+ All Categories
Home > Documents > A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow...

A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow...

Date post: 10-Sep-2018
Category:
Upload: truongduong
View: 219 times
Download: 0 times
Share this document with a friend
13
A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S. Globa and S.L. Barbour Abstract: An investigation and remediation of instability along cut slopes of an irrigation canal in salt-rich clayey col- luvium in southern Alberta is described. Random sloughing of the canal side slopes began to occur 2 years after con- struction over a 7500 m length of canal and was nearly continuous over a 2500 m length 3 years later. The sloughs are shallow seated and most had the appearance of a “mud” flow. Soil salinity also developed along the canal slopes due to capillarity and evaporation. Failure of the clay slopes in the cut sections is attributed to swelling, dispersion, and softening of the subgrade due to exposure of the salt-laden clays to fresh water. In contrast, 600 m of salt-rich clayey colluvium compacted as a clay liner on the canal side slopes showed no evidence of sloughing and salt loading after 9 years of service. Good soil compaction inhibited softening and slope instability in these salt-rich soils. Results from de- tailed site investigations and observations, and the interplay of soil chemistry and geology, facilitated the selection of a satisfactory method of repair. Repairs consisted of reconstructing 2500 m of canal by overexcavating the canal subgrade and replacing those soils with compacted soil. Key words: clays, salts, dispersive, sloughing, salinity, compaction. Résumé : On présente une étude et un comfortement de l’instabilité le long de pentes excavées d’un canal d’irrigation dans les colluvions argileux riches en sel dans le sud de l’Alberta. Des glissements superficiels aléatoires dans les pen- tes latérales du canal ont commencé à se produire 2 ans après la construction sur une longueur de 7500 m de canal et étaient quasiment continus sur une longueur de 2500 m après 3 ans. Les surfaces de glissement sont peu profondes, et la plupart avaient une apparence d’écoulements de boue. La salinité du sol s’est accentuée le long des pentes du canal à cause de la capillarité et de l’évaporation. La rupture des pentes argileuses dans les sections excavées est attribuée au gonflement, à la dispersion, et au ramollissement de la fondation par suite de l’exposition à l’eau douce des argiles chargées de sel. Par opposition, 600 m de colluvions argileux riches en sel compactés comme membrane argileuse sur les pentes latérales du canal n’ont montré aucune évidence de glissement de surface et d’absorption de sel après 9 ans de service. Un bon compactage du sol a empêché le ramollissement et l’instabilité de la pente dans ces sols riches en sel. Les résultats des études et observations détaillées du site, et l’interaction de la chimie et de la géologie ont facilité la sélection d’une méthode satisfaisante de réparation. Les réparations ont consisté en la reconstruction de 2500 m de canal par surexcavation de la fondation du canal et le remplacement de ces sols par un sol compacté. Mots clés : argiles, sels, dipersive, glissement de surface, salinité, compactage. [Traduit par la Rédaction] Globa and Barbour 677 Introduction and background This paper presents a case history of an unusual slope- stability problem along canal slopes excavated in salt-rich clayey colluvium. The canal, part of the Blood Tribe Irriga- tion Project, is located southwest of Lethbridge, Alberta, on the Blood Indian Reservation. The major components of the turnkey project, constructed between 1990 and 1994, consist of the 20 km Mokowan Ridge Canal, a 13 km long header canal, and a 19 m high earthen dam for an inland storage reservoir. The completion date for all components of the project is 2001. The canals are referred to herein as Mokowan Ridge Canal reaches 1–5 (MR1–MR5) as shown in Fig. 1 and header canal reaches 1–5 (HC1–HC5), which are not shown. A typical canal section in cut consists of 2.4 m high canal banks with 3:1 (H:V) side slopes, a bottom width of 5.5 m, and a water depth of 1.5 m. Maximum excavation depths were 8 m. Compacted clay sections made up about 600 m of the canal in MR3 and consisted of a 900 mm thick continu- ous clay liner placed across the entire canal section on side hill sections, and shorter sections of compacted two-bank fill Can. Geotech. J. 38: 665–677 (2001) © 2001 NRC Canada 665 DOI: 10.1139/cgj-38-4-665 Received July 5, 2000. Accepted January 2, 2001. Published on the NRC Research Press Web site at http://cgj.nrc.ca on Aug. 1, 2001. R.S. Globa. Agriculture and Agri-Food Canada Prairie Farm Rehabilitation Administration, 1800 Hamilton Street, Regina, SK S4P 4L2, Canada. S.L. Barbour. 1 Department of Civil Engineering, University of Saskatchewan, Saskatoon, SK S7N 5S3, Canada. 1 Corresponding author (e-mail: [email protected]).
Transcript
Page 1: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

A case history of shallow sloughing within cutslopes of an irrigation canal in salt-rich clayeycolluvium

R.S. Globa and S.L. Barbour

Abstract: An investigation and remediation of instability along cut slopes of an irrigation canal in salt-rich clayey col-luvium in southern Alberta is described. Random sloughing of the canal side slopes began to occur 2 years after con-struction over a 7500 m length of canal and was nearly continuous over a 2500 m length 3 years later. The sloughs areshallow seated and most had the appearance of a “mud” flow. Soil salinity also developed along the canal slopes dueto capillarity and evaporation. Failure of the clay slopes in the cut sections is attributed to swelling, dispersion, andsoftening of the subgrade due to exposure of the salt-laden clays to fresh water. In contrast, 600 m of salt-rich clayeycolluvium compacted as a clay liner on the canal side slopes showed no evidence of sloughing and salt loading after 9years of service. Good soil compaction inhibited softening and slope instability in these salt-rich soils. Results from de-tailed site investigations and observations, and the interplay of soil chemistry and geology, facilitated the selection of asatisfactory method of repair. Repairs consisted of reconstructing 2500 m of canal by overexcavating the canalsubgrade and replacing those soils with compacted soil.

Key words: clays, salts, dispersive, sloughing, salinity, compaction.

Résumé: On présente une étude et un comfortement de l’instabilité le long de pentes excavées d’un canal d’irrigationdans les colluvions argileux riches en sel dans le sud de l’Alberta. Des glissements superficiels aléatoires dans les pen-tes latérales du canal ont commencé à se produire 2 ans après la construction sur une longueur de 7500 m de canal etétaient quasiment continus sur une longueur de 2500 m après 3 ans. Les surfaces de glissement sont peu profondes, etla plupart avaient une apparence d’écoulements de boue. La salinité du sol s’est accentuée le long des pentes du canalà cause de la capillarité et de l’évaporation. La rupture des pentes argileuses dans les sections excavées est attribuée augonflement, à la dispersion, et au ramollissement de la fondation par suite de l’exposition à l’eau douce des argileschargées de sel. Par opposition, 600 m de colluvions argileux riches en sel compactés comme membrane argileuse surles pentes latérales du canal n’ont montré aucune évidence de glissement de surface et d’absorption de sel après 9 ansde service. Un bon compactage du sol a empêché le ramollissement et l’instabilité de la pente dans ces sols riches ensel. Les résultats des études et observations détaillées du site, et l’interaction de la chimie et de la géologie ont facilitéla sélection d’une méthode satisfaisante de réparation. Les réparations ont consisté en la reconstruction de 2500 m decanal par surexcavation de la fondation du canal et le remplacement de ces sols par un sol compacté.

Mots clés: argiles, sels, dipersive, glissement de surface, salinité, compactage.

[Traduit par la Rédaction] Globa and Barbour 677

Introduction and background

This paper presents a case history of an unusual slope-stability problem along canal slopes excavated in salt-richclayey colluvium. The canal, part of the Blood Tribe Irriga-tion Project, is located southwest of Lethbridge, Alberta, onthe Blood Indian Reservation. The major components of theturnkey project, constructed between 1990 and 1994, consistof the 20 km Mokowan Ridge Canal, a 13 km long headercanal, and a 19 m high earthen dam for an inland storagereservoir. The completion date for all components of the

project is 2001. The canals are referred to herein asMokowan Ridge Canal reaches 1–5 (MR1–MR5) as shownin Fig. 1 and header canal reaches 1–5 (HC1–HC5), whichare not shown.

A typical canal section in cut consists of 2.4 m high canalbanks with 3:1 (H:V) side slopes, a bottom width of 5.5 m,and a water depth of 1.5 m. Maximum excavation depthswere 8 m. Compacted clay sections made up about 600 m ofthe canal in MR3 and consisted of a 900 mm thick continu-ous clay liner placed across the entire canal section on sidehill sections, and shorter sections of compacted two-bank fill

Can. Geotech. J.38: 665–677 (2001) © 2001 NRC Canada

665

DOI: 10.1139/cgj-38-4-665

Received July 5, 2000. Accepted January 2, 2001. Published on the NRC Research Press Web site at http://cgj.nrc.ca on Aug. 1,2001.

R.S. Globa.Agriculture and Agri-Food Canada Prairie Farm Rehabilitation Administration, 1800 Hamilton Street, Regina, SKS4P 4L2, Canada.S.L. Barbour.1 Department of Civil Engineering, University of Saskatchewan, Saskatoon, SK S7N 5S3, Canada.

1Corresponding author (e-mail: [email protected]).

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:39 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 2: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

sections. All canal side slopes were covered with armourgravel to about 1.9 m above the invert for erosion protectionand along the invert only in compacted clay liner sections tominimize weathering of the compacted clay.

There were no incidences of slope instability on cutslopes in MR3–MR5 and HC1 during construction. How-ever, shallow sloughing began to occur late in the winter of1992 about 2 years after construction throughout MR3 tostation 15 000 in MR4. Sloughing occurred annually in thissection during late winter and was nearly continuous over a2500 m section by 1996. In contrast, random sloughing oc-curred as a one-time event in the rest of MR4 and in MR5and HC1. In most cases, the failures were typically shallow,“quick” or flow-type sloughs that occurred in fully softenedclay directly below the armour gravel (Fig. 2). Sloughingwas initially attributed to frost action partly because of theextremely silty appearance of the clay-rich soil and the rela-tively high water table. Observations showed no sloughingactivity in MR3 on slopes consisting of compacted clay.

The performance of the canal was not seriously affectedby this sloughing activity, but contractual obligations re-quired that the canal be restored to an as-built condition be-fore transfer of ownership to the Blood Tribe. Consequently,a detailed investigation was initiated in 1996 to understandthe mechanism of the unusual sloughing activity and to de-termine a solution to the problem. The major concerns werethe unusual form of the slope failures in clay and the differ-ences in the sloughing activity, which occurred under whatappeared to be similar conditions in MR3 to station 15 000

in MR4, and the one-time sloughing events that occurred be-yond station 15 000 in MR4 and in MR5 and HC1. Previousattempts to repair the slopes had been made between 1992and 1996 with limited success.

Observations

The following observations, made prior to the 1996 inves-tigation, helped to characterize the sloughing activity alongthe canal.

• Sloughing activity was commonly associated with thaw-ing conditions in late winter or during spring snowmelt.Construction staff assumed that sloughing was associatedwith frost heaving due to the silty nature of the clay-rich col-luvium. The onset of sloughing did not appear to be associ-ated with heavy precipitation events. Record rainfalls of upto 760 mm occurred in the summer months of 1993–1995with no related sloughing activity. However, occasionalsloughing did occur in subsequent summers at locationswhere water tended to pond in drainage ditches. This pondedwater contributed to salt loading on the surface of exposedditch cut slopes through capillarity and evaporation.

• Near-continuous sloughing activity of the canal sideslopes was observed in clay- rich colluvium of MR3 in com-parison to random sloughing that occurred in the clay-richalluvium and clay till of MR4 and MR5. The sloughingevents within MR4 and MR5 were isolated one-time events,with no failures in subsequent years.

© 2001 NRC Canada

666 Can. Geotech. J. Vol. 38, 2001

Fig. 1. Airphoto of canal reaches MR2–MR5, topographic characteristics, and geological units.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:46 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 3: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

• The observed slip surface tends to be shallow, occurringin the softened subgrade just below the armour gravel andgenerally involving the entire slope section covered with ar-mour gravel. The most common type of failure in MR3 andin MR4 up to station 15 000 was a “quick” type failure re-sembling a mudflow retrogressing from the toe. This wassometimes accompanied by multiple scarps above the slidearea. Sloughing in some areas occurred over a slickensidedsurface with heaving into the space from which the sloughdebris was removed. In the remainder of MR4 and MR5 thesize and location of the sloughs were randomly distributed,with no preferred orientation to the slip surfaces. Thesloughs were generally planar, parallel to the toe of theslope, and less than several metres in length and could befound almost anywhere along the armour-gravelled slopes inthe identified canal sections. In many instances, the sloughswere totally confined by armour gravel prior to failure.

• In some areas, the clayey subgrade near the toe of thecanal slopes softened to such an extent that the subgradecould not support the weight of a person.

• Long-term piezometric levels were at least 1 m belowthe canal invert, ruling out high piezometric levels as a con-tributing factor to slope instability.

• Surficial salt deposits were commonly observed in areasof slough debris, exposed slip surfaces, aggregate surfaces,and areas of exposed subgrade following dry periods. Thesesalt deposits developed as a result of upward migration ofsalts by capillary movement of water from groundwater andwater ponded at the toe of cut slopes in response to surfaceevaporation.

• Pinhole tests (American Society for Testing and Mate-rials (ASTM) Standard D4647-87) confirmed the presenceof potentially dispersive clays that could be a contributingfactor to random erosion or scouring along the canal invert.

• Sloughing activity and salt deposits were nonexistentwhere topsoil and grass were present on the canal slopesabove the armour gravel.

• Observations showed no sloughing activity occurred dur-ing or immediately after the release of fresh water down thecanal.

• Observations showed no sloughing activity in a 600 mlength of compacted clay liner in MR3 constructed fromclay-rich colluvium between 1990 and 1996.

• Site inspectors at times had difficulty distinguishing be-tween clay-rich soils consisting of colluvium, alluvium, orclay till. Although this was not a problem for construction, itdid point out the need to reevaluate site geology.

Site geology

The MR3 canal to station 15 000 of MR4 (Fig. 3) is in-cised into colluvial deposits overlying deposits of silt, siltysand, and bedrock. The colluvial deposits consist of low- tohigh-plasticity clay with some silt and silty sand layers de-rived from the Willow Creek Formation found atop theMokowan Ridge Butte, which is located several kilometresupslope of MR3, MR4, and MR5. The underlying silt andsilty sand deposits are more uniform and areally continuous.

© 2001 NRC Canada

Globa and Barbour 667

Fig. 2. Shallow sloughing of canal slopes downstream of station 13 878 in MR3 (summer of 1996).

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:49 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 4: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

These deposits may have been derived from coarser grainedlacustrine deposits.

The Willow Creek Formation and St. Mary River Forma-tion are composed of sandstone and clay shale. The WillowCreek Formation is nonmarine in origin, whereas the St.Mary River Formation is of marine origin. Soils derivedfrom the St. Mary River Formation have a greenish tinge asopposed to a red tinge for soils derived from the WillowCreek Formation. From stations 15 000 to 16 900 in MR4,the canal is incised into alluvial clay deposits. Stalker (1963)originally classified these deposits as lacustrine depositsconsisting of fine silt, sand, and clay that is locally varved ingenerally thin deposits over ground moraine. However, re-cent test pits described by T. Dash (personal communication,1997) noted minor pebble and sand layering, with no indica-tion of varved material as indicated by Stalker, which sug-gests these deposits are alluvial in origin. Shallow, clay-richslope wash overlies the clay till, which in turn overlies theSt. Mary River Formation beyond station 16 900, to the endof MR5 at station 20 400. Test pits and airphotos generallyconfirmed the boundaries between colluvium, alluvium, claytill, and bedrock (T. Dash, personal communication, 1997).

Construction records indicate the canal is incised into claycolluvium with some pockets and layers of silt and silty sandthroughout MR3 up to station 15 000 of MR4. Between sta-tions 15 000 and 16 900, the canal is incised into alluviumthat is predominately clay with some silts and sands. Beyond

station 16 900 to the end of MR5 at station 20 400, the canalis incised mainly in clay till that is mantled by 1–2 m ofslope wash rich in clay. In places, the canal was incised intoclay shale from the St. Mary River Formation.

The airphoto in Fig. 1 shows that the topographic charac-ter from MR3 to station 15 000 in MR4 is similar. The collu-vial deposits clearly show gully-type patterns caused byerosion, even though they are partly obscured by cultivation.In contrast, the topography between stations 15 000 and16 900 is smooth and the gully patterns are generally absent.The significance of gully erosion was not obvious during theoriginal canal investigation in 1989.

Soil investigations

Foundation conditions along the canal slopes, canal right-of-way, and compacted clay sections were investigated inmuch greater detail in 1996 than in the original 1989 investi-gation. In 1989, few problems were anticipated with thesesoils and effort was placed on locating the canal in impervi-ous soils and obtaining a balanced cut and fill. The emphasisof the 1996 investigation was placed on comparing physicaland chemical properties of the soils in the canal excavationin both failed and unfailed slope areas with those foundalong the canal right-of-away that extended about 20 m out-wards from the canal excavation.

© 2001 NRC Canada

668 Can. Geotech. J. Vol. 38, 2001

Fig. 3. Cross section along the centreline of Mokowan Ridge Canal reaches MR3 and MR4. F.S.L., Full Supply Level.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:53 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 5: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

The soils found in MR3 consist primarily of silty, me-dium-plasticity clay colluvium (clay-rich colluvium) andsome deposits of high-plasticity clay and lesser amounts ofsilts and sands. The soils in MR4 up to station 15 000 con-sist primarily of medium-plasticity clay colluvium withsome low-plasticity clay. The remainder of MR4 consistsprimarily of a low-plasticity clay alluvium. The soils foundin MR5 were mantled by up to 2 m of medium-plasticityclay, identified as slope wash, overlying medium-plasticityclay till and some protrusions of clay shale from the St.Mary River Formation. Pockets of silt and sand encounteredduring construction were replaced with compacted clay.General physical properties for the soils in canal reachesMR3–MR5 are shown in Table 1. The soil properties do notsuggest any unusual attributes for the clay-rich colluviumexcept that the dry densities are lower than for either clay al-luvium or clay till.

ConstructionCanal sections in MR3–MR5 were mainly in cut except

for a near-continuous 600 m section of compacted clay linerin MR3. Few difficulties occurred with excavation and han-dling of the clay-rich colluvium.

The compacted clay liner had been constructed from clay-rich colluvium obtained from excavating the canal in MR3.There were no unusual problems with the colluvium duringcompaction operations. Soil tests (Table 1) showed that thephysical properties of the clay used to construct the com-pacted clay liner were similar to those of the soils foundthroughout MR3. Two sets of compaction standards wereused to construct the clay liner. Clay above the upper eleva-tion of the armour gravel required compaction to at least92% of Standard Proctor maximum dry density at a permis-sible water content ranging from 2% dry to 3% wet of opti-

mum water content. Field density tests showed an average96.7% of Standard Proctor maximum dry density with anaverage water content of 1.0% above the optimum watercontent was obtained during construction. Clay compactedbelow the upper elevation of the armour gravel requiredcompaction to at least 95% of Standard Proctor maximumdry density at a permissible water content ranging from 0 to2% above optimum water content. Tests showed an average98.1% of Standard Proctor maximum dry density with anaverage water content of 0.6% above the optimum watercontent was obtained during construction. A sheepfoot rollerwith a foot contact pressure in excess of 2400 kPa was usedfor compaction.

Swell testingA limited number of swell tests under in situ surcharge

conditions were carried out on undisturbed samples of clay-rich colluvium obtained from canal reaches MR3 and HC1(Table 2). Samples were selected on the basis of soil pore-water chemistry. Limited testing suggested some correlationof swelling with percent sodium (%Na = Na/TDS, whereTDS = Ca + Mg + Na + K) and the sodium adsorption ratio(SAR = Na/[(Ca + Mg)/2]0.5) and a poor correlation withTDS. All ion concentrations are expressed in milliequi-valents per litre (mequiv./L). Samples with high %Na con-centrations showed total swells of 2.2 and 4.3% whenimmersed in distilled water. There was no apparent correla-tion between clay content (percent passing 2µm sieve size)and total swell, which suggests swelling may be related toclay mineralogy and sodium content. Although swelling ofthe clays in the subgrade contributed to slope failure, ex-treme spatial variation in clay mineralogy, ionic species, andconcentration, over very small sampling distances, indicatedadditional free swell tests would be pointless. It will be

© 2001 NRC Canada

Globa and Barbour 669

Canal reach Soil group γd (t /m3) WL (%) WP (%) IL % silt % clay

MR3 CI 1.52(26) 42(58) 21(58) 0.21(58) 59(63) 37(63)CI(clay liner) 1.54(4) 44(12) 21(12) 0.09(12) 55(11) 42(11)CH 1.46(6) 58(6) 24(6) 0.08(6) 42(6) 55(6)

MR4 CI — 41(12) 19(12) 0.12(12) 54(12) 38(12)CL — 29(4) 17(4) 0.41(4) 57(4) 23(4)

MR5 CIT 1.67(2) 38(6) 17(6) –0.02(6) 51(5) 28(5)CI 1.59(2) 39(6) 18(6) 0.03(6) 49(5) 31(5)CS 1.51(1) 53(3) 25(3) –0.02(3) 49(3) 32(3)

Note: Numbers in the table show the average test value, with the number of tests in parentheses. CL, CI, CH, CIT, and CS denote low-, medium-, andhigh-plasticity clay, medium-plasticity clay till, and clay shale, respectively;γ d, WL, WP, and IL are dry unit weight, liquid limit, plastic limit, and liquidityindex, respectively; % silt and % clay denote the percentage of material passing the 74 and 2µm sieve sizes, respectively.

Table 1. Physical soil properties for canal reaches MR3–MR5 from the 1996 investigation.

Canalreach

Soilgroup

WL

(%)WP

(%) IL

%silt

%clay

Ec

(dS/m) %Na SARTDS(mequiv./L)

Totalswell (%)

MR3 CI — — — 56 42 2.8 21 2 40 0.07MR3 CI 48 22 0.13 48 46 0.5 19 2 6 0.03MR3 CI 39 22 0.15 74 26 9.0 68 19 126 2.20HC1 CI 38 16 0.40 58 22 7.0 92 39 73 4.30

Note: Ec, electrical conductivity; %Na, sodium as a percentage of total cation concentration (Ca + Mg + Na + K); SAR, sodium adsorption ratio; TDS,total dissolved solids.

Table 2. Summary of total swell tests and soil pore-water chemistry.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:54 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 6: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

shown later that swelling and dispersion which contributedto slope failure can be correlated with high sodium and mag-nesium levels in the soil pore water.

Clay mineralogyTwo medium-plasticity clay colluvium samples from MR3

and a sample of high-plasticity clay colluvium from HC1were analyzed to determine their clay mineralogy. Analyticalresults are shown in Table 3. The clay fractions analyzed hadhigh percentages of smectite, even though this seemed to beinconsistent with the typical liquid limit values shown forthe respective soil groups. An attempt was made to wash thesalts out of the samples and then determine the effect on theAtterberg limits. This proved to be difficult, as the fine clayflocs tended to clog the filter paper used to retain the clay-sized particles. It is likely that centrifuging the samples mayhave been more effective, but this equipment was not avail-able at the time. Barbour and Yang (1993) demonstrated thatelevated salt concentrations could significantly reduce theliquid limit of clay soils while having little to no affect onthe plastic limit.

Field survey and test pit investigationVisual inspections and test pits were used to confirm site

geology, further evaluate soil structure, and determine soilconsistency of the subgrade in both failed and unfailed slopesections and compacted clay liner. General observations areas follows:

• Compacted clay liner sections in MR3 had not failed af-ter more than 6 years of service. Visual inspections showedthe compacted clay liner, although still intact, had weatheredto a depth of 300–600 mm in places and had developed anugget-like structure within this depth. There were no indi-cations of salt deposits on armour-gravel surfaces such asobserved on cut-slope surfaces in the remainder of MR3.

• In many areas the armour gravel was heavily contami-nated with fines deposited by wind and water. However, itwas also observed that subgrade material on the cut-slopesurfaces had moved upward into the armour gravel. It wasdifficult to determine whether the subgrade had softened orswelled up into the armour gravel.

• Thixotropic behaviour was observed in some clay-richcolluvium in MR3. For example, light, repeated tapping ofthe sides of a shallow hole located in the subgrade causedthe clay-rich colluvium to flow at the point of impact fol-lowed immediately by the presence of free water on the soilsurface. The clay firmed up immediately upon cessation ofthe tapping action. Examination showed the material wasmedium-plasticity clay with a stiff consistency.

• Fully softened clay was randomly encountered in testpits as pockets and layers to a depth of 2 m.

• Evidence of dispersed clay was encountered. Scourholes were present at random locations along the canal bednot protected with armour gravel. Scouring should not havebeen as extensive as observed based on the canal design ve-locity. Small pockets of colloidal-size clay particles werealso observed in places where the clay had flowed out frombeneath the armour gravel along the toe of the side slopes.

Pocket penetrometer surveysPocket penetrometer surveys were used to estimate the

depth of subgrade material that had to be excavated prior toreconstruction. The surveys also delineated the extent ofsubgrade softening due to swelling and allowed comparisonof the soil consistency in terms of the undrained shearstrength in both failed and unfailed cut slopes, as well asslopes constructed with a compacted clay liner. The test pitswere excavated across the entire length of the canal slope.

The data shown in Figs. 4–6 represent pocket penetrom-eter readings taken at a common vertical section at variouslocations along a test pit. Lines joining a common symbolrepresent this vertical section. Figure 4 shows pocket pen-etrometer readings from canal slopes in cut that were judgedstable at the time of the survey. Readings were combinedfrom three test pits to a depth of 1100 mm below the groundsurface. The data show subgrade softening to at least400 mm below the ground surface followed by increasingstiffness with depth. The survey showed that portions of theslopes were on the verge of failure where the shear strengthwas approaching 0 kPa. Figure 5 shows a pocket penetrom-eter survey of a failed cut slope located in MR4. Extremesoftening of the subgrade to a depth of 400 mm is apparentfollowed by an abrupt increase in the undrained shearstrength. Figure 6 shows a pocket penetrometer survey takenon a compacted clay liner section located in MR3. The sur-vey shows a region of lower undrained shear strength in theupper 400 mm of the subgrade followed by an increase withdepth. The undrained shear strength is comparably higherthan those indicated for unfailed canal slopes in cut thatwere judged to be stable at the time of the survey.

Soil pore-water chemistry and pinhole testsSurficial salt deposits developed on failed cut-slope sec-

tions after surface drying, along slip surfaces, on aggregatesurfaces, and on exposed subgrade areas not related to slopefailure following extended periods of surface drying. Thepresence of these surficial salt concentrations is the result ofan ongoing process of capillarity and evaporation. Implica-tions of this process were not fully appreciated until detailedsampling and testing of the subgrade were completed. Sam-

© 2001 NRC Canada

670 Can. Geotech. J. Vol. 38, 2001

Soilgroup

CEC(mequiv./100 g)

Specific surfacearea (m2/g)

Clay mineral composition (%)

Smectite Vermiculite Mica Kaolinite Chlorite Quartz

CI 57 41.2 30–40 30–35 15 10 5 <5CI 55 43.5 40–45 25–30 15–20 10 <5 <5CH 65 60.1 50–60 15–20 10–15 5–10 <5 <5

Note: CEC, cation exchange capacity in cmol/kg of clay particles <2µm (cmol/kg is equivalent to mequiv./100 g). Specific surface area is the specificexternal surface area of sodium-saturated clay.

Table 3. Clay mineralogy for clay-rich colluvium (Mermut 1997).

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:54 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 7: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

ples of both surficial soil deposits and the underlying claysubgrade to a depth of at least 6 m were obtained for soilchemistry analysis and pinhole tests from failed and unfailedcut-slope sections and compacted clay liner sections.

Pinhole testsA total of 36 pinhole tests with reference companion soil

chemistry tests were performed on soil samples obtainedfrom MR3–MR5. In addition, 59 pinhole tests without com-panion soil chemistry tests were also available from previouswork. Studies showed that 51% of 95 samples tested from arandom population were found to be potentially dispersive.A summary of the pinhole tests and soil pore-water chemis-try is provided in Table 4. The large standard deviation indi-cated results from spatial variations in soil pore-waterchemistry. In the pinhole test (ASTM Standard D4647-87), asample is classified as either dispersive (D1 or D2), moder-

ately to slightly dispersive (ND3 or ND4), or nondispersive(ND1 or ND2). A dispersiveness chart, as developed bySherard et al. (1976), showing pinhole test classification ver-sus %Na levels and TDS is shown in Fig. 7. A good correla-tion was obtained with samples classified as dispersive (D1and D2), but the chart was found to be less reliable for sam-ples classified as ND3/ND4 and ND1/ND2. In general, thedata in Fig. 7 suggest soils with a %Na level greater than60% are potentially dispersive.

Distribution of salts in canal reaches MR3,MR4, and MR5

Pinhole testing and soil pore-water chemistry showed thatpotentially dispersive soils and high sodium levels were gen-erally present in MR3–MR5. These high sodium levels in

© 2001 NRC Canada

Globa and Barbour 671

Fig. 4. Pocket penetrometer survey in three stable cut slopes in MR3 and MR4.

Fig. 5. Pocket penetrometer profile of a failed cut slope in MR4.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:54 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 8: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

conjunction with the high smectite content levels in theclays contributed to softening of the subgrade through swell-ing. However, this information did not explain the behav-ioural differences in the ongoing sloughing activity in MR3to station 15 000 in MR4 and the absence of ongoing

sloughing activity in the remainder of MR4 and MR5.Studies undertaken to analyze these behavioural differencesare outlined below.

High %Na levels are a major factor in clay dispersion andswelling. A total of 101 soil chemistry tests were performedusing a “saturation extract” technique as described by Janzen(1993). The major soil cations generally present in MR3 tostation 15 000 in MR4 in decreasing order of concentrationwere Na > Mg > Ca > K, except in MR5 where Ca levelsgenerally exceeded Mg levels. In some tests, the concentra-tion of Mg exceeded Na levels in MR3 and MR4. Table 5shows that %Na levels were even higher in MR4 and MR5than in MR3.

The spatial variation in the soil pore-water concentrationsresulted in a great deal of scatter in the test data. Nonethe-

© 2001 NRC Canada

672 Can. Geotech. J. Vol. 38, 2001

Fig. 6. Pocket penetrometer survey of compacted clay side slope in MR3.

Fig. 7. Relationship between susceptibility to dispersion and dissolved pore-water salts (after Sherard et al. 1976).

Pinhole testresult

No. oftests %Na SAR

TDS(mequiv./L)

D1 11 79(12) 28(29) 138(235)D2 5 81(8) 23(19) 55(53)ND3/ND4 9 56(18) 16(15) 217(338)ND1/ND2 11 44(13) 4(2) 47(49)

Table 4. Summary of pinhole tests and soil pore-water chemistry.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:55 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 9: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

less, there were a sufficient number of tests to delineate apattern of salt distribution. The data shown in Figs. 8–10also include soil pore-water chemistry obtained from thecompacted clay liner samples in MR3, as shown in Table 6.There are numerous locations along the canal where surfacesalinity is visible. Based on these observations, it was ini-tially thought that the salt concentrations would increasenear the surface, due to the upward movement of water as aresult of evaporation.

Figure 8 shows that high levels of %Na were foundthroughout MR3–MR5 and it was interesting to note thatthere was no dramatic attenuation of the sodium concentra-tion with depth as anticipated. The high %Na levels shownin MR4 are located upstream of station 15 000. Figure 9shows a correlation of TDS concentration with depth. Al-though not identified individually, TDS concentrationsdownstream of station 15 000 in MR4 were similar to thosein MR5. The low TDS concentrations downstream of station15 000 in MR4 and MR5 are in sharp contrast to the highsodium levels present in these canal reaches and may be onereason for less swelling and dispersion in these sections.

Another aspect in the behavioural differences of the cutslopes may be seen in the magnesium to calcium cation ratio(Mg/Ca) shown in Fig. 10. Most clayey soil on the Prairies

tends to be dominated by Ca and Na cations. Tests showedthat Na and Mg cations were generally dominant in the claysfound in MR3 and MR4 but not those in MR5. St. Arnaud(1979) indicates a high Mg/Ca ratio is indicative of trans-ported soil salinity and capillary rise from a shallow watertable. Emerson (1979) indicates exchangeable sodium is themain cause of dispersion, but its effect is enhanced by thepresence of exchangeable magnesium rather than calcium.Mermut (1997) suggests Mg can lead to dispersion andswelling much like that due to the presence of Na.

The Mg/Ca ratios in MR3 show a gradual increase withan increase in depth, with peak concentrations occurring at adepth of 2.0–2.5 m. This depth corresponds closely to the lo-cation of the groundwater table. The two extremely high val-ues shown in MR4 were obtained from surface samples inwhich salts have concentrated as a result of evaporation.These deep profiles of elevated Mg levels indicate that thehigh salt concentrations are the result of historical salinityand regional groundwater flow and not simply due to evapo-ration and the surface accumulation of salts.

Groundwater and surface-water qualityWater samples were obtained for chemical analysis from

standpipes located adjacent to the canal in MR3 and MR4, adugout where clay was borrowed to construct the compactedclay liner during construction, and water ponded in the canalin MR3. The standpipes were purged and allowed to stabi-lize before water samples were obtained. Chemical analysisshowed the dominant cation species from the groundwater,dugout, and canal were, respectively, Na > Mg > Ca > K,Na > Ca > Mg > K, and Mg > Na > Ca > K. Sulphate (SO4)was the dominant anion found in all three water sources.The TDS levels of groundwater samples obtained from MR3ranged from 8000 to 18 000 mg/L. Downstream of station15 000 in MR4 the TDS levels were generally less than

© 2001 NRC Canada

Globa and Barbour 673

Fig. 8. Distribution of percent sodium with depth.

Canal reach No. of tests %Na

MR3 77 51(19)MR4 18 70(19)MR5 10 67(31)

Note: Average test values are given, with standarddeviation in parentheses.

Table 5. Distribution of percent sodium in canalreaches MR3–MR5.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:55 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 10: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

5000 mg/L. TDS levels for MR5 were not available but arejudged to be similar to or less than those obtained in the lat-ter part of MR4. The TDS levels of the water samples ob-tained from the dugout and canal were diluted due to surfacerunoff and were 800 and 220 mg/L, respectively.

Analyses and discussion of results

The observations associated with slope instability alongthe canal are consistent with existing literature on dispersiveclays.

Literature descriptionDispersive clays are fine-grained soils in which the clay

particles spontaneously detach from each other and go intosuspension. These soils are susceptible to erosion and are of-ten associated with problems of piping in dams and the for-mation of features such as sinkholes and gullies (Mitchell1993). Sherard et al. (1977) describe problems such as tun-nel and gully erosion of the banks and degradation (erosion)of the canal bottom associated with canals in dispersivesoils. The formation of these dispersive clays is associatedwith low- to high-plasticity clays which contain montmoril-lonite (>10%) (Sherard et al. 1976). These clays are often

© 2001 NRC Canada

674 Can. Geotech. J. Vol. 38, 2001

Fig. 9. Total dissolved cation concentration with depth.

Fig. 10. Mg/Ca ratio with depth.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:55 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 11: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

colluvial or alluvial in origin, being formed as floodplains,slope wash, or lake beds. In addition, these clays often con-tain high concentrations of dissolved salts, particularly so-dium. High salt contents are related to the fact that theseclays are deposited in arid and semiarid climates in whichlow infiltration rates and high evaporation rates promote theaccumulation of salts.

When exposed to fresh water, the diffuse double layer sur-rounding the clay particles swells, increasing the repulsionbetween individual clay particles and resulting in swelling.If unconstrained, the particle may be separated from otherparticles, resulting in complete particle dispersion. Once re-leased, the colloidal clay particles are free to be transportedby moving water. Early work by Australian researchers dem-onstrated that the onset of dispersion was a function of boththe ionic concentration of the eroding water and the dis-solved salts present in the clay. Charts, like that shown inFig. 11, were developed to demonstrate the chemical condi-tions required for the onset of instability. The chart illus-trates that dispersive clays may be stable under elevatedpore-water salt concentrations. As the pore water is diluted,the clays undergo swelling and become potentiallydispersive. If the samples are constrained, then swelling willlead to a new stable but deflocculated structure. However, ifunconstrained while undergoing deflocculation, the clayswill become unstable and disperse. It is also important tonote that clays with exchangeable sodium percentages aslow as 10% can be destabilized if the eroding water has alower salt concentration than the soil pore water.

Application to canal sloughingThe information gathered to date provides conclusive evi-

dence that the colluvium along MR3 is a potentiallydispersive clay. The genesis of these materials is typical ofthe development of dispersive clays in arid and semiarid en-vironments. Thecolluvial soils along MR3 are clay rich(-40%) and contain a significant component of expansiveminerals. These soils are also salt rich and contain high per-centages of sodium (>50%). The pinhole dispersion tests andthe soil chemistry also demonstrate that these materialswould be classified as dispersive. In addition, the behaviourof these materials upon exposure to fresh water, both in thefield and in laboratory testing, is typical of descriptions inthe literature of erosion, swelling–softening, and sloughing–quick conditions upon exposure to fresh water. MR3 is moretypically characterized by piping of the subgrade fines intothe armour gravel and the numerous shallow flow-typesloughs.

Some of the key questions, however, relate to the lack ofongoing failures along the cut slopes downstream of station15 000 in MR4 and along MR5 and the effectiveness ofcompacted liners in MR3 in preventing the sloughing fail-ures.

The geologic interpretation suggests that the canal isfounded primarily in clay till along MR5 and within allu-vium, rather than colluvium, downstream of station 15 000in MR4. Although the clays within this reach of MR4 aresimilar to those in MR3, there are some significant differ-ences in chemistry. Elevated percent sodium levels are pres-ent in all three reaches; however, the TDS levels of soil–extract solutions downstream of station 15 000 in MR4 andMR5 are much lower. This may be due to a differentdepositional environment or may be indicative of less salin-ity development in these areas. In any event, these lower ini-tial salt concentrations would lead to less swelling (and alower potential for dispersion) when exposed to fresh water.Another significant factor in this regard is the profile ofMg/Ca ratios in MR3 and MR4 to station 15 000 relative tothe remainder of MR4 and MR5. Elevated Mg/Ca ratios inMR3 and MR4 are indicative of transported salt associatedwith the development of soil salinity. In addition, elevatedMg concentrations tend to make a clay potentially moredispersive, with a greater swelling potential, relative to Ca.

RemediationIngles and Aitchison (1969) state quite succinctly that

from a chemical perspective the methods of remediationmust either increase the bond strength between the clay par-ticles or decrease the dispersive power of the eroding water.The purpose of a freshwater canal precludes the use of thelatter. Increasing the bond strength would include treatmentof the soil with some form of divalent cation through the useof chemical additives such as calcium hydroxide (lime) orcalcium sulphate. There is evidence that compaction wet ofoptimum can minimize erosion of dispersive clays (Landau

© 2001 NRC Canada

Globa and Barbour 675

Fig. 11. Zones of soil-water instability conditions for montmoril-lonite (after Ingles and Aitchison 1969).

ParameterNo. oftests

Average andstandard deviation

Ec (dS/m) 10 6(3)TDS (mequiv./L) 10 79(38)%Na 10 52(17)SAR 10 10(5)Mg/Ca 10 2(1)

Table 6. Soil pore-water chemistry obtained fromcompacted clay liner and embankment in MR3.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:56 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 12: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

and Altshaeff 1977). Ingles and Aitchison recommend theuse of compaction wet of optimum and design proceduresthat protect the clay from drying (shrinkage) which wouldproduce macropores and the potential for piping. Sherardand Decker (1977) also note that compaction may have somebenefit. The field evidence from the compacted sections ofMR3 suggests that compaction does have a significant bene-fit in controlling swelling–softening and dispersion.

Studies are currently being planned to evaluate the mecha-nism for this increased stability; however, the increase islikely the result of the increased density and strength. Theclay may initially also have a decreased hydraulic conductiv-ity due to compaction, but it would likely undergo an in-crease in hydraulic conductivity as it undergoes freeze–thawand desiccation. The weathering of the clay liner did pro-duce a nugget-like soil structure in the upper 300–600 mmof the compacted clay. This structure may result in theweathered zone of the compacted clay having less ability totransmit water due to capillarity than the natural clays. Thisin turn would minimize the movement of salts into the sur-face soils as the result of evaporation. It was observed thatclays located on the excavated canal slopes did not exhibitthe same degree of nugget-like structure from weathering. Itis felt that nugget-like structure observed in these clays hadpartially healed from swelling and dispersion.

Repair optionsContractual obligations required that the canal be restored

to a new condition before being transferred to the BloodTribe, since the work was still under warranty. Although ob-servations showed that sloughing occurred over a 7500 msection of canal, it was only near continuous over a 2500 msection of canal in MR3 and that part of MR4 located inclay-rich colluvium where sloughing had occurred on anearly annual basis. Random sloughing had occurred as aone-time event in the remainder of the canal, with no furtheroccurrences in the past 6 years. Based on the results of theinvestigation and observations, it was decided to repair onlythe 2500 m section of canal located in MR3 and MR4, withlocal repairs in the remainder of the canal. This resulted in acost saving of about 2 million dollars for canal reconstruction.

A number of repair options were considered, including aplastic geomembrane, surcharge loading, cation exchange,grassing the side slopes, and a compacted clay liner. Repair,in all cases, except for the cation exchange option and seed-ing the slopes to grass, would have required removal and re-placement of the existing armour gravel and soft subgrade toa minimum depth of 1 m or more. Reconstruction was car-ried out in the fall of 1997 and spring of 1998 using a com-pacted clay liner. In some isolated pockets, up to 2.5–3.0 mof clay subgrade was removed due to very soft consistency.

Consideration was given to placing a 20 mil (1 mil =25.4 µm) plastic geomembrane directly under the armourgravel to prevent the development of surficial salt bycapillarity. The geomembrane would prevent upward move-ment of salt water by capillarity and downward infiltrationof fresh water into the underlying subgrade, thus reducingswelling of the subgrade. Although a geomembrane couldprevent future swelling of the subgrade, a risk exists that theintegrity of the geomembrane and its ability to prevent waterfrom seeping into subgrade from various points of entry

would not be satisfactory. Even a small amount of swellingcould eventually destroy the membrane.

Surcharging the subgrade with gravel to counteract swell-ing was given consideration, since this method has beenused successfully on other Prairie Farm Rehabilitation Ad-ministration (PFRA) constructed projects. A series of swelltests was planned to determine the surcharge load to coun-teract swelling pressures. This was a difficult task consider-ing the spatial variability of the sodium and magnesiumlevels and clay mineralogy of the clays encountered alongthe canal. The swell testing program was consequently aban-doned when a previous repair was investigated. A 25 mlength of a failed cut slope had been repaired by replacingthe subgrade with up to 750 mm of gravel. The slopeshowed signs of distress after 4 years of service. The repairsuggested that possibly up to 1 m ormore of gravel could berequired to counteract the swelling pressure. The cost of thisrepair option was too expensive.

Chemical treatment of the soil through cation exchangewith lime could be relied upon to initially stabilize thesubgrade and prevent dispersion and swelling. However,lime treatment was judged to be a short-term solution, sincecontinual upward migration of sodium and magnesiumwould eventually result in replacement of the calcium. Theinvestigation showed there was a limitless supply of Na saltsdeep into the subgrade.

Observations showed that grass-covered cut slopes ap-peared stable with no evidence of salt loading on the groundsurface. This is in contrast to bare slopes above the armourgravel that failed. The root system apparently functions as acapillary break, preventing salt transport to the surface.Grassing was considered viable, but contract obligations re-quired that the canal side slopes be lined with armour gravel.In addition, there were concerns about increased mainte-nance costs and the possibility of the grass dying when sub-ject to sustained flooding during canal operation.

Observations showed that compaction of salt-rich claycolluvium with a sheepfoot roller resulted in stable canalslopes. The existing compacted clay liner showed no evi-dence of slope instability or salt loading after 6 years of ser-vice prior to reconstruction of MR3 and MR4. These slopesare still stable today (9 years later). It was decided to repairthe failed canal sections with compacted clay based on theperformance of the existing clay liner. Repairs were carriedout by overexcavating the canal to obtain a compacted zoneof clay that was at least 1 m in thickness normal to the canalside slopes and invert. The clay liner was constructed byplacing soil in 200 mm, uncompacted, horizontal lifts andcompacting with a sheepfoot roller with a minimum contactpressure of 2400 kPa. The results of 88 quality-control testsindicated a final compaction of 98.9% (±3.8%) of StandardProctor maximum dry density with a final water content1.6% (±2.4%) wet of optimum. The clay was to be com-pacted to at least 95% of Standard Proctor maximum drydensity with a water content ranging from 0 to 2% wet ofthe optimum water content.

Conclusions

Sloughing of cut slopes within a canal constructed in salt-rich, clayey colluvium in southern Alberta has been investi-

© 2001 NRC Canada

676 Can. Geotech. J. Vol. 38, 2001

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:56 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen

Page 13: A case history of shallow sloughing within cut slopes of ...engr · A case history of shallow sloughing within cut slopes of an irrigation canal in salt-rich clayey colluvium R.S.

© 2001 NRC Canada

Globa and Barbour 677

gated. The dispersive nature of the clays was effectivelycharacterized through the use of soil chemistry, clay miner-alogy, and pinhole dispersion testing. Sloughing appeared tobe concentrated in a colluvial slope where historically salin-ity development had developed as a result of a shallow watertable. Further development of this salinity was enhanced byconstruction of the canal. Compaction was observed to beeffective in preventing salt loading and swelling and soften-ing of the side slopes. Approximately 2500 m of the canalwas reconstructed in MR3 and MR4 which was located inthe clay-rich colluvium. Performance of the reconstructedsection has been satisfactory to date.

Acknowledgements

The authors would like to express their thanks to Dr.Ahmet Mermut, Department of Soil Science, University ofSaskatchewan, for his generous advice and to the PrairieFarm Rehabilitation Administration for permission to pub-lish this case history.

References

Barbour, S.L., and Yang, N. 1993. A review of the influence ofclay–brine interactions on the geotechnical properties of Ca-montmorillonitic clayey soils from western Canada. CanadianGeotechnical Journal,30: 920–934.

Emerson, W.W. 1979. Aggregate stability.In The encyclopedia ofsoil science. Part 1. Physics, chemistry, biology, fertility andtechnology. Edited by R.W. Fairbridge and C.W. Finkl, Jr.Dowden, Hutchinson and Ross, Stroudsburg, Pa. pp. 22–24.

Ingles, O.G., and Aitchison, G.D. 1969. Soil-water disequilibriumas a cause of subsidence in natural soils and earth embankments.

In Land Subsidence, Proceedings of the Tokyo Symposium,September 1969, International Association of Scientific Hydrol-ogy, United Nations Economic, Social and Cultural Organisa-tion, Tokyo, Japan, Vol. 11, pp. 342–352.

Janzen, H.H. 1993. Chapter 18 Soluble Salts.In Soil sampling andmethods of analysis.Edited byC.M. Carter. Canadian Society ofSoil Science, Lewis Publishers, Boca Raton, pp. 161–166.

Landau, H.G., and Altshaeff, A.G. 1977. Conditions causing pipingin compacted clay.In Dispersive clays, related piping and ero-sion in geotechnical projects.Edited byJ.L. Sherard and R.S.Decker. American Society for Testing and Materials, SpecialTechnical Publication STP 623, pp. 240–259.

Mermut, A. 1997. Clay mineralogy analysis. Department of SoilScience, University of Saskatchewan, Saskatoon, Sask.

Mitchell, J.K. 1993. Fundamentals of soil behaviour. John Wiley &Sons, Inc., New York.

Sherard, J.L., and Deckar, R.S. 1977. Summary–An evaluation ofsymposium on dispersive clays.In Dispersive clays, related pip-ing and erosion in geotechnical projects.Edited byJ.L. Sherardand R.S. Decker. American Society for Testing and Materials,Special Technical Publication STP 623, pp. 467–480.

Sherard, J.L., Dunnigan, L.P., and Decker, R.S. 1976. Identificationand nature of dispersive soils. Journal of the Geotechnical Engi-neering Division, ASCE,102(GT4): 287–301.

Sherard, J.L., Dunnigan, L.P., and Decker, R.S. 1977. Some engi-neering problems with dispersive clays.In Dispersive clays, re-lated piping and erosion in geotechnical projects.Edited byJ.L.Sherard and R.S. Decker. American Society for Testing and Ma-terials, Special Technical Publication STP 623, pp. 3–13.

Stalker, A. 1963. Surficial geology of the Blood Tribe Indian Re-serve. Geological Survey of Canada, Paper 63-25.

St. Arnaud, R.J. 1979. Nature and distribution of secondary soilcarbonates within landscapes in relation to soluble Mg/Ca ratios.Canadian Journal of Soil Science,59: 87–98.

I:\cgj\Cgj38\Cgj04\T01-003.vpThursday, July 26, 2001 9:25:56 AM

Color profile: Generic - CMYK US Negative ProofingComposite Default screen


Recommended