The effect of compaction and shear deformation of saturated soil on hydraulic conductivity

27 July 2021

AperTO - Archivio Istituzionale Open Access dell'Università di Torino

The effect of compaction and shear deformation of saturated soil on hydraulic conductivity / W.R. Whalley; G.P.

Matthews; S. Ferraris. - In: SOIL & TILLAGE RESEARCH. - ISSN 0167-1987. - 125(2012), pp. 23-29.

Original Citation:

The effect of compaction and shear deformation of saturated soil on hydraulic conductivity

Published version:

DOI:DOI: 10.1016/j.still.2012.05.020

Terms of use:

Open Access

(Article begins on next page)

Anyone can freely access the full text of works made available as "Open Access". Works made available under a

Creative Commons license can be used according to the terms and conditions of said license. Use of all other works

requires consent of the right holder (author or publisher) if not exempted from copyright protection by the applicable law.

Availability:

This is the author's manuscript

The

effect

of

compaction

and

shear

deformation

of

saturated

soil

on

hydraulic

conductivity

W.R.

Whalley

a,

*

,

G.P.

Matthews

b

,

S.

Ferraris

c a_{RothamstedResearch,Harpenden,HertfordshireAL52JQ,UK}

b

EnvironmentalandFluidModellingGroup,UniversityofPlymouth,Plymouth,DevonPL48AA,UK

c

InteruniversityDepartmentofRegionalandUrbanStudiesandPlanning,Universita` degliStudidiTorinoePolitecnicodiTorino,vialeMattioli39,Torino10125,Italy

1. Introduction

When soil is damaged in the field, because of traffic by agricultural machinery, it is often referred to as ‘‘compacted’’. However,inpracticeinthetopsoil,so-calledcompactionevents areusuallyacombinationofdensificationandsheardeformation.

Greenetal.(2003)havereviewedtheeffectsofvarioustypesofsoil deformation on hydraulic conductivity. Their conclusion, in agreement with many other studies (Matthews et al., 2010; Reynoldsetal.,2008;WangandHuang,1984),isthatsaturated hydraulic conductivityis very sensitive tochanges in porosity. Manyoftheeffectsofsoiltypeandcompactioncanbetakeninto accountwiththeuseofaneffectiveporosity.Carman(1939)was oneofthefirsttoproposethisapproachtoallowthesaturated hydraulicconductivityksatinclaysandsandstobedescribedbya

singleequation.TheKozeny–Carman(KC)equationiswidelyused to explore the effects of changes in porosity on hydraulic conductivity.Itiswrittenas:

ksat¼

t

1 S2

r

2 wg

nr

2 s e3 1þe (1)

whereSisthespecificsurfacearea,

t

isatortuosityfactor,gisthe accelerationduetogravity,

r

wand

r

sarethedensityofwaterand

soilrespectively,

n

isthedynamicviscosityofwaterandeisthe voidratio(note:e=

f

/(1

f

),where

f

isporosity)(seeChapuis and Aubertin, 2003). However, Chapuis and Aubertin (2003)

observed that the KC equation is rarely used because of the difficultyofmeasuringtheparameters,especiallyS.Anadditional complicationisthat

t

isaconstantthattakesaccountofdifferences intortuositybetweendifferentsoils,andifksatistobepredicted

t

needs to be known. However, Carman (1939) expected that

t

wouldonlyhaveanarrowrangeofvalues.

Therelationshipbetweensoilporosityandhydraulic conduc-tivityisoftendescribedbyanempiricalpowerlaw(Ahujaetal., 1989;Aminrunetal.,2004;Greenetal.,2003):

ksat¼B

f

ne (2)

whereBandnarefitted,and

f

eistheeffectiveporositywhichis

definedasthedifferencebetweenthewatercontentatsaturation

u

sandatamorenegativematricpotential.Thereisageneralview

thatEq.(2)isaKozeny–Carmenderivedrelationship,althoughthe resemblanceisnotverystrongasfarastheformoftheequationis concerned.Infactthefunctionalrelationshipisdifferenttotheone describedbyCarman(1939).Typicallynisassumedhaveavalue between4and5,andBmayvarywithsoiltype(Ahujaetal.,1989). The approach recognises that the hydraulic conductivity of saturated soil depends mainly on the larger voids (i.e. those affecting the saturated end of the water retention curve). The choiceofthelowervalueofmatricpotential,usedtoestimate

f

e,is

ARTICLE INFO Keywords: Consolidation Criticalstate Claysoil Empiricalmodels ABSTRACT

Thispaperdescribeshowinaclaysoil,consolidationandthensheardeformationataconstantporosity

affectthehydraulicconductivityofthesaturatedsoil.WeusedaBishopandWesleytriaxialcellto

consolidatethesoilalongthenormalconsolidationlineandthentoshear-deformthesoilataconstant

porositytothepointwherethecriticalstateconditionhadbeenreached.Therelationshipbetween

hydraulicconductivity(ksat)andsoilporosityforsoilconsolidatedonthenormalconsolidationlinewas

similarto previouslypublisheddata. However,sheardeformation ofsoilwhenheldataconstant

porositygreatlyreducedksatespeciallyathighporosity,whereksatwasreducedto5%ofitsoriginalvalue.

Indensesoiltheeffectsofsheardeformationonksatweresmaller.Weusedpreviouslypublishedwater

releasedataforthevariouslycompactedandsheardeformedsoiltoestimatewaterreleasecurvesforthe

soilinourexperiment. Weshowedthatanempiricalmodel topredictksatgaveagoodfit toour

experimentaldatacollectedinthelaboratory.Wetestedtheempiricalmodelonawidersetoffielddata

obtainedfromtheHYPRESdatabase.

ß2012ElsevierB.V.Allrightsreserved.

*Correspondingauthor.

E-mailaddress:richard.whalley@rothamsted.ac.uk(W.R.Whalley).

ContentslistsavailableatSciVerseScienceDirect

Soil

&

Tillage

Research

j o urn a l hom e pa g e :ww w . e l se v i e r. c om / l oca t e / st i l l

0167-1987/$–seefrontmatterß2012ElsevierB.V.Allrightsreserved.

arbitrary. Ahuja et al. (1989) recommend 33kPa whereas

Aminrun et al. (2004) suggest that a more negative value of 66kPamayworkbetterforsoilwithhigherclaycontent.Han etal.(2008)estimated

f

efromthepointofinflexionofthewater

retention curve, which was determined from a fit of the Van Genuchten function (van Genuchten, 1980). With this general approach, it does seem possible to obtain an approximately commonrelationshipbetweenksatand

f

efordifferentsoilsaswell

asdescribetheeffectsofsoildamagebycompactiononksat(Green et al., 2003). However, individual soil types have systematic variationsfromtheaveragecurvefittedtoanumberofsoils(see Fig.4ainAhujaetal.,1989).AssoulineandOr(2008)outlinehow, inadditiontoporosity,theairentrypotentialcanbeusedinan improvedapproachtotakeaccountoftheeffectsofcompactionon hydraulicconductivity.

Matthews et al. (2010) have proposed a simple model for saturated hydraulic conductivity using the water retention characteristic.They considered theinterconnectionof cylinders ofdifferentsizesacrossasectionofaporousmaterial.Inaspecial caseitwasassumedthattheseinterconnectionswererandom.A distributionF(r)ofcylindricalporeradiirwasderivedfromthe waterretentioncurve(

u

,

c

)as:

FðrÞ¼d

u

dr¼ d

c

dr d

u

d

c

(3)

where

c

isthematricpotential.Assumingthattheflowofwater througha cylinderwas proportionaltoits cross-sectional area, then, ksat/½ Z 1 0 rFðrÞdr 2 (4)

Commonly this expressionis incorporatedinto a model for relative permeability or unsaturated hydraulic conductivity, appearingasadenominatorinsuchamodel(Mualem,1976).

A constantofproportionality isrequiredtoachieve absolute valuesofksatfromEq.(4)whichmaybewrittenas:

ksat/ Z 1 0 1

c

d

u

" #2 (5)

Matthewsetal.(2010)usedthevanGenuchtenfunctionforthe waterretentioncharacteristic(vanGenuchten,1980)toevaluate thisintegralandusedthestandardconstraintm=11/nforthe twoshapeparametersofthewaterretentioncharacteristic.

Withthechangeofvariable

Q

¼

u



u

r

u

s

u

r

(6)

where

u

sand

u

raretherespectivesaturatedandresidualwater

contents,Eq.(5)became:

ksat/ð

u

s

u

rÞ2 Z 1 0 1

c

d

Q

" #2 (7)

whichwasrewrittenas:

ksat/

a

2ð

u

s

u

rÞ2 Z 1 0

Q

1=m 1

Q

1=m !1m d

Q

2 4 3 5 2 (8)

where

a

wasthethirdfittingparameterfortheVanGenuchten function.

TheintegralinEq.(8)equatesto1anditcanbewrittenas: ksat/

a

2ð

u

s

u

rÞ2 (9)

Aconstantofproportionalityneededtoachieve absolutevalues maybeinsertedinEq.(9)togive:

ksat¼

r

w g 8

v

2

g

r

wg  2

a

2 ð

u

s

u

rÞ2 (10)

where

n

istheviscosityofwaterand

g

isthesurfacetension of water.Hereweassumenorequirementforacorrectionfactorto accountfortortuosity.Eq.(10)simplifiesto:

ksat¼

g

2

2

vr

wg

a

2_ð

_u

s

u

rÞ2 (11)

Matthews et al. (2010) explored the effects of isotropic compression on thehydraulic conductivity of soil. They found thatbothEq.(11)andaporenetworkmodelgavegoodpredictions of the effects of consolidation on the change in hydraulic conductivity, although the absolute predicted values were in someerror.

Soildeformationscanbemorecomplexthanmerelyanincrease insoildensityandmayincludetheeffectsofsheardeformation. The critical state description of soil behaviour allows the relationships between compression,shear deformation and soil porositytobedefined.Duringcriticalstatedeformationthesoilis completelydestructured(MitchellandSoga,2005)totheextent thatthereisnoevidenceofanyaggregationandtheentirepore spacecanbeassumedtobetextural.Atthecriticalstate,theshear stressanddensityremainconstantwhileshearstrainincreases. Dependingontheinitialstateofthesoil,itsporositywilleither increaseordecreaseasitissheardeformedtoapproachthecritical state.KirbyandBlunden(1991)consideredtheeffectofboth of thesetransitionstothecriticalstateonsoilhydraulicproperties. Innewlycultivatedsoilsitislikelythatporositywilldecrease during shear. We were interested to discover if relationships betweensoilporosityanditshydraulicconductivityforsaturated soilweredifferentfornormallyconsolidatedsoilcomparedwith soil which had beenshear deformed. Althoughtheeffects of a changeinporosityonhydraulicconductivityarewidelyreported, there are very little published data on the effect of shear deformationon soil hydraulicconductivity(Kirbyand Blunden, 1991).

2. Materialsandmethods

2.1. Soilsampleandpreparation

WeusedRowdensoil(Gregoryetal.,2010a,b;Matthewsetal., 2010)whichisaclaysoilfromNorthWyke,Devon,UK.Basicsoil

Table1

Basicsoilproperties.

Location NorthWyke,Devon

Field Rowden

Gridreference GBNationalGrid SX652994

Longitude 03,54,50W

Soiltype Latitude 50,46,43N

SSEWgroup Stagnogleysoil

SSEWseries Hallsworth

FAO GleyicLuvisol

Landuse Permanentgrass

Sand2000–63mm gg1 drysoil 0.147 Silt63–2mm gg1 drysoil 0.396 Clay<2mm gg1 drysoil 0.457

Texture SSEW Clay

Particledensity gcm3 _2.466

Organicmatter gg1_{drysoil 0.076}

W.R.Whalleyetal./Soil&TillageResearch125(2012)23–29 24

propertiesweremeasuredbystandardmethods(Table1).Theair driedsoilwascrumbledtopassa4mmsieveandlefttoair-dryina singlebulked sample.Beforeusethesoil wasequilibratedtoa watercontentof0.55gg1_{byaddingwaterandstoringinafridge.}

Thesoilwaspackedintoamould38mmindiameterand78mm long in thin layers, approximately 1cm deep. Each layer was compressed uniaxiallywitha pneumaticpress at a pressureof 10kPa.

2.2. Measuringksatinconsolidationandsheardeformedsoil

Weuseda Bishopand Wesleytriaxialcell(Ngand Menzies, 2007)todeformthesoilandmeasurehydraulicconductivity(see

Fig. 1). Once in the triaxial cell, the soils were saturated by increasingthepressureofthewaterconfiningthesoilsampleand thepressureofthesoilwaterto600and590kParespectivelyover aperiodof24h. Thesoilswerethenconsolidatedtoarangeof effectivestressesbetween10(initialcondition)andapproximately 500kPa. Isotropic compression was achieved by adjusting the relativevaluesoftheporepressureandcellpressuretoincrease the effective stress on the saturated soil sample. The normal consolidationlinewasdeterminedfromanumberofindependent consolidation tests. Following consolidation of samples, shear

deformationwasappliedbyraisingthepedestal(seeFig.1).During sheardeformation,thevolumeofthesamplewaskeptconstantby fixingthevolumeofthepore-pressurepressurevolumecontroller andthechangeintheporepressurewasmonitored(seeFig.1). During shear deformation,thesoilsample shape changedfrom cylindrical(attheendofnormalconsolidation)tobarrel-shaped (attheendofsheardeformation).

Thetriaxial cellweusedhad anadditional pressurevolume controllerconnectedtothebaseofthesample,sowewereableto measurethefluxofwaterthroughasampleatanygivenhydraulic gradient. We applieda hydraulic gradient of20kPa acrossthe sample as described by Matthews et al. (2010). The hydraulic conductivityofthecylindricalsampleswasreadilycalculated.

To convert the flux of water measured through the barrel-shapedshear-deformedsamplesintoahydraulicconductivitywe assumed that the soil had thediameter ofa cylinderwiththe volumeofthebarrelandwiththesameheight.Todeterminethe magnitude of any error arising from the assumption of an equivalentvolume cylinder,theDarcy equation wasintegrated inanaxialsymmetricaldomainbyafinitevolumecode(Manzini andFerraris,2004)fortwogeometries;abarrelshapedsampleand acylinderofthesameheightwithanequivalentvolumediameter. Foragivenvalueofksat(1.14cm/s)andhydraulicgradient(20kPa)

Fig.1.BishopandWesleycellusedtoconsolidateandshearthesoilandthentomeasurethesaturatedhydraulicconductivityofsoil.Ahydraulicgradientcanbeappliedby adjustingthepressureofthewateratthetopofthesampleP1relativetothepressureofthewateratthebaseofthesampleP2.

the modelled fluxes from these two geometries were similar (30.4cm3/h for the barrel compared to 34.1cm3/h for the equivalentdiametercylinder).

2.3. Useofpublisheddata

Empirical models to predict ksat usually require the water

retentioncharacteristictobeknowntosomeextent.Toestimate thewaterretentioncharacteristiccurvesofthedeformedsoilsin thispaperweusedthedatapublishedbyGregoryetal.(2010a). Theyreportedwaterrelease characteristicdataforRowdensoil whichwascompresseduniaxiallyandalsosoilsheardeformed,by remoulding,whichwasfittedtothefollowing:

u

¼

u

s

u

r ð1þð

f

b

c

ÞnvÞ1ð1=nvÞ 2 4 3 5 þ

u

r (12)

where

f

isthesoilporosityandbisaparameterthatdependson soildamage.Parameters

u

s,

u

r,

a

(here

a

=

f

b)andnvarethoseof

thevanGenuchten function.The valueof bincreaseswithsoil damage.We assumedthat remouldingofsoil byGregoryet al. (2010a,b)hadthesameeffectsassoildeformationatthecritical stateandhencethatthevaluesofbforremouldedsoilappliedto soilatthecriticalstate.

Totestempiricalmodelsforksatmorewidelyweusedhydraulic

conductivitydataforsaturatedfromtheHYPRESdatabasealong withestimatesof theVan Genuchten parameters for thesame soils.Themostcomprehensive(‘L’)datasetoftheHYPRESdatabase (Wo¨stenetal.,1999)wasused.Theextracteddatawererestricted totopsoils(notallwerecultivated)andsubsoils,todepthsof2.3m. Buried,organicorambiguoushorizonswereomitted.

3. Resultsanddiscussion

3.1. Soildeformation

The normal consolidation curve is shown in Fig.2A. This is consistentwithdatapublishedforthesamesoilbyMatthewsetal. (2010). Thecritical stateline is alsoshown inFig.2Aand it is consistentwithpreviouslypublisheddataforclaysoils(O’Sullivan andRobertson,1996)inthatitsslopeisslightlysteeperthanthatof thenormal consolidation line. In Fig. 2B thedeviator stress is plottedagainstthemeanstress.Thedeviatorstressvarieslinearly withthemeanstress.Inallcasessheardeformationresultedin visiblebarrellingofaninitiallycylindricalsoilsample(seeMitchell andSoga,2005).

3.2. Deformationandhydraulicconductivity

Thehydraulicconductivityofsoilisplottedagainstporosityin

Fig.3.Fornormallyconsolidatedsoilthedataaretypicalofwidely reportedcorrelations betweenconsolidation andhydraulic con-ductivity(e.g.Matthewset al.,2010). When thesoilwasshear deformed to the critical state, ksat was smaller for any given

porositycomparedtoitsinitialvaluewhenthesoilwasnormally consolidated(seeFig.3).Wefitted

log10ksat¼R

f

þQ (13)

to the data in Fig. 3 where

f

is porosity. The values of the

parameters, R and Q are given in Table 2, together with the standarderrors(SE).Thefittedcurvesexplained90percentofthe variance(P<0.002)inksatforthenormallyconsolidatedsoiland

85percentofthevariance(P<0.005)inksatforsoilatthecritical

state.

Inthisclaysoilataporosityof62percent,sheardeformation reduced ksat to 5 percent of its original value. In the most

compactedsoiltheeffectofsheardeformationonksatwassmaller.

IfthefittedlinesinFig.3areextrapolatedtolowerporosities,they meetat56percent.Therefore,ifsoilataporosityofaround56%, wassheardeformed,theeffectonksatislikelytobeverysmall.

Fig.2.(A)Porosityasafunctionofthelogarithmofmeaneffectivestressforsoil which iseither normallyconsolidated(open symbols)orinthe criticalstate condition(closedsymbols).(B)Deviatorstressplottedagainstthemeanstressfor soilwhichhasbeensheardeformedtotheextentthatitisinthecriticalstate condition.

Table2

FittedvaluesofRandQinEq.(13)andtheirstandarderrors.

Soilcondition Parameter PercentageofvarianceaccountedforandPvalue

R SE Q SE

Normallyconsolidated 0.4836 0.0683 29.35 4.20 90.8,P<0.002

Criticalstate 0.3024 0.0547 19.22 3.36 85.5,P<0.005

W.R.Whalleyetal./Soil&TillageResearch125(2012)23–29 26

3.3. Empiricalpredictionsofksatbasedonthewaterrelease

characteristic:laboratorydata

Mostempiricalmethodsusedtopredicthydraulicconductivity, includingEqs.(2)and(9),requiresomeinformationfromthewater release characteristic. We used published data for variously damagedRowdensoil(Gregoryetal.,2010a)whicharelistedin

Table3toestimatewaterreleasecharacteristicswithEq.(12)(see

Fig.4).Theseestimatedwaterreleasecharacteristicswereusedto estimate

f

e and

a

(

u

s

u

r) of the loose soil, dense soil, shear

deformedloosesoilandsheardeformeddensesoil(seeTable3). The effective porosities were determined from the difference betweenthesaturatedwatercontentsandthewatercontentsata matricpotentialof33kPa(seeTable3).

Thevaluesofksatcorrespondingtothoseporositiesofsoilsfor

whichwater releasedatawereavailableweredeterminedwith Eq.(13)andRandQvaluesgiveninTable2.Thefittedvaluesofthe exponentandconstantinEqs.(2)and(9)aregiveninTable4(see

Fig.5). ThevalueoftheexponentinEq.(2)was4.04,similarto thosepreviouslyreported(Ahujaetal.,1989;Greenetal.,2003), althoughMavkoandNur(1997)observedthatthenexponentin Eq.(2)isanadhocsolutiontotheproblemofempiricallyreducing ksatwithporosityandthatnvariesfrom3to8dependingonsoil.

ThefittedexponentinEq.(9)was1.651anditwasnotstatistically significantly differentto the expected value of 2 (Table 4 and

Fig.5B).

3.4. Empiricalpredictionsofksatbasedonthewaterrelease

characteristic:fielddata

WeexploredthepossibilityforthemoregeneraluseofEq.(9)

usingdatafromawiderrangeofsoilsfromtheHYPRESdatabase. InFig.6thefittedksatdataareplottedagainstthemeasuredvalues

withtherelationship:

log10ksat¼2log10ð

a

ð

u

s

u

rÞÞþ4:119: (14)

We kept theexponentatits theoreticalvalueof2 and only adjustedthelinearconstantduringthefittingprocess.

Thelaboratorydatacollectedinthisstudyandfielddatafrom HYPRESdidnotseemtobeingoodagreement.Fig.6showsthat Eq. (14) overestimated the saturated hydraulic conductivities measuredinthisworkbyupto2½ordersofmagnitude,although theerrorswerewithinthescatteroftheHYPRESdata.However, Eq.(14)correctlypredictedthetrendinthemeasureddata.The differences could be attributed to the fact that oursoils were

Fig.4.Estimatedwaterreleasecurvesfornormallyconsolidatedlooseandcompact soilaswellassheardeformedlooseandcompactsoil.Thesecurveswereestimated usingparametersofEq.(12)inTable3forvariouslycompactedandsheardeformed RowdensoilpublishedbyGregoryetal(2010).

Table4

ValuesoftheparametersfittedtoEqs.(2)and(9)usingthedatainFig.5.

Equation UnitsofKs Unitsofspecified

parameter

Fittingparameter Valueoffitting parameter

Percentageofvariance accountedforandPvalue

ksat¼Bfne mmday1 B 348,096 89.5%,P=0.036

n 4.04

ksat¼CavðusurÞv mmday1 a(kPa1) C 243 97.5%,P=0.023

v 1.651a

a

Thisexponentiswrittenas2inEq.(9).Herethefittedvalueisgiven. Table3

ParametersderivedfromtheestimatedwaterreleasecharacteristicsshowninFig.4alongwiththeksatdatacalculatedfromtheappropriateporosity,i.e.usthesaturated

watercontentwithEq.(12).

Soilcondition Effectiveporosity usur b us nv a

i.e.fb

a(usur) ksat(mmday1calculatedwithEq. (12)forappropriateporosityus)

Loose 0.135 0.451 1.25 0.650 1.187 1.713 0.7727 121.388

Looseandsheardeformed 0.030 0.205 6.72 0.650 1.187 0.0554 0.0114 0.0624

Compact 0.045 0.303 4.04 0.582 1.187 0.1123 0.0339 2.7289

Compactandsheardeformed 0.014 0.138 6.71 0.582 1.187 0.0264 0.0036 0.0239

Fig.3.Thelogarithmofhydraulicconductivityfornormallyconsolidatedsoil(open symbols)andforsoilinthecriticalstate(closedsymbols)plottedagainstporosity. ThedataarefittedtoEq.(13)andthefittedparametersaregiveninTable2.

repackedwhereasthesoilsusedtogeneratetheHYPRESdatabase wereundisturbed,orthatweusedestimatedandnotmeasured waterreleasedata(Fig.4).NeverthelessitisclearthatEq.(14), withonly onearbitraryfittingparameterandotherparameters obtainable from the water retention characteristic, provides a

usefulfittotheHYPRESdataandtheexperimentaldataobtained fromrepackedsoil.

3.5. Theimplicationsofourfindings

Ourdataimpliesthatsoilmacrostructureinloosesoilisvery sensitivetodamagebysheardeformation.Inthefield,isotropic consolidationislikelytobearareoccurrence.Uniaxial compres-sion is often used in the laboratory to replicate the effects of compaction (Gregory et al.,2006), but this is a combination of consolidationandsheardeformationandwillprobablyresultina ksatforanygivenporositythatisbetweenthelimitsofthosevalues

for‘‘normalconsolidation’’andthe‘‘criticalstate’’showninFig.3. Thetwoempiricalmodelsforksatappearedtobeeffectiveat

describingtheeffectsofcomplexdeformationsonksat.However,

the requirement for prior knowledge of the water release characteristic makestheseempirical approachesdifficulttouse inpractice.Nevertheless,wehaveshownthatthereisscopetouse relativelysimpleempiricalapproachestodescribetheeffectofsoil deformationonthewaterreleasecharacteristicwhichcanthenbe usedtoexplainhowksatvarieswithsoildeformation.Theeffective

porosity approach (Eq. (2)) required the water content at saturation and one other more negative matric potential. The useofEq.(9)wasfurthercomplicatedbecausethevalueof

a

also neededtobeknownorestimated.Whenwaterreleasedatawere available, a relatively simple model for ksat (Eqs. (9) and (14))

appearedtobeusefuloverawiderangeofsoilswhichhadbeen variouslydamaged.

4. Conclusions

Sheardeformationofloosesoilataconstantporositycanresult inalarge reductioninthesaturatedhydraulicconductivityksat.

Reductionsinksatfollowingcompactionand/orsheardeformation

arepredictableiftheeffectofsoildeformationonthewaterrelease characteristiciseitherknownorcanbeestimated.Wehavetested anempiricalequationtoestimateksatusingtheHYPRESdatabase,

andestablisheditsutilityforpredictingksatinthefield.

Acknowledgements

ThisworkwasfundedbyagrantfromtheUKEngineeringand PhysicalSciencesResearchCouncil(EPSRC)(EP/H040064/1)and ItalianGovernmentgrantPRIN2007‘‘Experimentalmeasurement of the soil-vegetation-atmosphere interaction processes and numerical modellingof their response toclimate change’’. The facilities at Rothamsted Research were funded by the UK BiotechnologyandBiologicalSciencesResearchCouncil(BBSRC). We are indebted to Allan Lilly of the James Hutton Institute, AberdeenwhoextractedhydraulicdataofsoilfromtheHYPRES databaseon27thJuly2011.

References

Ahuja,L.R.,Cassel,D.K.,Bruce,R.R.,Barns,B.B.,1989.Evaluationofspatial distribu-tionofhydraulicconductivityusingeffectiveporositydata.SoilScience148, 404–411.

Aminrun,W.,Amin,Ms.M.,Eltaib,S.M.,2004.Effectiveporosityofpaddysoilsasan estimateofitssaturatedhydraulicconductivity.Geoderma121,197–203. Assouline,S.,Or,D.,2008.Airentry-basedcharacteristiclengthforestimationof

permeabilityofvariablycompactedearthmaterials.WaterResourcesResearch 44,W11403,http://dx.doi.org/10.1029/2008WR006937.

Carman,P.C.,1939.Permeabilityofsandssoilsandclays.JournalofAgricultural Science29,262–273.

Chapuis,R.P.,Aubertin,M.,2003.OntheuseontheKozeny–Carmanequationto predictthehydraulicconductivityofsoils.CanadianGeotechicalJournal40 616-613.

Green,T.R.,Ahuja,L.R.,Benjamin,J.G.,2003.Advancesinpredictingagricultural managementeffectsonsoilhydraulicproperties.Geoderma116,3–27. Fig.5.Thelogarithmofhydraulicconductivityksatplottedagainstthelogarithmof

effectiveporosity(panelA)andthelogarithmofa(usur)(panelB).Boththe

effective porosity and a(usur) were determined form the water release

characteristicsshowninFig.4(seeTable3).ThefittedparametersofEq.(2)for thetoppanelandEq.(9)forthebottompanelaregiveninTable4.

Fig.6.Fittedksatplottedagainstmeasuredksat(onalogscale).Opencirclesaredata

fromtheHYPRESdatabase.FitteddatawereobtainedfromEq.(14).Alsoplottedare thedatapredictedwithEq.(14)forthecompactedandshearedsoilsamplesinthis study(closedcircles)usingthefittingparameterobtainedfromtheHYPRESdata.

W.R.Whalleyetal./Soil&TillageResearch125(2012)23–29 28

Gregory,A.S.,Whalley,W.R.,Watts,C.W.,Bird,N.R.A.,Hallett,P.D.,Whitmore,A.P., 2006.Calculationofthecompressionindexandtheprecompressionstressfrom soilcompressiontestdata.SoilandTillageResearch89,45–57.

Gregory,A.S.,Bird, N.R.A., Whalley, W.R., Matthews, G.P.,Young, I.M., 2010a. Deformationandshrinkageeffectsonthesoilwaterreleasecharacteristic.Soil ScienceSocietyofAmericaJournal74,1104–1112.

Gregory,A.S.,Bird,N.R.A.,Whalley,W.R.,Matthews,G.P.,2010b.Estimatingrelative hydraulicconductivityfromthewaterreleasecharacteristicofashrinkingclay soil.SoilScienceSocietyofAmericaJournal74 1113-1113.

vanGenuchten,M.T.,1980.Aclosed-formequationforpredictingthehydraulic conductivityofunsaturatedsoils.SoilScienceSocietyofAmericaJournal44, 892–898.

Han,H.,Gimenez,D.,Lilly,A.,2008.Texturalaveragesofsaturatedsoilhydraulic conductivity predicted from water retention data. Geoderma 146, 121–128.

Kirby,J.M., Blunden,B.G.,1991.Interactionofsoildeformations,structureand permeability.AustralianJournalofSoilResearch29,891–904.

Mavko,G.,Nur,A.,1997.TheeffectofapercolationthresholdintheKozeny–Carman relation.GeoPhysics62,1480–1482.

Manzini,G.,Ferraris,S.,2004.Mass-conservativefinite-volumemethodson2-D unstructuredgridsfortheRichardsequation.AdvancesinWaterResources27, 1199–1215.

Matthews, G.P.,Laudone,G.M., Gregory,A.S.,Bird, N.R.A., Matthews,A.G.D.G., Whalley,W.R.,2010.Measurementandsimulationoftheeffectofcompaction ontheporestructureandsaturatedhydraulicconductivityofgrasslandand arablesoil.WaterResourcesResearch46,W05501,http://dx.doi.org/10.1029/ 2009WR007720.

Mitchell,J.K.,Soga,K.,2005.FundamentalsofSoilBehaviour.Wiley&Sons,NJ,USA, 577p.

Mualem,Y.,1976.Anewmodelforthepredictingthehydraulicconductivityof unsaturatedporousmedia.WaterResourcesResearch12(3),513–5221976. Ng,C.W.W.,Menzies,B.,2007.AdvancedUnsaturatedSoilMechanicsand

Engi-neering.TaylorandFrancis,Abingdon,Oxfordshire,UK.

O’Sullivan, M.F., Robertson,E.A.G.,1996.Criticalstate parametersfromintact samplesoftwoagriculturaltopsoils.SoilandTillageResearch39,161–173. Reynolds,W.D.,Durury,C.F.,Yang,X.M.,Tan,C.S.,2008.Optimalsoilphysical

qualityinferredthroughstructuralregressionandparameterinteraction. Geo-derma146,466–474.

Wang,M.C.,Huang,C.C.,1984.Soilcompactionandpermeabilitypredictionmodels. ASCEJournalofEnvironmentalEngineering110,1063–1083.

Whalley,W.R.,Jenkins,M.,Attenborough,K.,2011.Thevelocityofshearwavesin saturatedsoil.SoilScienceSocietyofAmericaJournal75,1652–1657. Wo¨sten,J.H.M.,Lilly,A.,Nemes,A.,LeBas,C.,1999.Developmentanduseofa