hydraulic

conductivity has the effect of doubling both the SGD rate and associated nutrient

load. Further discussion on this topic is presented in Section 5.7.



Table 23: Estimates of nutrient discharge into Cockburn Sound via SGD

Reference Nutrient Discharge Location and Method Assumed Values

Appleyard (1990) in Appleyard (1994)

N-flux = 547 kg.d-1 P-flux = 11 kg.d-1

Coastline between Cape Peron and Fremantle

-

Mackie Martin (1992) in Appleyard (1994)

N-flux = 959 kg.d-1 As above -

Appleyard (1994) N-flux = 909 kg.d-1 (superficial) N-flux = 159 kg.d-1 (sand) N-flux = 751 kg.d-1 (limestone) N-flux = 0.0001-0.1 kg.d-1.m-1 (sand) N-flux = 0.0005-0.7 kg.d-1.m-1 (limestone) P-flux = 5.5 kg.d-1 (superficial) P-flux = 2.7 kg.d-1 (sand) P-flux = 2.7 kg.d-1 (limestone)

Cockburn Sound, flownet coupled with groundwater nutrient concentrations from 39 bores

k = 10-50, i = 0.001-0.003

DEP (1996) N-flux = 940 kg.d-1 (superficial)

Cockburn Sound -

PPK (2000) N-flux as DIN = 74 kg.d-1 (superficial)

Weston Bioproducts plume into Northern Harbour

PPK (2000) N-flux as DIN 34 kg.d-1 (superficial)

Woodman Point Waste Water Treatment Plant plume into Northern Harbour

PPK (2000) N-flux as DIN = 50 kg.d-1 (superficial)

Regional (market garden) inputs into Northern Harbour; excluding plumes

Measured nutrient concentrations in groundwater from 28 bores

D.A. Lord (2001) N-flux 1978 = 493 kg.d-1 (superficial) N-flux 1990 = 1243 kg.d-1 (superficial) N-flux 2000 = 600 kg.d-1 (superficial)

Cockburn Sound -

Symbols and Units k: saturated hydraulic conductivity [m.d-1] i: hydraulic gradient [1]



5.5 ESTIMATION OF NUTRIENT DISCHARGE BASED ON SGD RATES FROM THE

INSHORE WATER BALANCE MODEL AND MEASURED NUTRIENT

CONCENTRATIONS IN SUBMARINE POREWATER

This sections presents estimates of nutrient mass flux into Cockburn Sound based on

SGD rates estimated from the inshore water balance model (Section 4.4) and

groundwater nutrient concentrations derived from the submarine porewater survey

along Cockburn Sound foreshore (Section 3.4).

5.5.1 METHOD

An initial task was to assign groundwater nutrient concentrations to each 100-m

section of coastline corresponding to the 100-m-wide coastal boundary cells of the

model. This was carried out for the section of coastline from Woodman Point to the

southern boundary of the model (i.e. for 137 boundary cells). The estimates of

groundwater nutrient concentrations developed in Section 3.4 were spatially

interpolated. An average value (arithmetic mean) was assigned where more than one

porewater sample had been collected within the section of coastline spanned by a

particular boundary cell. For cells in which no porewater sample had been collected,

a groundwater concentration was assigned subjectively based on the nearest

estimated concentrations. Note that the porewater sampling density was increased in

areas of known groundwater nutrient contamination (Figure 41). Note, also, that

some sections of the coast could not be accessed for porewater sampling (Figure 41).

The total nutrient mass flux from each 100-m section of coastline was calculated by

multiplying the model-simulated values of SGD [L3.T-1] by the assigned

groundwater nutrient concentrations in each coastal boundary cell

gQ NSGDN ×= (4)

where: NQ is the nutrient mass flux in groundwater [M.T-1] and Ng is the nutrient

concentration in groundwater [M.L-3]. Therefore, the total nutrient load NT [M] from

the modelled-section of coastline during time period ∆t is

∑=

∆=n

iQiT NtN

1

(5)

where n is the number of coastal cells.



5.5.2 RESULTS

Figure 56 and Figure 57 depict model-simulated SGD for the low-recharge and high-

recharge scenarios, respectively. Figure 58 and Figure 59 show the estimated

nitrogen loads into Cockburn Sound for these estimates of SGD, and Figure 60 and

Figure 61 depict the corresponding phosphorus loads. Solid bars indicate nutrient

loads estimated from measured porewater concentrations and hatched bars indicate

interpolated values. Total nitrogen discharge into Cockburn Sound was estimated to

be 394 kg.d-1 (144 tonnes.yr-1) for the low-recharge scenario and 836 kg.d-1

(305 tonnes.yr-1) for the high-recharge scenario. Because total SGD for the high-

recharge scenario was approximately twice as large as total SGD for the low-

recharge scenario, the estimated total nitrogen load was also approximately doubled.

A break down of nitrogen loads along nominated sections of the coastline is

presented in Table 24. As identified by Appleyard (1994), a large portion of the total

nutrient load (approximately 75% in the current estimate) is discharged from the

coastline south of James Point. Approximately 10% of estimated nitrogen discharge

was from the short section of shoreline north of James Point and around 7% was

discharged from the area adjacent to the Naval Base caravan Park.

Table 24: Nitrogen loads for nominated sections of coastline

Nitrogen Load

Scenario Northern Harbour

[tonnes.yr-1]

Naval Base [tonnes.yr-1]

James Point North

[tonnes.yr-1]

James Point South

[tonnes.yr-1]

Total [tonnes.yr-1]

Low-recharge 6 (4%) 10 (7%) 15 (10%) 106 (73%) 144 High-recharge 10 (3%) 20 (7%) 34 (11%) 224 (74%) 305

1. Northern Harbour: 1.7 km length (6443650-6441950 mN AMG84) 2. Naval Base: 2.0 km length (6439950-6437950 mN AMG84) 3. James Point North: 1.0 km length (6435450-6434450 mN AMG84) 4. James Point South: 3.0 km length (6433950-6430950 mN AMG84)

Nutrient enriched porewater was not detected in Northern Harbour, with this section

of coastline contributing only 6-10 tonnes.yr-1 (4%) of the total nitrogen load. This

compares to estimates of around 70 tonnes.yr-1 (30%) in Appleyard (1994) and

approximately 58 tonnes.yr-1 in PPK (2000). Thus, the Northern Harbour plumes

probably discharge further offshore, beyond the shoreline strip in which the

porewater samples were collected.



Figure 56: Model simulated SGD for low-recharge scenario



Figure 57: Model simulated SGD for high-recharge scenario



Figure 58: Estimated nitrogen load into Cockburn Sound based on model simulated SGD (low-recharge scenario) and groundwater nutrient concentrations estimates from porewater sampling; solid bars indicate measured concentrations; hatched bars indicate interpolated concentrations



Figure 59: Estimated nitrogen load into Cockburn Sound based on model simulated SGD (high-recharge scenario) and groundwater nutrient concentrations estimates from porewater sampling; solid bars indicate measured concentrations; hatched bars indicate interpolated concentrations



Figure 60: Estimated phosphorus load into Cockburn Sound based on model simulated SGD (low-recharge scenario) and groundwater nutrient concentrations estimates from porewater sampling; solid bars indicate measured concentrations; hatched bars indicate interpolated concentrations



Figure 61: Estimated phosphorus load into Cockburn Sound based on model simulated SGD (high-recharge scenario) and groundwater nutrient concentrations estimates from porewater sampling; solid bars indicate measured concentrations; hatched bars indicate interpolated concentrations



5.6 ESTIMATION OF NUTRIENT DISCHARGE BASED ON SGD RATES FROM THE

INSHORE WATER BALANCE MODEL AND GROUNDWATER NUTRIENT

CONCENTRATIONS MEASURED BY APPLEYARD (1994)

Appleyard (1994) measured contemporaneous nutrient concentrations in 39 bores

along Cockburn Sound foreshore (Figure 55). Thirty-one of these bores were along

the section of coastline considered above; 17 in the Safety Bay Sand and 14 in the

Tamala Limestone. Here, nutrient discharge into Cockburn Sound is estimated based

on SGD rates from the inshore water balance model (Section 4.4) and the

groundwater nutrient concentrations measured by Appleyard (1994).

5.6.1 METHOD

The following methodology was used:

1. Nutrient discharges from Safety Bay Sand and Tamala Limestone were estimated

separately and then summed to calculate the total nutrient load.

2. It was assumed that 85% of total SGD was from Tamala Limestone and 15%

from Safety Bay Sand. This rough estimate is subject to a large uncertainty.

There are several difficulties with this approach, which are discussed in more

detail in Section 5.7.

3. The nutrient concentrations measured in groundwater bores were assigned

uniformly throughout the aquifer within the bores’ discharge widths (length of

coastline); these were chosen to match those used by Appleyard (1994).

Groundwater concentrations in Safety Bay Sand were based on the 17 shallow

bores installed into sand and concentrations in Tamala Limestone were based on

the 14 bores installed into limestone.

4. Nutrient discharge from each 100-m section of coast was calculated as follows

)85.0()15.0( TLSBSQ NSGDNSGDN ××+××= (6)

where NSBS and NTL are the assigned groundwater nutrient concentrations [M.L-3]

in Safety Bay Sand and Tamala Limestone, respectively.



5.6.2 RESULTS

Figure 62 and Figure 63 depict the estimated nitrogen loads into Cockburn Sound

and Figure 64 and Figure 65 show the corresponding phosphorus loads. The red bar

on each plot near to Northern Harbour indicates nitrogen and phosphorus loads that

were calculated in a previous study by Mackie-Martin PPK (1993). These values

were adopted and used in the analysis by Appleyard (1994) and are included here for

consistency. They are estimates of nutrient discharge into Northern Harbour from

Weston Bioproduct’s (formerly NB Love Starches) wastewater injection operation.

Total nitrogen discharge into Cockburn Sound was estimated to be 868 kg.d-1

(317 tonnes.yr-1) for the low-recharge scenario and 1,603 kg.d-1 (585 tonnes.yr-1) for

the high-recharge scenario. A break down of nitrogen loads along the nominated

sections of the coastline, as above, is presented in Table 25. Approximately 75-80%

of the total nitrogen load was estimated to discharge from the area of contaminated

aquifer located south of James Point; however, the absolute values of nutrient loads

were approximately double those calculated based on nutrient concentrations in

porewater.

Insignificant discharges of nutrients occurred north of James Point North and in the

area adjacent to the Naval Base caravan Park because elevated nitrogen levels in

these areas were not detected in the groundwater bore that were sampled.

Table 25: Nitrogen loads for nominated sections of coastline

Nitrogen Load

Scenario Northern Harbour

[tonnes.yr-1]

Naval Base [tonnes.yr-1]

James Point North

[tonnes.yr-1]

James Point South

[tonnes.yr-1]

Total [tonnes.yr-1]

Low-recharge 76 (24%) 1 (<1%) 6 (1%) 234 (74%) 314 (99%) High-recharge 84 (14%) 2 (<1%) 6 (1%) 486 (83%) 578 (99%)

1. Northern Harbour: 1.7 km length (6443650-6441950 mN AMG84) 2. Naval Base: 2.0 km length (6439950-6437950 mN AMG84) 3. James Point North: 1.0 km length (6435450-6434450 mN AMG84) 4. James Point South: 3.0 km length (6433950-6430950 mN AMG84)

Between 76 and 84 tonnes.yr-1 (14 to 24%) of the total nitrogen load was discharged

into Northern Harbour. Of this, 67 tonnes.yr-1 came from the adopted estimate of

nutrient discharge from Weston Bioproducts’ wastewater injection plume (Mackie-

Martin PPK, 1993).



Figure 62: Estimated nitrogen load into Cockburn Sound based on model simulated SGD (low-recharge scenario) and groundwater nutrient concentrations measured by Appleyard (1994)



Figure 63: Estimated nitrogen load into Cockburn Sound based on model simulated SGD (high-recharge scenario) and groundwater nutrient concentrations measured by Appleyard (1994)



Figure 64: Estimated phosphorus load into Cockburn Sound based on model simulated SGD (low-recharge scenario) and groundwater nutrient concentrations measured by Appleyard (1994)



Figure 65: Estimated phosphorus load into Cockburn Sound based on model simulated SGD (high-recharge scenario) and groundwater nutrient concentrations measured by Appleyard (1994)



5.7 SYNTHESIS OF RESULTS

Table 26, and Figure 66 and Figure 67 present summaries of estimated total nitrogen

discharge into Cockburn Sound via SGD. Previous estimates are labeled according

to date (i.e. the year) of the estimate, and estimates from the present study and

Appleyard (1994) are labeled “a” to “h” according to which combination of SGD and

groundwater nitrogen concentrations was used to evaluate the mass flux.

To facilitate the comparison of results, the mean rate of groundwater discharge,

SGDmean, for each estimate was calculated as

L

QSGDmean = (7)

where Q is the total groundwater discharge [L3.T-1] along a stretch of coastline with

length L. The mean nitrogen concentration in groundwater, Nmean, was then

calculated according to

Q

NN Q

mean = (8)

where NQ is the corresponding estimate of total nitrogen discharge in groundwater

[M.T-1] along the same stretch of coastline. These calculations were performed to

highlight differences between estimates that were not obvious if only the final values

of nutrient loads were considered.

Estimates labeled “a” to “d” are those calculated in Sections 5.5 and 5.6 above. They

are based on the SGD values derived from the inshore water-balance model

(Section 4.4) and groundwater nitrogen concentrations determined from the

submarine porewater survey (Section 3.4). Estimates “e” and “f” are from Appleyard

(1994), and “g” and “h” are worst-case and best-case estimates, which provide limits

to the analysis.

Figure 67 is a “bubble plot” that depicts the magnitudes of total nitrogen loads

according to the mean values of SGD and groundwater nitrogen concentration used

to calculate them. It illustrates the range of uncertainty in both quantities and

emphasises the fact that similar estimates of nutrient load have been derived from

dissimilar estimates of SGD and nutrient concentrations. Compare estimates “b” and



“c” for example. In estimate “b”, SGDmean ≈ 6 m3.d-1.m-1, Nmean ≈ 11 mg.l-1 and NQ =

305 tonnes.yr-1. In estimate “c”, mean discharge is approximately half, SGDmean ≈

3 m3.d-1.m-1, but mean concentration is approximately doubled, Nmean ≈ 23 mg.l-1,

such that NQ = 317 tonnes.yr-1. Thus, the estimates are approximately equivalent but

for different reasons.

Overall, the considered estimates of SGDmean and Nmean both vary by a factor of

approximately 3. This results in a potential factor of 9 (i.e single order of

magnitude) variation in NQ, as represented by estimates “g” and “h”. This type of

simple sensitivity analysis has not been considered in previous studies. In some

instances, the groundwater flow rates and concentrations that were used to estimate

nutrient mass flux were not reported.

180

454

350 343

219

144

305 317

585

230

430

965

108

0

200

400

600

800

1000

1200

1978 1990 1992 1996 2000 a b c d e f g h Estimate

N Q [tonnes.yr -1 ]

Previous Estimates This Study and Appleyard (1994)

Figure 66: Estimates of total nitrogen discharge into Cockburn Sound via groundwater



Table 26: Estimates of total nitrogen discharge into Cockburn Sound via groundwater, and equivalent average values of SGD and nitrogen concentrations in groundwater

Label NQ

[tonnes.yr-1] L [km]

SGDmean [m3.d-1.m-1]

Nmean [mg.l-1]

SGD Estimate Nitrogen Concentration

Estimate

a 144 13.7 2.8 10.3 This study – low recharge This study - porewater b 305 13.7 5.5 11.1 This study – high recharge This study - porewater c 317 13.7 2.8 22.6 This study – low recharge Appleyard (1994) d 585 13.7 5.5 21.3 This study – high recharge Appleyard (1994) e 230 16.0 2.1 18.8 Appleyard (1994) Appleyard (1994) f 430 16.0 2.1 35.1 Appleyard (1994) Appleyard (1994) g 965 13.7 5.5 35.1 This study – high recharge Appleyard (1994) h 108 13.7 2.1 10.3 Appleyard (1994) This study - porewater

1978 180 - - - D.A. Lord (2000) D.A. Lord (2000) 1990 454 - - - D.A. Lord (2000) D.A. Lord (2000) 1992 350 - - - Mackie Martin (1992) Mackie Martin (1992) 1996 343 - - - DEP (1996) DEP (1996) 2000 219 - - - D.A. Lord (2000) D.A. Lord (2000)

Min. 2.1 10.3 Max. 5.5 35.1

1. Yellow-highlighted values were used to calculate blue-highlighted values

144 305

317 585

230

430

108

965

0

5

10

15

20

25

30

35

40

45

1 2 3 4 5 6 7 SGD mean [m 3 .d -1 .m -1 ]

N mean [mg.l -1 ]

a b c d e f g h

Total Nitrogen Discharge

[tonnes.yr -1 ]

Figure 67: Estimates of total nitrogen discharge into Cockburn Sound via groundwater with data plotted according to equivalent average values of SGDs and nitrogen concentrations in groundwater



5.7.1 UNCERTAINTY

Uncertainty in the above estimates of nutrient mass flux stem from the uncertainties

in estimating SGD and determining correct spatial distributions of nutrient

concentrations in the aquifer.

To estimate SGD, relatively accurate determinations of hydraulic gradient can be

made; however, the hydraulic conductivity of sediments is uncertain and varies

typically over several orders of magnitude. This introduces a same order of

magnitude uncertainty in SGD and nutrient flux estimates. By taking into

consideration the inshore groundwater balance, and ensuring that SGD rates are

consistent with realistic estimates of recharge and abstraction from the aquifer, the

level of uncertainty can be minimised. This approach was used in the present study;

however, the SGD rates that were calculated based on low and high estimates of

groundwater recharge still varied by a factor of at least 2.

Additional uncertainty was introduced during the process of deciding which value(s)

of nutrient concentration were most representative of nutrient distributions in

groundwater. For example, results from Northern Harbour highlight a potential

inadequacy of the submarine porewater survey. Because porewater samples were

collected within 10 m of the shoreline, the groundwater component of these samples

was most likely dominated by shallow groundwater from the upper part of the

aquifer (see conceptual models of SGD in Section 4.1). Nutrient concentrations in

shallow groundwater may or may not be representative of nutrient concentrations in

deeper groundwater discharging further from the shore. The nutrient plume from

Woodman Point Wastewater Treatment Plant, which is detected in groundwater

monitoring bores, was not detected in the porewater samples collected along the

shoreline of Northern Harbour.

In general, unless a porewater survey is extended offshore and over the full width of

SGD—assuming this is known—then there is potential that a single porewater

sample collected immediately along the shoreline could misrepresent the average

porewater concentration at that location. To resolve this uncertainty would require a

precise knowledge of where groundwater discharges so that porewater samples could

be collected at these locations. Nevertheless, patterns of groundwater discharge are



complex and strongly influenced by the coastal hydrogeology, which is poorly

defined at the scale at which porewater measurements are made.

Despite these limitations, the porewater survey had several advantages over previous

attempts to estimate nutrient mass flux based on concentrations measured in inshore

groundwater bores. Porewater samples were collected within 0.5 m of the seabed

and it is reasonable to assume that measured nutrient concentrations are an accurate

reflection of the concentrations in groundwater discharging at the measurement

locations. In addition, porewater samples were collected at more locations along the

coastline—compared to previous sampling of groundwater bores—and the density of

sampling was increased at locations where significant nutrient contamination was

detected; for example, north and south of James Point where porewater samples were

collected at 20-m intervals. Where possible, porewater samples were also collected

in offshore transects.

The main difficulty in estimating nutrient mass flux from nutrient concentrations

measured in groundwater bores is the interpolation within the aquifer of this data.

The spatial resolution of measured concentrations is typically poor. Deciding how

representative of the surrounding aquifer these measurements are, and how the data

should be interpolated spatially, is non-trivial and largely subjective. This is

compounded by the fact that it is rarely obvious which value of SGD should be

coupled with the measured or interpolated nutrient concentration at a particular

location in the aquifer.

For example, the available hydrogeological data indicates that most groundwater in

the Safety Bay Sand drains vertically into high-permeability Tamala Limestone and

then flows toward the coast via preferential flow paths. In this situation, it is unclear

whether nutrient concentrations measured in shallow groundwater bores that are

screened within the Safety Bay Sand should be coupled with flow rates in the sand or

limestone. This issue has not been addresses in previous investigations, which

simply assume that all groundwater in the Safety Bay Sand flows laterally through

the sand and discharges at the coast (i.e. vertical draining of groundwater into the

underlying Limestone is ignored). Because the contrast in hydraulic conductivity

between the limestone and overlying sands is so large, absence of vertical drainage

into the limestone is highly unlikely except where the sand and limestone are

separated by clay aquitard that prevents vertical flow.



To improve our understanding of groundwater nutrient transport into Cockburn

Sound, annual SGD nutrient loads computed in this study and previous studies

should be compared with estimates of mass loadings at the source locations (e.g.

based on loss of product or leaching rates). In each case, the following question

should be considered: Is the estimated nutrient mass loading at the source location

consistent with associated estimates of nutrient mass discharge to Cockburn Sound

over time? For example, if the estimate of nutrient mass discharge for a particular

nutrient plume, over a twenty-year period, was found to be significantly greater than

the estimate of nutrient mass loading at the source, then this would indicate that the

SGD rates used to calculate the mass discharge loadings were possibly over

estimated. Alternatively, the groundwater nutrient concentrations may have been

underestimated. This highlights one advantage of the porewater survey, which

detects nutrient concentrations directly below the seabed. Porewater sampling

affords a reasonably high level of confidence in estimates of nutrient concentrations

at the location of SGD; therefore, any inconsistencies between nutrient mass loadings

at source and discharge locations are more likely to be the results of errors in the

estimated SGD rates.

Spatial variability in groundwater recharge and flow, control the transport pathways

and travel times for nutrients to move from source areas inshore from Cockburn

Sound to the ocean. Nutrient contamination along the foreshore strip of Cockburn

Sound is mobilized into the groundwater system by local groundwater recharge. The

aquifer beneath these areas is then ‘flushed’ by groundwater flow from further

inshore as it moves toward Cockburn Sound. The presence of preferred flow paths

mean that some areas of the aquifer are likely to contain less active—slower

flowing—areas of groundwater compared with other parts of aquifer containing

preferred flow paths. Nutrients that are mobilized into relatively inactive areas of the

flow system are likely to take longer to reach Cockburn Sound. In addition, these

areas of aquifer are likely to contain nutrients for longer periods following

contamination events since the flushing time will be longer. The reverse case applies

in areas where the groundwater flow system is relatively active.

5.7.2 COMPOSITE ESTIMATE

A composite estimate of nitrogen discharge into Cockburn Sound, based on the

following assumptions and logic, is presented in Table 27:



• Compared to SGD rates derived from the inshore water-balance model, the

groundwater discharge rates used by Appleyard (1994) now appear to be

relatively small. Therefore, the model-derived values of SGD were used to

estimate nutrient mass flux.

• It is likely that there have been significant changes in concentrations of nutrients

in groundwater since the study by Appleyard in 1994, and an updated set of

nutrient concentrations in groundwater bores along the Cockburn Sound coastline

is not available. Therefore, nutrient concentrations in groundwater that were

derived from the submarine porewater survey were used here. An advantage is

that previously undetected discharges of nutrient-enriched groundwater from

north of James Point and adjacent to the Naval Base caravan park are included.

A disadvantage is the absence of comparable data sets with which to evaluate the

uncertainty in these measurements.

• Groundwater nutrient concentrations derived from the porewater survey appear to

underestimate the true concentrations in groundwater inshore from Northern

Harbour. Therefore, nitrogen mass flux into Northern Harbour computed by

PPK (2000, Table 3.2, Scheme 4) is used here. The estimate by PPK allows for a

51% reduction in nitrogen loading to the harbour due to plume recovery

operations.

The resultant estimate of total nitrogen load to Cockburn Sound was 234

±88 tonnes.yr-1 along the length of coastline considered in this study.

Approximately 71% was contributed from south of James Point, 12% from Northern

Harbour, 11% from north of James Point and 6% from the shoreline adjacent to the

Naval Base caravan Park.

Table 27: Composite estimate of total nitrogen discharge into Cockburn Sound via groundwater

Northern Harbour1

[tonnes.yr-1]

Naval Base2 [tonnes.yr-1]

James Point North2

[tonnes.yr-1]

James Point South2

[tonnes.yr-1]

Total [tonnes.yr-1]

29±14 (12%)

15±5 (6%)

25±10 (11%)

165±59 (71%)

234±88

1 PPK (2000, Table 3.2) 2 This study, estimates “a” and “b”



REFERENCES

Appleyard, S. J., 1994, The flux of nitrogen and phosphorus from groundwater into Cockburn Sound, Perth Metropolitan Region, Geological Survey of Western Australia (unpublished), Hydrogeology Report 1994/39.

Bodard, J. M., 1991a, Assessment of the BP refinery site, Kwinana, Western Australia: Volume I Hydrological framework and site work recommendations, BHP Research, Melbourne Laboratories.

Bodard, J. M., 1991b, Assessment of the BP refinery site, Kwinana, Western Australia: Volume II Hydrological framework and site work recommendations, Figures and attachments, BHP Research, Melbourne Laboratories.

Corporate Environmental Consultancy, 2002, Hismelt (Operations) Pty Limited Commercial Hismelt Plant Kwinana Western Australia Public Environmental Review, April 2002.

D.A. Lord, 2001, Cockburn Sound Management Council, The state of Cockburn Sound: A pressure-state-response report, prepared by D.A. Lord & Associates Pty Ltd in association with PPK Environment and Infrastructure Pty Ltd, Report No. 01/187/1, June 2001.

Davidson, W. A., 1995, Hydrogeology and groundwater resources of the Perth Region Western Australia, Geological Survey of Western Australia, Bulletin 142.

Davis, G. B., Johnston, C. D. and Patterson, B. M., 1994, Deep drilling data, multiport installation specifications, groundwater chemistry and soil gas composition for the first bioremediation field trial at the Kwiana refinery site, Report to BP (Kwinana), Report No. 94/34, September 1994.

DEP, 1996, Southern Metropolitan coastal waters study (1991-1994) final report, Department of Environmental Protection Western Australia, Report 17, November 1996.

Department of Land Administration, 5m and 1m Topographic data (supplied by WRC).

Department of Minerals and Energy, Perth Metropolitan Region 1:50,000 Environmental Geology Series Fremantle, Part Sheets 2033 I and 2033 IV.

Department of Minerals and Energy, Perth Metropolitan Region 1:50,000 Environmental Geology Series Rockingham, Part Sheets 2033 II and 2033 III.

Department of Minerals and Energy, 1:50000 Environmental Geology digital data, (supplied by Lorraine)



Gerbaz, S., 1999, Investigation of saltwater intrusion at BP Refinery (Kwinana). A thesis presented for Degree of Bachelor of Engineering, The University of Western Austalia, October 1999.

Glover, R. E., 1959, The pattern of fresh-water flow in a coastal aquifer. Journal of Geophysical Resaerch 10: 1039-1043.

HGM (Halpern Glick Mansel), 1998, Investigation of water quality in the Jervoise Bay Northern Harbour; December 1997 – March 1998, report to Department of Commerce and Trade.

Hirschberg, K-J., 1990, Lake Coogee Area assessment of monitoring data 1981-1987, Geological Survey of Western Australia, Hydrogeology Report 1989/38 (unpublished?).

Krupa, S., Gefvert, C., Brock, J., Grall, C. and Krupa, A., 2000, Field report for the International Oceanographic Committee, Submarine groundwater intercomparison experiment, Cockburn Sound, Western Australia, November 25 - December 6, 2000 (unpublished).

Lee, D. R., 1977, A device for measuring seepage flux into lakes and estuaries, Limnology and Oceanography, 22, 140-147.

Nield Consulting, 1999, Modelling of the Superficial Aquifer in the Cockburn Groundwater Area (Draft), unpublished report to Water and Rivers Commission and Kwinana Industries Council.

Mackie-Martin PPK, 1993, NB Love Starches Consultative Environmental Review: Proposal to increase quantity of wastewater disposed of by shallow bore injection.

Passmore, J. R., 1970, Shallow coastal aquifers in the Rockingham district Western Australia, Water Research Foundation of Australia, Bulletin No. 18.

Patterson, B.M.; Grassi, M.E.; Davis, G.B.; Robertson, B. and McKinley, A.J., 2002, The use of polymer mats in series for sequential reactive barrier remediation of ammonium-contaminated groundwater: laboratory column evaluation, Environmental Science and Technology, 36, 3439-3445.

Patterson, B.; Smith, A.; Davis, G.; Robertson, B.; Hick, W.; Grassi M., Herne, D. and Turner J., 2002, Quantifying submarine groundwater discharge and demonstrating innovative clean-up to protect Cockburn Sound from nutrient discharge (Project No. WA9911), final report to Coast and Clean Seas National Heritage Trust, Centre for Groundwater Studies and CSIRO Land and Water, November 2002.

Paulsen, R.J., C.F. Smith, and T.-f. Wong, 1997, Development and evaluation of an ultrasonic groundwater seepage meter, in Geology of Long Island and Metropolitan New York, 88-97.



PPK Environment and Infrastructure, 1999, Remediation strategy for nitrogen-rich groundwater at Jervoise Bay, report prepared for Department of Commerce and Trade, 98K148A:PR2:4454:RevB, April 1999.

PPK Environment and Infrastructure, 2000, Jervoise Bay Project groundwater recovery plan, report prepared for Department of Commerce and Trade, 98L084A:PR2:5426, April 1999.

Rust PPK Environment and Infrastructure, 1996. Kwinana Works Groundwater Monitoring Manual, report prepared for Westfarmers CSBP Ltd., 98H065A:PR2:0639:RevA, May 1996.

Searle, D. J., Semeniuk, V. and Woods, P. J., 1988, Geomorphology, stratigraphy and Holocene history of the Rockingham-Becher Plain, South-western Australia, Journal of the Royal Society of Western Australia, 70(4), 89-109.

Smith, A.J. and Hick, W.P., 2001, Hydrogeology and aquifer tidal propagation in Cockburn Sound, Western Australia, CSIRO Land and Water Technical Report 06/01, February 2001.

Stieglitz, T., 2001a, Seismic Survey in Northern Harbour & Cockburn Sound December 2000, unpublished report to SCOR/LOICZ Working Group 112.

Stieglitz, T., 2001b, Ground conductivity transects in Northern Harbour, Cockburn Sound, December 2000, unpublished report to SCOR/LOICZ Working Group 112.

Stieglitz, T., Ridd, P. V. and Hollins, S., 2000, A Small Sensor for Detecting Animal Burrows and Monitoring Burrow Water Conductivity. Wetlands Ecology and Management, 8, 1-7.

Walker, N., 1994, Tidal influences on coastal groundwater levels. A thesis for the Degree of Bachelor of Engineering, The University of Western Australia, October 1994.

WAWA, 1993, Cockburn Groundwater Area Management Plan, June 1993, produced by Groundwater and Environment Branch, Water Authority of Western Australia, ISBN-0-7309-5260-6.

APPENDIX A Technical Report - January 2003


APPENDIX A: AQUIFER TIDAL PROPAGATION AT NORTHERN

HARBOUR

CSIRO Land and Water APPENDIX A


THE TIDAL METHOD

The tidal method (Ferris, 1951; Carr and Van Der Kamp, 1969; Townley, 1995) is a

simple technique for estimating aquifer diffusivity—the ratio of aquifer

transmissivity to the aquifer storage coefficient—based upon the response of the

aquifer to tidal forcing at a boundary. Similar to conventional pump test techniques,

in which the piezometric head response of an aquifer to artificial pumping is used as

a basis for estimating aquifer hydraulic properties, the tidal method takes advantage

of natural tidal forcing and obtains an estimate of diffusivity from the attenuation of

a tidal signal as it propagates into an aquifer. The tidal signal can be expressed as a

linear combination of sinusoidal terms or tidal constituents (Forman and Henry,

1989, p.109) that are differentially attenuated as they travel though the aquifer; the

degree and rate of attenuation depends upon the aquifer hydraulic properties. There

are two possible mechanism for tidal propagation: (1) flow of tidal water into and out

of an unconfined aquifer, (2) compression and expansion of a confined aquifer due to

the load of the incoming and outgoing tides. In an aquifer that has tidal forcing at a

lateral boundary, the attenuation of the tidal signal is normally described by the two

quantities tidal efficiency and lag.

Tidal efficiency, TE, is a normalised measure that relates the amplitude of head

fluctuations in the aquifer to the amplitude of fluctuations at the tidal boundary. In

general,

ak

akk F

RTE = (9)

where Fa is the amplitude of forcing at the tidal boundary, Ra is the amplitude of the

response at a point in the aquifer and k denotes the tidal constituent or frequency.

The tidal efficiency has a value between zero and one.

Lag is an inverse measure of the velocity of propagation of a tidal constituent as it

moves through the aquifer

kkk FRlag θθ −= (10)

where Fθ is the phase of tidal forcing and Rθ is the phase of the tidal response at a

point in the aquifer. Note that the slower the speed of propagation, the larger the lag.



Ferris (1951) presents analytical expressions that describe the tidal efficiency and lag

in a one-dimensional, semi-infinite and homogeneous aquifer with uniform

transmissivity (see also Carr and Van Der Kamp, 1969). Inverse solutions are

obtained for 1D-transient groundwater flow with a tidal or harmonic boundary

condition at x = 0.

t

h

T

S

x

h

∂∂

∂∂ =

2

2

(11)

( )θω HtHth a −= cos),0( (12)

Here, Ha is the amplitude of head fluctuations at the tidal boundary [L] and Hθ is the

phase measured in radians. Since (11) and (12) are both linear, the groundwater flow

problem they describe can be decomposed into a steady-state flow problem and a

harmonic flow problem that consists of one or more frequencies. The harmonic

solution is rearranged to yield the following expressions for tidal efficiency and lag

⎟⎟⎠

⎞⎜⎜⎝

⎛−=

kk TP

SxTE

exp

π (13)

T

SPxlag k

k 4π= (14)

The coordinate x represents distance from the tidal boundary, S is the aquifer storage

coefficient [1], T is aquifer transmissivity [L2T-1]—which is assumed uniform, Pk is

the period [T] of the tidal constituent k and ωk = 2π/Pk is the corresponding angular

frequency [T-1]. Since a particular tidal signal may contain several dominant

frequencies, e.g. semi-diurnal and diurnal constituents, several independent

estimations of diffusivity are possible from a single tidal signal. In general, longer-

period constituents will propagate further into the aquifer and yield estimates of

diffusivity that apply to larger regions of the aquifer.

From (13) and (14), efficiency-based and lag-based expressions for aquifer

diffusivity are straightforward to derive

( ) kk PTE

x

S

T2

2

ln

π= (15)



( )2

2

4 k

k

lag

Px

S

T

π= (16)

Equating (15) and (16) gives the non-dimensional relationship

kk

k TEP

lagln

2

1⎟⎠⎞

⎜⎝⎛−=

π (17)

where the left-hand term is a normalised lag. It follows that, on a semi-log plot, the

normalised lag and tidal efficiency are related by the straight line with slope equal to

–1/2π. The tidal efficiency decreases exponentially with increasing distance from the

tidal boundary, whereas the lag increases linearly.

FOURIER ANALYSIS

The term Fourier analysis refers to any data analysis technique that describes or

measures fluctuations in a time series by comparing them with sinusoids. Well

known methods include filtering, least squares regression on sinusoids and harmonic

analysis (Bloomfield, 1976; Chatfield, 1975; James, 1995). Fourier techniques are

commonly used to detect frequency components that are hidden within a 'noisy'

signal.

Harmonic analyses fits uniformly sampled time series data to a set of harmonic

frequencies; that is, a set of frequencies that are integer multiples of one another.

The fitted harmonics are normally in the frequency range Nνν ≤≤0 , where

tPN

N δν

2

11 == (18)

is the Nyquist frequency [T-1], PN is the Nyquist period [T] and δt is the data

sampling interval [T].

The Nyquist condition simply states that the smallest period that can be reasonably

fitted to the data is equal to twice the sampling interval. This corresponds to fitting,

at most, one local maxima or minima between each pair of data point in the time

series. From a practical point of view, the Nyquist conditions tells us that it is

unreasonable to fit, say, a semi-diurnal frequency (12 hour period) to data that has



only been sampled 24 hourly. The smallest period and highest frequency that should

be fitted to diurnal data are 48 hours and 0.5 cycles per day, respectively.

DISCRETE FAST FOURIER TRANSFORM (FFT)

The discrete FFT is a computationally efficient algorithm that enables regularly

sampled time series data to be fitted by a set of harmonic frequencies (Chatfield,

1975; Bloomfield, 1976; James, 1995). For a time series of n data values, the FFT

fits n/2 frequencies over the frequency range NN vvnv ≤≤/2 . The FFT of the time

series is the set of n/2 complex valued Fourier coefficients Φ(vj), where the moduli

|Φ(vj)| are equal to the amplitudes of the fitted frequencies and arguments the

arg(Φ(vj)) are equal to the phases.

The Inverse Fast Fourier Transform (IFFT) is a reverse operation that converts a set

of Fourier coefficients Φ(vj) back to the original time-series data y(ti). This

relationship is denoted symbolically as (James, 1995)

)()( ji vty Φ⇔ (19)

A measure of the energy or power at each FFT frequency is provided by the Power

Spectral Density (PSD) function

)( )()( *jjj vvvS ΦΦ= (20)

Here, Φ*(vj) denotes the complex conjugates of the Fourier coefficients. A large

value of S(vj) indicates more energy at frequency vj. This means that the dominant

frequencies in a time series are represented as larger spikes on a PSD plot.

FREMANTLE TIDE DATA

The tide station at Fremantle—maintained by the Department of Transport, Western

Australian—is the nearest tidal monitoring location to Cockburn Sound. Historic

tide data is available upon request to the Maritime Division at a minimum

continuous sampling interval of five minutes. Fremantle tide data (Figure 68) used

in this report are sampled at 10-minute intervals over the period 01/09/2000 to

16/01/2001.



Fremantle has a mixed tide that is characterised by diurnal (24-hour) and semi-

diurnal (12-hour) constituents. The K1† (period = 23.9345 hours) and O1‡ (period =

25.8193) diurnal constituents are the dominant frequencies (Department of

Transport, unpublished harmonic analysis). The semi-diurnal frequencies vanish for

part of each month, which makes them unsuited for use with the tidal method.

FREMANTLE TIDE DATA

Tide data was logged continuously at 10-minute intervals for the duration of the

SGD intercomparison experiment. Water levels were measured using a capacitive

water level probe (Dataflow Systems) and model 392 data recorder (Dataflow

Systems) suspended within a 50-milimetre PVC stilling well and mounted off a sea

wall at the intercomparison site. The same equipment was used to record water

levels—also at 10-minute intervals—in each of the five monitoring bores. Logger

internal clocks were synchronized using the same laptop computer so that the tide

and water levels in all bores were recorded at the same ten-minute intervals.

The tide data collected during the intercomparison is presented in Figure A2.

Because the tidal logger failed on 12 December and the length of tidal record

obtained to that point in time was too short, the K1 and O1 tidal constituents could

not be separated in a harmonic analysis. Department of Transport tide data for

Fremantle (Figure 69) was used to ‘repair’ the Cockburn Sound data set and extend

the length of the time series to match that obtained for the groundwater monitoring

bores.

Harmonic Analysis

All data sets presented in this report are standardised to a ten-minute sampling

interval. The corresponding 72 cycles per day Nyquist frequency is more than

adequate to detect the expected diurnal or 1 cycle per day tidal constituents. The

number of harmonics fitted in each analysis depends upon the number of data values

in the time series being considered. Since the discrete FFT fits n/2 harmonics to n

data points, a larger number of data points result in more frequencies and a higher

resolution PSD plot. When a standardised sampling interval is used, the total length

of the time series determines the number of data points and Fourier frequencies. All

† K1 is the lunisolar diurnal constituent; with O1 it expresses the effect of the Moon's declination ‡ O1 is the lunar diurnal constituent



data sets used in this report contained 4628 data points, corresponding to the period

from 25/11/2000 @ 17:10 to 27/12/2000 @ 20:20. This means that 2314

frequencies are fitted in the harmonic analyses presented.

Foreman and Henry (1989, Table A1) specify that the minimum total length of time

series required for separation of the K1 and O1 tidal constituents in a harmonic

analysis are 24 hours and 328 hours (13.67 days), respectively. The data sets used in

this report are approximately 32 days long, which satisfies the above crtieria.

Harmonic analyses are performed using Mathematica 3.0 (Wolfram Research, 1988-

1996) by first detrending the time series—that is, subtracting the mean data value

from each data value—and then implementing the intrinsic FFT function

“Fourier[list]”. The PSD function is then calculated as nvvvS jjj /)( )()( *ΦΦ= .

Figure 70 to Figure 75 depict water levels and the PSD functions for six of the seven

intercomparison monitoring bores; NH1a, NH2a, NH2b, NH3, NH4 and NH5. A

continuous record was not obtained in NH1b due to problems with the data logger.

The ‘spectral signatures' of these data sets are consistent with the tide data and the

K1 and O1 tidal constituents are clearly discernable. Peaks on the PSD plots at

lower frequencies ( 25.0≤jv ) correspond to the passage of weather systems,

associated fluctuations in barometric pressure and 'piling up' of water at the shoreline

due to prevailing onshore winds. These weather effects are non-stationary, however,

and are complicated by the fact that barometric pressure fluctuations affect the

aquifer storage directly through atmospheric loading, as well as indirectly through

tidal loading. They are therefore not suited for use with the tidal method.



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) Fremantle Tide Data

20 40 60 80 100 120

Frequency (cycles/day)

Va

lue

(b) Fremantle Power Spectral Density

0.5 1 1.5 2 2.5 3

0

20

40

60

80

100

120

O1

K1

Figure 68: Tidal data and Power Spectral Density for Fremantle for the period 1/9/2000 to 16/1/2001; sampling interval is 10 minutes



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) Cockburn Sound Tide Data

5 10 15 20 25 30 35

Loggerfailure

-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(b) Cockburn Sound ‘Repaired’ Tide Data

5 10 15 20 25 30 35

Figure 69: Cockburn Sound tidal data and ‘repaired’ Cockburn Sound tidal data for the period 25/11/2000 to 3/1/2001; sample interval is 10 minutes



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) NH1a Tide Data

0 5 10 15 20 25 30


Va

lue

(b) NH1a Power Spectral Density

0.5 1 1.5 2 2.5 3

O1

K1

0.5

1.0

1.5

2.0

2.5

3.0

0.0

Figure 70: Hydrograph and Power Spectral Density for groundwater monitoring bore NH1a for the period 25/11/2000 to 27/12/2000; sampling interval is 10 minutes



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) NH2a Tide Data

0 5 10 15 20 25 30


Va

lue

(b) NH2a Power Spectral Density

0.5 1 1.5 2 2.5 3

O1

K1

0.0

0.5

1.0

1.5

2.0

2.5

Figure 71: Hydrograph and Power Spectral Density for groundwater monitoring bore NH2a for the period 25/11/2000 to 27/12/2000; sampling interval is 10 minutes



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) NH2b Tide Data

0 5 10 15 20 25 30


Va

lue

(b) NH2b Power Spectral Density

0.5 1 1.5 2 2.5 3

O1

K1

0.0

0.5

1.0

1.5

2.0

2.5

Figure 72: Hydrograph and Power Spectral Density for groundwater monitoring bore NH2b for the period 25/11/2000 to 27/12/2000; sampling interval is 10 minutes



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) NH3 Tide Data

0 5 10 15 20 25 30


Va

lue

(b) NH3 Power Spectral Density

0.5 1 1.5 2 2.5 3

O1

K1

0

1

2

3

4

5

6

Figure 73: Hydrograph and Power Spectral Density for groundwater monitoring bore NH3 for the period 25/11/2000 to 27/12/2000; sampling interval is 10 minutes



-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) NH4 Tide Data

0 5 10 15 20 25 30


Va

lue


0.5 1 1.5 2 2.5 3

O1

K1

0

2

4

6

8

10

12




-0.4

-0.2

0.0

0.2

0.4

0.6

0.8

1.0

Time (days)

Ele

vation

(mA

HD

)

(a) NH5 Tide Data

0 5 10 15 20 25 30


Va

lue


0.5 1 1.5 2 2.5 3

O1

K1

0

5

10

15

20




AQUIFER TIDAL PROPAGATION

At each monitoring bore, applying the tidal method to the K1 and O1 tidal

constituents gives four estimates of aquifer diffusivity; each constituent yields an

efficiency-based and lag-based estimate of diffusivity. Results for the

Intercomparison monitoring bores are summarised in Tables A2 and A3.

Tidal efficiency and lag are calculated from FFT’s of the ocean tide data and

groundwater level responses in the monitoring bores. At each bore

ocean1

bore11 )(

)(

K

K

KTEνν

ΦΦ

= ocean1

bore11 )(

)(

O

O

OTEνν

ΦΦ

= (21)

and

⎟⎟⎠

⎞⎜⎜⎝

⎛ΦΦ

=ocean1

boren11 )(

)(arg

K

KKlag

νν

⎟⎟⎠

⎞⎜⎜⎝

⎛ΦΦ

=ocean1

bore11 )(

)(arg

O

OOlag

νν

(22)

Once tidal efficiency and lag are determined, diffusivity is calculated using equations

(15) and (16).

The validity of applying the ‘Ferris model’ to the Cockburn Sound data set is tested

by plotting normalised lag verses tidal efficiency on a semi-log graph (Figure 76).

The data are compared with each other and the theoretical relationship from equation

(17). Most notably, all the bores diverge from the theoretical relationship between

tidal efficiency and normalised lag that is predicted by the Ferris model, and there is

a clear separation of the deep and shallow bores.

Shallow Bores

There is evidence of a logarithmic relationship between tidal efficiency and

normalised lag in the shallow bores but the slope is clearly different to that of the

Ferris model (-1/2π). In general, the observed attenuation of tidal amplitude is

greater than the model-predicted attenuation, for the same speed of propagation

through the aquifer. Correspondingly, the observed speed of propagation is faster

than the model-predicted speed of propagation, for the same attenuation of

amplitude. This difference is most likely a consequence of the modelling



assumptions and results in estimates of lag-based diffusivity that are consistently

larger than efficiency-based estimates.

Despite this deficiency of the model, the estimates of diffusivity determined from the

shallow bores fall within the range 3,600 to 11,800 m2.d-1, which is within a factor of

3.3. Assuming that the saturated depth of the sand units overlying the Tamala

Limestone is around 17 metres (based on the drilling data) and assuming a specific

yield of approximately 0.25, the estimated range in hydraulic conductivity of the

sands is 53 to 174 m.d-1, which is characteristic of medium to coarse sand (Fetter,

1994; Bouwer, 1978).

Deep Bores

Figure 76 indicates that the relationship between tidal efficiency and normalised lag

in the deep bores is significantly different compared to both the shallow bores and

the Ferris model. Again, this is likely to be a consequence of the model assumptions,

which do not account for the high-transmissivity Tamala Limestone underlying the

sand. The assumptions of 1-D flow and a homogeneous aquifer, in particular, are

poor representations of the true field conditions.

During the SCOR/LOICZ SGD Intercomparison, it was observed that some of the

submarine discharge being measured by benthic flux meters on the seabed had

seawater salinity. This is consistent with the fact that the ocean level was falling

during the experiment and indicates that salt water, previously forced into the aquifer

during a period of higher sea level, was draining back out of the aquifer. The period

of the intercomparison corresponds to the first ten days in Figure 70 to Figure 75 in

which the trend of falling sea level is clearly depicted. It is believed that there is very

active exchange of seawater between the ocean and Tamala Limestone, which is

manifest as ‘tidal pumping’.

The current hypothesis is that during periods of rising sea level, saltwater flows into

the Tamala Limestone, which has the effect of displacing both the saltwater interface

and free surface in the overlying sands upward. Downward displacement of the free

surface and saline interface occurs during periods of falling sea level. In effect, rapid

propagation of tidal oscillations through the limestone extends the ocean boundary

inshore and underneath of the freshwater flow system. The fact that the saltwater

wedge extends more than two kilometres inland from the coast is an indication of

how transmissive the limestone is (Figure 16). In addition, groundwater levels



several hundred metres inland from the coast are strongly correlated with monthly

average sea level (PPK, 2000).

Because the deep bores are screened just above the limestone, this affects their tidal

responses. Compared with the Ferris model, the attenuation of tidal amplitude in the

deep bores is much larger than expected for the same speed of propagation.

Therefore, the lag-based estimates of diffusivity, which are in the range 8,400 to

28,100 m2.d-1, are much larger than the efficiency-based estimates, which are in the

range 51,800 to 441,400 m2.d-1. For the reasons discussed, none of these values is

considered realistic and it is concluded that the Ferris model is not suited to the

particular field conditions.

Table 28: Summary of tidal analysis for K1 tidal constituent

Bore ID Distance to Shore [m]

TEK1 Efficiency-based

Diffusivity [m2.d-1]

LagK1 Lag-based Diffusivity [m2.d-1]

NH1a 179 0.152 27,300 0.094 296,900 NH2a 93 0.171 8,400 0.117 51,800 NH2b 96 0.076 4,200 0.293 8,800 NH3 45 0.301 4,200 0.146 7,800 NH4 28 0.451 3,600 0.090 7,700 NH5 18 0.611 4,000 0.064 6,500

Table 29: Summary of tidal analysis for O1 tidal constituent

Bore ID Distance to Shore [m]

TEO1 Efficiency-based

Diffusivity [m2.d-1]

LagO1

Lag-based Diffusivity

[m2.d-1]

NH1a 179 0.147 28,100 0.075 441,400 NH2a 93 0.174 9,100 0.044 347,600 NH2b 96 0.075 4,400 0.246 11,700 NH3 45 0.321 5,000 0.115 11,800 NH4 28 0.493 4,900 0.081 9,000 NH5 18 0.639 5,200 0.050 10,200



Figure 76: Non-dimensional plot of normalised lag verses tidal efficiency for groundwater monitoring bores NH1 to NH5

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.001 0.01 0.1 1

Tidal Efficiency

Nor

mal

ised

Lag

Ferris

K1

O1

ShallowBores

DeepBores

APPENDIX B Technical Report - January 2003


APPENDIX B: BENTHIC FLUX METERS

CSIRO Land and Water APPENDIX B


Three types of benthic flux meter (BFM) are referred to in this study:

Lee-type

Ultrasonic

Heat-pulse

A brief description of their design and operation is presented below.

LEE-TYPE

Lee-type meters (Lee, 1977) consist of an open drum with a flexible plastic bag

connected by a short length of conduit. With the open end of the drum pushed

securely into the sediment and the meter fully submerged, the plastic bag is checked

periodically and the time-integrated benthic flux calculated from the change in

volume of water in the plastic bag. An increase of water volume indicates a mean

flux of porewater out of sediment into the water column during the measurement

period; a decrease indicates a mean flux of seawater into sediment. Benthic flux, in

units L.T-1, was calculated by diving the change in volume [L3.T-1] by the ‘footprint’

area of the collection drum [L2].

ULTRASONIC

The ultrasonic BFM used during the SGD intercomparison was developed in the

USA (Paulsen et al., 1997). Flow velocity is measured based on the difference in the

speed of propagation between ultrasonic waves traveling in the direction of flow and

against the direction of flow. In operation, a collection device is pressed into the

seabed and the flow amplified by passing it through a relatively small diameter tube

in which the flow velocity is measured using ultrasonics. A distinct advantage of

this instrument is its speed. Essentially instantaneous measurements of SGD can be

made because the time required for ultrasonic waves to travel the length of the

analyser tube is very short.



HEAT-PULSE

The heat-pulse BFM referred to in this study was developed by CSIRO Land and

Water. The operating principle is based on using heat as a tracer of flow (Figure 77).

Flow through the collection device on the seabed is amplified through a relatively

small diameter tube that contains a small heater and electronic temperature sensors

(thermistors). The heater is positioned mid-way in the thermistor array and is

controlled electronically. To make a measurement, the heater is turned on for 0.2

seconds and the direction and velocity of flow in the analyser tube estimated based

on the direction and speed of travel of the heated water, which is detected by the

thermistor array. Temperature at each thermistor is recorded continuously at

0.2-second intervals. The objective is to heat the water to a temperature that is

detectable by the thermistors but not to the point that natural convection perturbs the

flow.

Figure 78 depicts example heat-pulse data measured during instrument calibration.

Each curve corresponds to the temperature time series at the labeled thermistor

(Figure 77). Thermistors were equally spaced at 10 mm intervals in the instrument

used in this study. A temperature count of 50 was approximately equivalent to 1

degree Celsius. In Figure 78, the heater was ‘fired’ at 5 seconds and the heat pulse

moved down the analyser tube in the direction of thermistors T4, T5 and T6. In

operation, this would indicate a flux of porewater from sediment into the water

column. The apparent flow velocity in the analyser tube is calculated by dividing the

distance between two thermistors by the time taken for the heat pulse to travel

between them. Two estimates can be made using the inner pair of thermistors, T2-

T3 or T4-T5, and the outer pair, T1-T3 or T4-T6, dependent on the direction of flow.

The travel time between thermistors is calculated as either the time difference

between peak temperatures at the thermistors or the time difference between the

arrivals of the temperature front at thermistors. The latter method is preferred if

there are flow reversals in the tube since this causes multiple temperature peaks that

are difficult to distinguish.

The true flow rate is obtained from calibration curves that relate the apparent flow

velocity in the analyser tube to the true flow rate determined by electronic mass

balance. Calibration curves for the inner and outer pairs of thermistors are presented

in Figure 79 and Figure 80. The correspondence between the apparent and true flow

rates is not one-to-one because the transport of heat in the analyser tube is not purely

advective; heat is also transported by conduction, which is a diffusive process. In



general, the calibration model is less precise and more uncertain for higher flow rates

since travel times between thermistors are smaller and this number is the divisor in

the calculation of apparent flow rate. The minimum flow rate that was practical to

detect in the laboratory was approximately 1 mm.s-1. Below this rate, the heat

transport was achieved mainly by conduction (i.e. the Péclet number was large). At

lower flow rates, the heat pulse traveled away from the heater at approximately the

same rate in both directions.

T1T2 T3 T4

T6T5

+ve

-ve

Thermistors

and flow

direction

Heater

Figure 77: Heat-pulse BFM



BFM Data - LaboratorySteady Flow 19/2/01, 12:50:00 PM

2140

2160

2180

2200

2220

2240

2260

2280

2300

0 10 20 30 40 50 60 70

Time [s]

Tem

p [

cou

nts

]

T1 T2 T3 T4 T5 T6

Figure 78: Example heat-pulse data from laboratory calibration; flow rate = 2.8 mm.s-1

Calibration on Inner Thermistor Pair

y = 0.965x + 2.1076

R2 = 0.99

0

1

2

3

4

5

6

7

8

9

10

11

12

0 1 2 3 4 5 6 7 8 9 10 11 12

Mass Balance Flux [mm/s]

BF

M F

lux

[mm

/s]

1st Thermistor Pair Linear regression

Figure 79: Calibration curve for the heat-pulse BFM using inner thermistor pairs T2-T3 and T4-T5



Calibration on Outer Thermistor Pair

y = 1.2647x + 0.9252

R2 = 0.9961

0

1

2

3

4

5

6

7

8

9

10

11

12

0 1 2 3 4 5 6 7 8 9 10 11 12

Flux from Mass Balance [mm/s]

Flu

x fr

om

BF

M A

nal

yser

[m

m/s

]

2nd Thermistor Pair Linear regression

Figure 80: Calibration curve for the heat-pulse BFM using outer thermistor pairs T1-T3 and T4-T6

APPENDIX C Technical Report - January 2003


APPENDIX C: MODELS OF TIDE AND WAVE AFFECTS ON BENTHIC

FLUX

CSIRO Land and Water APPENDIX C


EFFECT OF SEA LEVEL DYNAMICS

As groundwater in an unconfined aquifer moves toward and eventually discharges at

the coast, the direction of groundwater flow changes from predominantly horizontal

flow in the aquifer inshore from the coastline to predominantly vertical flow directly

beneath the seabed. This change from horizontal to vertical flow generally takes

place near to the shoreline or area of SGD.

Harmonic solutions of 1D, unsteady groundwater flow (e.g. Ferris, 1951 and

Townley, 1995) are commonly applied to simulate aquifer tidal responses that are

due to lateral propagation of ocean water-level fluctuations into an aquifer. Tidal

components are typically diurnal (1-day period) and can induce water level responses

and lateral oscillations in groundwater flow at significant distances inshore from the

beach if the aquifer is sufficiently permeable. In such cases, the aquifer is simulated

as a horizontal flow system and vertical flow components near to the beach are

ignored.

The same harmonic solutions also can be applied to simulate vertical oscillations in

benthic flux across the seabed. These are induced by tidal and longer-period sea

level fluctuations but also arise due to higher-frequency fluctuation in sea level such

as swell and waves. Coastal swell and waves typical have periods in the order of

seconds to ten-of-seconds. The affects of these high-frequency fluctuations in sea

level do not propagate far into the aquifer because the boundary condition is

perturbed only for a short period. Predominantly small-scale vertical oscillations are

induced in groundwater flow directly beneath the seabed. Although the volume of

water that is exchanged across the seabed is relatively small compared to tidal and

longer-period exchanges, the instantaneous fluxes caused by waves and swell can

still be significant because the period of exchange is small (i.e. flux = volume/time).

Two analytical solutions (Ferris, 1951; Townley, 1995) are suitable for quantitatively

exploring the affects of tides, waves and swell on benthic flux. Both solutions

assume a 1D homogeneous aquifer with a harmonic head (sea level) boundary

condition at the ocean. The Ferris solution is for a so-called semi-infinite aquifer

that is infinitely deep and extends an infinite distance inshore. The Townley solution

allows for a finite depth and length aquifer with impermeable base.



The purpose of the following analyses is to demonstrate that the theoretical

magnitudes of tide-induced and wave-induced benthic fluxes are significant in

comparison with the estimates rates of fresh terrestrial SGD into Cockburn Sound.

TIDE-FORCED MODEL

The tide-pumping model explores lateral propagation of tidal flows into and out of

the aquifer. It applies the Townley solution for 1D-transient groundwater flow in an

aquifer with a tidal or harmonic boundary condition at x = 0. For each tidal

constituent the governing flow equations are

t

h

T

S

x

h

∂∂=

∂∂

2

2

(23)

)cos(),0( tHth a ω= (24)

0),( =

∂∂−

x

tLhT (25)

where h is hydraulic head [L], S is the aquifer storage coefficient [1], T is

transmissivity [L2/T], L is aquifer length [L], x is distance inshore from the tidal

boundary, t is time, Ha is the amplitude [L] of sea level fluctuation at the tidal

boundary, ω = 2π/P is the angular frequency of sea level fluctuation [T-1] and P is the

period [T]. The solution, expressed by Townley using complex variables, is

α

α

cosh

)(cosh

)(⎟⎠⎞

⎜⎝⎛ −

= L

xL

Hxh ap (26)

TP

SLi

2

2πα = (27)

where hp is a complex head, α is complex and 1−=i . The amplitude of

fluctuation in head is equal to |hp| and the phase is equal to arg(hp). Note that in (26)

the group of terms to the right of Ha is an efficiency function that is equal to 1 at x =

0 where the amplitude of groundwater level fluctuations is equal to the amplitude of

sea level fluctuations. The value of this function decays with increasing distance



away from the shoreline and the rate of decay is determined by the value of the non-

dimensional ratio L2S/TP.

The issue of interest is the groundwater flux across the coastal boundary at x = 0

αα tanhL

THQ ap = (28)

where Qp is a complex flux [L2/T]. Because the inshore boundary of the aquifer is a

long way from the coast, equation (28) can be simplified by recognising that

tanh α ≈ 1 when L is large

P

STHQ a

a

2

2π= (29)

In other words, the amplitude of tidal flux, Qa, across the coastal boundary is a

function of the ratio Ha2ST/P. This relation can be applied to typical parameter

values for the coastal aquifer adjacent to Cockburn Sound, which are listed below

0.05 < Ha < 0.25 [m]

0.1 < S < 0.3 [1]

200 < T < 2000 [m2.d-1]

P ≈ 1 [d]

0.2 < Ha2ST/P < 150 [m.d-1]

Figure 81 is a plot of Qa for these ranges and it is clear that the simulated tidal

groundwater flows into and out of the aquifer are the same order of magnitude as the

estimated values of terrestrial SGD into Cockburn Sound. The approximate range of

values in Figure 81 is 1 < Qa < 30 m3.d-1.m-1. It should be expected that benthic flux

meters that are capable of detecting the terrestrial component of SGD would also

detect tidal fluxes.

WAVE-FORCED MODEL

The wave-pumping model simulates vertical fluctuations in groundwater flow

beneath the seabed in response to relatively high-frequency oscillations in sea level

that do not propagate significant distances laterally into the aquifer. For a particular

wave period the governing equations for confined groundwater flow are



t

h

K

S

z

h o

∂∂=

∂∂

2

2

(30)

)cos(),0( tHth a ω= (31)

0),( =

∂∂−

z

tDhK (32)

where So is the aquifer specific storativity [L-1], K is vertical hydraulic conductivity

[L/T], D is the thickness of aquifer beneath the seabed [L] and z is distance beneath

the seabed. The solution (Townley, 1995) is

β

β

cosh

)(cosh

)(⎟⎠⎞

⎜⎝⎛ −

= D

xD

Hzh ap (33)

KP

SDi o

2

2πβ = (34)

In this case, the aquifer tidal efficiency is determined by the value of the non-

dimensional ratio D2So/KP. Benthic flux across the seabed at z = 0 is

ββ tanhD

KHq ap = (35)

where qp has units [L/T]. The amplitude of flux is shown below to be a function of

two ratios, Ha2SoK/P and D2So/KP, but the solution is insensitive to values of

D2So/KP that are greater than 2

⎟⎟⎠

⎞⎜⎜⎝

⎛++

−=γγγ

γγπ2cos2exp24exp1

2cos2exp412

2

P

KSHq oa

a (36)

KP

SD o2

πγ = (37)



Typical parameter values for Cockburn Sound are

0.1 < Ha < 0.5 [m]

10-3 < So < 10-5 [m-1]

10 < K < 100 [m.d-1]

10 < P < 15 [s]

25 < D < 35 [m]

0.01 < Ha2SoK/P < 230 [m2.d-2]

1 < SoD2/KP < 1000 [m2.d-2]

Values of qa are plotted in Figure 82, where the approximate range is

1 < qa < 40 m.d-1. Note that these values will be larger when integrated over an area

of seabed and it is clear they are the same order of magnitude and potentially larger

than tidal groundwater flows and estimates of the terrestrial component of SGD into

Cockburn Sound. This result predicts that there can be significant fluctuations of

benthic flux into the out of the seabed over time scale of tens of seconds.

H TS2

/Pa

Qa[m

/d/m

]3

0 20 40 60 80 100 120 140

5

10

15

20

25

30

Figure 81: Tide-pumping model; amplitudes of tidal groundwater fluxes across the coastal boundary calculated for typical values of tidal range and aquifer properties



H KS P2

o/a

Qa[m

/d]

0 50 100 150 200

5

10

15

20

25

30

35

40

Figure 82: Wave-pumping model; amplitudes of wave-induced groundwater fluxes across the seabed calculated for typical values of wave height and aquifer properties

Date post:	05-Oct-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

Quantifying Submarine Groundwater Discharge and Nutrient ...Anthony J. Smith, Jeffrey V. Turner,...

Documents