1

June22,2017

Measurementofsoilcarbonstocks

BenEllertandCharlesRice

1) AgricultureandAgri-FoodCanada2) DepartmentofAgronomy,KansasStateUniversity

Soilversusnon-soil

Sincethedawnofagriculturalscience,humanshavebeeninterestedinmeasuringsoilcarbon.Thistypicallyinvolvedcollectingasoilsamplefromthefieldandanalyzingthecarboncontentinthelaboratory.Thelateralandverticalvariabilityofsoilsrequiresthatthesamplinglocations,samplingdepths,andeventhesamplingtimearetakenintoconsideration.Topreservethesoilorganiccarbonatthetimeofsampling,soilstypicallyareairdriedtocurtailbiologicalactivity.Evenwithinthesoilsample,theheterogeneousdistributionofsoilrequiressomeformofprocessingbeforeanalysescanberepeatedwithconfidence.Thestandardmethodofprocessingsoilsamplesistopassthemthroughasievewith2mmopenings.Thematerialthatpassesthroughthesieveisregardedassoil,whilethatwhichdoesnotoftenisregardedasnon-soil.Thenon-soilconsistsofstonesorsolidmineralfragmentslargerthan2mm,butitinvariablycontainsrootsaswell.Toooftenthenon-soilhasbeenmerelydiscardedassomethingirrelevanttothestudyofsoilscience,butaswillbecomeapparent,thenon-soilcomponentsoftenneedtobetakenintoaccountwhenestimatingtheecosystemquantitiesofCandothernutrients.

Below-groundplantcarbon

Oncethesoilsampleisremovedtothelaboratory,carefulconsiderationmustbetakennotonlyonthephysicaldimensionsofthesamplebutalsoonthenatureofthematerials(bothsoiland“non-soil”)containedwithinit.Seldomdiscussedistheextenttowhichplantrootsshouldbeincludedinthesoilsample.Manyregardrootsthatfailtopassthe2mmcut-offasnon-soilandexcludethelargermaterials,someattempttoexcludemostrootandsurfaceplantresiduesfromsoilsamples,whilesomeofuscutandshredthelargerpiecesoforganicmattersothatitmightpassthe2mmsieveandthusincludedwiththesoil.Rootcarbonandabove-groundplantresiduesonthesoilsurfacearethemostproblematicpoolstoquantifyinterrestrialcarboncycles.ThesizeofthesoilorganicC(SOC)poolinterrestrialecosystemsusuallyisexpressedastheSOCmassperunitareatoaspecifieddepthormassofsoil.ThebalancebetweenSOCinputs(ultimatelyfromplants)andoutputs(decompositionofSOCbacktoCO2emittedfromthesoil)determinesSOCstockorpoolsize.MeasuringplantCinputsfromtheabove-groundisstraight-forwardrelativetothosefrombelow-ground(mainlyroots,butalsorhizomesandotherplantstructuresformedbelowthesoilsurface).Below-groundplantCinputsandsomeportionofabove-groundresiduesonthesoilsurfaceinvariablyarecollectedwithsoilsamples.Oftenthelargerandmostobviousportionoftheseconstituentsaresimplydiscarded,othertimestheseconstituentsarecarefullyremovedfromandexcludedfromtheSOCpool,andsomearguethatthese

2

shouldbeincludedwiththesoilsample.Whilesuchconstituentstypicallyaccountforlessthan20%oftheSOCpool,theyusuallyarethefractionthatismostsensitivetolanduseandmanagement.Wesuggest,therefore,thatthesecomponentsshouldeitherbequantifiedasaseparatepoolorincludedinthe>2mmsoilsample.Thebelow-groundplantCinputsandsoilsurfaceresidues(onlypresentinsoilsamplesfromtheupper-mostsoillayer)mustnotbediscardedinahaphazardfashionbecausetheyareacriticalanddynamicpartoftheterrestrialCstock.

Oncesoilsampleshavebeenremovedfromthefield,air-driedandcrushedtopass2mm,theyarefundamentallydistinctfromthesoilsintheiroriginalnaturalstate.Theliquidphasehasbeenlargelyeliminated,thegasphasehasbeenhomogenizedandlargelydissociatedfrombiologicalactivity,andthestructuralarrangementofthesoilparticleshasbeenextensivelydisrupted.Whatisregardedassoil,especiallythearrayoforganicconstituentsincludedassoil,isoperationally-definedbythesampleprocessingprocedures.Methodsthatseektoseparatemacro-organicmatterfromthesoilmustbeuniformlyandquantitativelyimposeacrossallsamples,recognizingthatbelow-groundplantCinputsmayencompasscoarserootsaswellasfinerootletsandrootexudates(nowdriedontothesoilparticles)andevenmycorrhizalhyphae.Againthedistinctionrarelyisclear-cut,andsofractionssuchasmacro-organicmatter,aredefinedbytheproceduresusedtoisolatethemfromthesoil.Oftenitmaybepreferabletoincludemacro-organicmatterwiththe<2mmsoilsample,butinmanysettings,itisinappropriatetosimplydiscardthematerial.Beforesoilsamplesmaybeanalyzedtodeterminechemicalcomposition,theymustbeprocessed(dryingandsieving,atminimum).Sinceprocessedsoilsamplesoftenbearlittleresemblancetothenaturalsoilprofileorpedonfromwhichtheywereobtained,itiscriticaltoaccountforchangesduringprocessingsoilthatanalyticaldatamaybeplacedbackintothecontextofthepedon.

Soilorganiccarbonversusorganicmatter(SOCvsSOM)

Soilorganicmatterreferstothenon-mineralportionofthesoil,whereassoilorganicCreferstocarbonatomsthatarepresentwithinorganicmolecules.Organicmatterconsistsofadiversityofmolecules,rangingfromsimpleorganicanionslikeacetatethroughtolargehetero-polymersknownashumicsubstances.ConfusionpersistsregardingthedistinctionbetweenSOMandSOC,andmanyfarmersandextensionpersonnelstillseemtorefertoSOM,eventhoughanalysesofitarearchaic.Conceptually,thereisnothingfundamentallyerroneousaboutSOManddiscussinghowitmightbemanagedandhowitmightinfluenceecosystemfunction.Analytically,however,soilorganicmatterisdifficulttoassessreliably,andmostassaysaredefinedbythemethodusedratherthanthesubstancesoughttobeassessed.SOMtypicallyisdeterminedviawetoxidationwherebysoilismixedwithanacidicsolutionofpotassiumdichromate(K2Cr2O7).Overtheyearsadiversityofoxidationconditions(>2mmor>0.25mmsoilparticles;withorwithoutheating)andassayshavebeenusedtoestimatethemassfractionofSOMinthealiquotofsoiltakenfortheassay.Oftentheamountofunreacteddichromatewasdeterminedbyredoxtitration,suchthattheamountofoxidantremainingwasinverselyproportionalSOMcontent.CorrectionfactorstoaccountforincompleteoxidationofSOMusuallywererequired.Notonlyisthisanalyticalapproachempirical,withoxidantconsumptionequatedtoSOM(despitethecommonoccurrenceofothersoilconstituentsthatconsumetheoxidant),butitalsogenerateschromium-ladenwastes(includinghexavalentchromium)thatareknowntobetoxic.Althoughdeficienciesinwet

3

oxidationmethodsforSOMwerenotedasearlyas1930byEnglishresearchers(WalkleyandBlack1934),andhavebeenacknowledgedrepeatedly(e.g.Chatterjeeetal.2009),suchmethodscontinuetoberegardedassimpleandinexpensivetoperform.

ThecontinueduseofwetoxidationmethodsbysoiltestinglabslikelycontributetothepersistencereferencetoSOM.ProducersandextensionagentsoftenaremorefamiliarwithSOMcontentsthanwithcorrespondingSOCcontents.MorecarefulconsiderationsofhowmanagementmightinfluenceSOMsoonraisequestionsaboutthedifferencesinthechemicalnatureandmolecularstructureofSOMrelativetotheplantmaterialsenteringthesoil.Comparedtorecentlydepositedplantresidues(roots,stems,leaves,chaff,flowers,seedcoats,etc.)amajorityofSOMhasbeenmodifiedbysoilbiologicalprocessesandchemicalinteractionswithsoilminerals.Toconstructcarbonbudgetstocomparesoilinputsandoutputswiththeamountpresent,thecarbonatomisconvenientlyadoptedasthecommondenominator.BecausetheSOCconcentrationandstockprovideusefulinsightstoCcyclingandinterpretationsofmanagementeffects,SOCoftenisestimatedasSOM÷1.724.ThelatterfactoroftenisreferredtoasthevonBemmelenfactorafteraGermanchemistwhoin1890assertedthatSOMcontained58%C.Ofcourse,therecanbenosingleconversionfactortoenableastraight-forwardinterconversionbetweenSOCandSOMbecauseSOMisvariableandheterogeneous.Relativetothefirsthalfofthe20thcentury,nowthereisamuchgreaterappreciationfortheassociationandevenorgano-mineralcomplexationbetweenmineralandorganicsoilparticles.IsolatingSOMfrommineralsoilstypicallyinvolvessomechemicalalterationofit,sodataontheelementalcompositionofSOMisolatedfrommineralsoilarerare.Anevensimplerbutrelatedproblemistheelementalcompositionofplantrootsbecauseelementalanalysesofrootsamplesfrommineralsoilsindicatethattherootorganicmatterhasbeendilutedbymineralparticles

Soilcarbonanalysis

ContemporarymethodsofsoilCanalysisinvolverequiringtheuseofautomatedcombustionanalyzersthatconvertcarbontoCO2whichsubsequentlyisanalyzed.ManyoftheseanalyzersalsodeterminetotalsoilNaswellviaautomationoftheDumasmethod,namedforJeanBaptisteAndreDumas,aFrenchChemistworkinginthemid-1800s.ThehightemperaturesrequiredforcompletecombustionororganicmatterandthermaldecompositionofcarbonatesalsoconvertsoilNtogaseousoxides,whichsubsequentlyarereducedtoN2bypassingthroughreducedcoppergrainsinaheattube.TheabilitytoreliablydeterminebothCandNcontentsinasingleanalysisisveryusefulbecausetheresultingC/Nratioprovidesvaluableinformationofthechemicalnatureofthesecrucialelements,aswellasimportantchecksonanalyticalperformance.MostoftheseinstrumentscanbeconfiguredtodetermineCandNatmuchgreaterconcentrationsthanthoseencounteredinmineralsoils,andsotheymayalsobeusedforanalyzingplantandanimaltissues,andotherorganicmaterials.MostcontemporaryCNanalyzersareequippedwithanautomatedsamplerwhichpermitsunattendedanalysesofsamplesetsthatmayvaryinsizefrom30to300individualsamples.Mostoftheanalyzersusepuregases,suchasHe,O2,N2orAr,typicallysuppliedfromcompressedgascylindersorgasgenerators.ThepuregasesareessentialtoflushawayambientCO2andespeciallyN2inthebackgroundatmosphere,andO2oftenisrequiredtoensurecompletecombustionoforganicmaterialstoCO2.Whenoperatingproperly,theanalyzershouldconvertallsoilCtoCO2,regardlesswhetheritispresentasorganicmatter,charor

4

inorganiccarbonate.In2016thecostsofsuchanalyzerstypicallylieintherangeof$50,000to$150,000USD,dependingonconfiguration.Theanalyzersvaryconsiderablyinthesamplesizetypicallyrequiredforasingleanalysis(5to5000mgmineralsoil),incombustionprocess(tubeorientation,reagents/catalysts,temperatures,ashremoval,etc.),intheapproachesusedtoseparatethecombustiongases,andfinallyintheanalyticalprincipleusedtoquantifytheamountofCO2orN2inthecombustiongases.

AutomatedsoilCNanalyzersmeasuretotalcarbonandtotalnitrogenconcentrationoriginallypresentinthesoilsampleandleavebehindnon-combustibleashinthecombustiontubeorsamplecrucible/boat.Manysoilscontainappreciableamountsofcarbonate,sometimesatthesurface,butofteninthesub-surfacelayers.SoilcarbonateorinorganicCisfundamentallydistinctfromSOC,asitissubjecttoformationanddissolutionbyabioticprocesses,albeitwithconsiderableinfluenceofbiologicalprocesses.Usuallythetime-scaleassociatedwithchangesinsoilinorganicCareconsiderablylonger(100sto1000sofyearsormore),thanwithchangesinSOC(10sto100sofyears).Ofcourse,soilinorganicCmaychangerapidlyinresponsetoapplicationofagriculturallimetomanagesoilacidity,ortoothercarbonate-richamendments,includingcarbonate-richirrigationwater.Ongeologicaltimescales,precipitationanddissolutionprocessesdominatethecirculationofcarbonamongthelithosphere,hydrosphere,andatmosphere.Onthetimescaletypicallyassociatedwithanthropogeniclandmanagement,theprimaryfocustendstobeonSOCratherthaninorganicC.Consequently,itisimportanttoanalyticallydistinguishsoilinorganiccarbonfromSOCortotalsoilC.Wetoxidationtechniquesareunreliable,astheyrarelyrecoveralloftheSOC.OnlyslightlybetteraredualcombustiontemperatureapproachesthatattempttoexploitthefactthatmanycarbonatemineralsremainstableattemperaturesabovethoseatwhichSOCcombuststoCO2.Dualcombustiontechniquescannotbeimplementedonanalyzersthatemploydynamicflashcombustionbecausehightemperatures(1000°Candabove)andexothermicreactionstypicallyexceedthethermalstabilityofcarbonates.Thefundamentalproblemwiththedualtemperatureapproachesisthatsoilchemistryismessy,andratherthanbeingpresentassomewell-definedcrystalstructure,soilcarbonateisco-precipitatedwithanarrayofotherelements,andthestructurerangesfromcrystallinetoamorphous.Consequently,soilcarbonatemaydecomposeatlowertemperaturesthanwell-orderedminerals.Conversely,SOCmaybeoccludedorprotectedfromcombustionbycloseassociationwithmineralsurfaces,suchthatgreatertemperaturesarerequiredforcombustionthattypicallyexpectedfornon-complexedSOC.

SoilinorganicCmaybeeliminatedfromsoilsamplesbyacidificationbeforedirectautomatedcombustionanalysisofSOCpersistingafteracidification,orinorganicCmaybedetermineddirectlyasCO2produceduponacidificationofthesoil,andSOCthenmaybedeterminedastotalminusinorganicC.FullyautomatedinstrumentationspecificallydesignedfordeterminingsoilinorganicCisuncommon(withtheexceptionofthosesoldbyUICInc.inJolietILUSA).Manyreliabletechniques(includingmanualapparatus)fordeterminingsoilinorganicCareavailable,anddetectionofevolvedCO2oftenhasbeenautomatedthroughtheuseofnon-dispersiveinfraredgasanalyzersorgaschromatographs,andrelatedmodesofdetection.Variousacids(H2SO3,H3PO4,HCl)andassortedacidificationprocedureshavebeenusedtoeliminatesoilinorganicCbeforeusinganautomatedCNanalyzertodeterminetotalCwhichisequivalenttoSOC,providedinorganicChasbeeneliminatedsuccessfully.Potentially

5

confoundingeffectsofimproperacidificationproceduresisthesolubilisationandsubsequentlossoforganicC(e.g.causedbywashingorrinsingthesamplewithaqueousacid)andbyweightchangesassociatedwithacidificationreactions.Stoichiometrically,forexample,thereactionbetweencalciumcarbonateandHClwillliberateCO2,theresultingCaCl2willhaveagreatermassthanthecarbonateitreplaced.Togetaroundtheseissues,insteadofacidifyingabulksoilsample,manyanalystsperformasmall-scaleacidificationofonlythealiquotofsoiltobeintroducedtotheautomatedCNanalyzer.Inthisway,themassoftheoriginalnon-acidifiedsampleisusedinthecalculationofSOCconcentration,andanymasschangescausedbyacidificationareimmaterialbecausetheseoccuraftertheoriginalmassisrecorded.ManyanalystsuseHCl,asafteracidificationanyexcesscanbeeliminatedbygentleheatinginavacuumoven,andsolubilisationisanon-issuebecausethesampleisdriedbeforeanalysiswithoutrinsingorwashingthatmightremovesolubilizedSOC.

Soilsamplecollection

Soilsamplecollectionisthemostcriticalaspectofmeasuringsoilcarbonstock.ThemostsophisticatedinstrumentproducingthemostaccurateandprecisedeterminationsofsoilcarbonconcentrationinthelaboratoryislikelytoprovidequestionableandevenmisleadingassessmentsofSOCstockifthesoilshavenotbesampledproperly.AssortedapproachestoevaluateSOCinsitu,withoutcollectingsamplesandanalyzingtheminalaboratoryarebeinginvestigated,butthestandardreferenceapproach,forallbutthemostcomplexsetting,involvescollectingsoilsamplesinthefieldandreturningthemtothelaboratoryforprocessing,preparation,andanalysis.Soilsarehighlyvariableinspace,andthesizeormorespecifically,theareacapturedinthesoilsampleoftenissmallrelativetotheextentofthepedonorsoilmanagementunit.Inevitablyandjustifiably,spatialvariabilitycombinedwiththesmallareacapturedbysoilsamplingleadstoconcernsabouthowwellthesamplerepresentstheareaofinterest.Whentheprimarygoalistomeasurethetemporalchangeinsoilcarbonstorage,contributionsfromspatialvariabilityareevenmoreconcerning.Astrategyofsimplificationmaybeadoptedtolessentheinfluenceofspatialvariabilityontemporalchangesinsoilcarbonstocks.Inthisinstance,thescopeisnarrowedtofocusontemporalchangesatamicrosite(e.g.anareaof10to15m2).SufficientmicrositesmustbesampledtoprovideestimatesofthedispersioninmeasuredSOCstocksaboutthemean.Withinanindividualmicrosite,however,itisassumedthatspatialvariabilityislessimportantthantemporalchanges.

Itiscrucialthattheinvestigatorisfullyawareofwhatisbeingmeasured.Thesoilsamplingapproachdiscussedherefocussesontheeffectsofbiologicalprocesses,namelythebalancebetweenplantCinputsandheterotrophicdecomposition.Insomelandscapes,SOCstocksalsoareinfluencedbygeomorphologicalprocessesorthebalancebetweensoildepositionandsoilremovalbyerosion.Theapproachpresentedhereisthesimplestcasewhereerosionanddepositionmaybenegligible.QuantifyingchangesinSOCstocksinlandscapeswithappreciabledepositionorerosionrequiresmoreinvolvedtechniquesappliedatthelandscapeorcatchmentsscale,andarebeyondthescopeofthisdiscussion.FailuretorecognizethecomplexityofgeomorphologicalprocessesinfluencingSOCstocksinerosionanddepositionallandscapeshascontributedmuchconfusiontoassessmentsofland-atmosphereCexchange.ThebiggesterrorstendtobeassociatedwithinadequateconsiderationofthefateofSOCinerodedsoil.UsuallyitisassumedtoberapidlydecomposedbacktoatmosphericCO2,

6

whereasmorecarefulassessmentsoftensuggestasizablefractionofSOCbecomesstabilizedindepositedsoiland/orsediment,andthattherateofSOCaccumulationatrecentlyerodedsitesmayexceedthepre-erosionalrateofaccumulation.ThemicrositeapproachdiscussedherewillmeasuretemporalchangesinSOCatsiteswhereerosionanddepositionarenegligible.Theapproachisalsoapplicabletorareinstanceswheretherateofsoildepositionorerosionatthesiteispreciselyknown.

Soilsamplesmustbecollectedinsuchamannerthattheanalyticalconcentrationsdeterminedforahighlyprocessedsub-samplecombustedinautomatedanalyzersmaybeplacedbackintothecontextofthefield.Todothis,soilsamplesmustbecollectedfromawell-definedareaandacarefullymeasureddepthincrement.Theareaanddepthincrementdefinethevolumesampled,andthesoilweightperunitvolumetypicallyiscalledsoilbulkdensity.TheSOCstockiscalculatedastheproductofsoilbulkdensity,layerthickness,andSOCconcentration.Forexample,iftheSOCconcentrationis2%or20kgSOCMg-1soilfora15cmthicksoillayerwithabulkdensityof1.4Mgm-3,theSOCstockmaybecalculatedas:

!"$%&'()*%+,-.

×0.15𝑚× ).!*%+,-.)56 = 8.9$%&'(

5:

InvestigatorsusuallyrecognizetherelevanceofsoilthicknesstoestimatingSOCstocks,butmaynotfullyappreciatehowdifficultitmaybetoaccuratelymeasurethicknessinthefield,orthatsimpleestimateofsoilthicknessrarelyprovidequantitativeinsighttosoilerosion.SimplecalculationslikethatgivenabovealsohavealertedinvestigatorstotheimportanceofmeasuringsoilbulkdensitytoestimateSOCstocks.Unfortunately,thisawarenesssometimeselicitscomplicatedandindependentmethodstomeasureinsitusoilbulkdensitythatmightberelevanttocharacterizesoilphysicalprocesseslikewaterinfiltrationorgasdiffusion.ForthepurposeofestimatingSOCstocks,bulkdensitymeasurementsdonothavetobecomplicated,butthemeasurementsshouldbedeterminedfortheverysamesamplescollectedtodetermineSOCconcentrations.

Inpractice,mechanically-drivensoilcoringequipmentprovidesthemostefficientmeansofcollectingsamplesfordeterminingSOCstocksinagroecosystems.Thecrosssectionalareaoftheinsidediameterofthesoilcoretubebitdefinestheareaofsoilbeingcollected.Dependingonthesetting,acorediameterof5to10cmprovidesforanadequatevisualassessmentofsoilcharacteristics,andmanageableamountofsoilforprocessinginthelaboratory.Inthefield,thecoretubecontainingthesoilisremovedfromthehydraulicapparatususedtoinsertthetubeandwithdrawthesoilcore,thetubeisplacedinanhorizontalorientation,andthesoilcoreiscarefullypushedfromthetubeandcutintoappropriatedepthincrements.Asitispushedfromthesamplingtube,thesoilcoreisinspectedforbreakage,excessivecompaction,adhesiontothecoretubewalls,channelscarvedbystonesorwoodyplantmaterials,etc.Theendofthecoretubemaybeusedasacuttingguidetoimprovetheaccuracyofthicknessmeasurements.Thecoremaycrumbleasitistransferredintotraysorbagsforsubsequentprocessing,buttheoriginalvolumemustbewell-definedandpreciselyknown.Thedrymassofsoilandanystones(>2mm)willbedeterminedinthelaboratory,andthestone-freedrymassperunitvolume(bulkdensity)willbeusedtoestimateSOCstocks.Therecommendeddepthofsoilsamplingshouldbeadjustedtocomplywithstudyobjectivesandenvironmentalsetting,buttypicallywillbebetween40and120cm,withincrementsof15cmorsmaller.Ideally,allincrementsforaparticularsamplingpoint

7

willbedrawnfromthesamecoreorcores.Sometimestwoorthreecoresmaybecombinedforasinglesamplingpoint,therebyincreasingtheareasampledandlesseningthepotentialeffectsofspatialvariability.Theimportantaspectisthatbulkdensityisdeterminedforthecoresactuallycollectedandthattheentirelayersamplediscollectedinacontiguousfashion(i.e.layerssectionedcarefullysothatnoincrementisexcluded).

Theconfigurationofsoilsamplingpointswithinamicrositemayvarydependingonthedesirednumberofcoressampledfromwithineachmicrosite,andthenumberoftimesthemicrocosmwillbesampled.Forexample,witha4x7mrectangularareasixinitialsamplesmightbecollectedalongtworowsat2mintervals.Atsomesubsequentsamplingtime(say5to10yearslater)sixadditionalcoresarecollected,butthesewillbeoffsetby1mbutinterspersedwiththecorescollectedinitially.Topreciselymarkthelocationswheretheinitialcoreswerecollected,electromagneticmarkers,suchasthoseusedbyurbanutilitycompaniesmaybeburiedwellbelowthedepthoftillage.Tominimizemicrositedisturbance,andespeciallyburialoftopsoilfromwithinthemicrosite,theholesleftbytheinitialsamplingshouldberefilledwithcorescollectedfromsomedistance(perhaps10m)outsideofthemicrocosm.ThesixcoresmaybecombinedsothemeanSOCstockmightbeestimatedfortheinitialsamplingtimesateachmicrocosm,andsothatanalyticalresourcesmaybedevotedtosamplingagreaternumberofmicrocosmswithinthefieldorsoilmanagementunit.Inpractice,carefulsamplingisexpensive,andinvestigatorsmayanalyzeSOCstocksforeachcorewithinthemicrosite.

Soilsampleprocessingandpreservation

Soilcoringequipmentusuallyworksbestinslightlymoist,butnotwetorverydryconditions.OntheCanadianPrairies,ourpreferredtimeofsoilsamplingisinthefallaftersoilmoisturehasbeendepletedandthecrophasbeenharvested.Thesoilsamplesareinafieldmoistcondition,inmostyearsambienthumidityisconducivetoair-dryingasameanstoslowmicrobialtransformationsandpreservethesoilsample.Soonafterthesoilsamplesaretransferredfromthefieldtothedryingroom,thetotalwetweightisrecordedandthecoresarecrumbledtoobtainrepresentativesub-samplestodeterminefieldmoisturecontent(bydryingasub-sampleat105°C;subsequentlydiscarded)andtoretainasampleatfieldmoisturecontentifrequiredformicrobialanalyses(preservedbyrefrigerationorfreezing).Typically,bulksamplesaredriedinAlfoiltrays(re-usable,withinlimits),andsamplesarestirredonadailybasistoacceleratedryingatambienttemperature(20°C).

Soilstypicallyarecrushedtopassasievewith2mmaperturestoseparatesoilfromthestones.Asdiscussedabove,procedurestohandlecoarsefragmentsofplanttissues,especiallyinsoilsfromthesurfacelayermustbedefinedattheoutset.Werecommendagainstarbitrarydistinctionsbetweensoilandnon-soilC,andinstead,prefertoretainallmaterialscollectedinthesoilcoreatsampling.Somelabscontinuetouseperforated(2mmroundholes)drummillssimilartotheRukuhiagrinderdescribedbyWatersandSweetman(195?).Themineralsoilaggregatesandwellasamajorityoftheplantmaterialiscrushedtoparticlessmallerthan2mm.

SincemostautomatedCNanalyzerstypicallycombusta5to500mgaliquotofmineralsoiltorepresentanentiresamplethatmayweightontheorderof1kg(evenmoreifsamplesarecomposited),fine-

8

grindingisessential.FinegrindingmustreduceparticlesizetoensuretheSOCandtotalNishomogenouslydistributed,andthatthesmallaliquotintroducedtotheanalyzerrepresentstheentiresample.Ageneralguidelineisthatarepresentativesubsampleiscollectedfromthe2mmmaterial,andcrushedtopassaNo.60or100meshsievewithopeningsof250or150um.Representativesub-samplingandcarefulfine-grindingisessentialtoobtainreliableresults,andsometestingofsub-samplingvariabilitytoprobessourcesandamountsofsamplingvariabilitymaybeinstructive.Thesizeofthesub-sampleandconsequentlytheamountgroundmayvarywithavailablegrindingproceduresandequipment.Forexample,forhighvolumeprojectsweoftenusearollermillinwhicha7gsub-sampleislefttotumblefor18hoursinmetalcanisterscontainingmetalbarstoeffectpulverization.Thismillcangrindmorethan100samplessimultaneously,butthe7gsub-samplemustrepresenttheentiresample.Forsomeprojectsweuseavibratory‘dishringandpuckmill’,similartothoseusedtocrushoresamplesformetalanalyses.Dependingondishsize,sub-samplesof70gormorearepulverizedtoafinepowderyconsistency,sosub-samplingtendstobelessproblematic.Samplethroughput,however,isreducedandlaborrequirementsareincreased,becausesamplesmustbegroundoneatatime.Inmanycasesthe2mmsoilinthecatchtrayfortheperforateddrummillmaybethoroughlymixedandrandomscoopstakentocollectarepresentativesub-sample.Inothercasestheconeandquarteringtechniqueormanualriffledividersmaybeusedtoobtainrepresentativesub-samples.Sincethesamplemayhavetogothroughsuchaprocessmultipletimestoobtainamanageablesub-sample,suchproceduresmaybecomecostly,dustyanderror-prone.Intheseinstancessomesortofmechanicalsampledivider,suchasarotatingtubesub-sampler,maybeusedtocreateaflowingstreamofdrysoilsothatrepresentativeandunbiasedsubsamplesmaybeobtainedfromit.

Soilanalysis

Afterfinegrinding,soilsareretainedinsamplevialsorevenhigh-gradepapercoinenvelopes.Despiteasmallriskofcontributionsfromthepaperfiber,suchenvelopesareinexpensive,easytolabel,exposealargesurfaceareatodryingorequilibrationtoambienthumidity,andmostimportantarelesspronetothebuild-upofstaticchargewhichotherwisemaythwarttheaccuracyofmicro-analyticalweighing.AnalysesoftotalCandNarecalibratedusingpureorganicchemicals(e.g.acetanilide,glutamine,EDTA),matrixrelevantreferencematerials(e.g.soils,sediments,ores)developedin-houseorpurchasedfromcommercialsuppliers,orreferencematerialswithconcentrationsassignedbyrecognizedgovernmentagenciesorcertificationbodies(e.g.agenciesresponsibleforchemicalmetrology).

CalculationoftemporalchangeinSOCstocks

Assumingthatsampling,processingandanalysisofthesoilsampleshasbeendonecorrectly,andfurtherassumingthestone-freesoilbulkdensitiesattheinitialandsubsequentsampletimeareidentical,SOCstocksforsuccessivesoilthicknessesorvolumesmaybecalculatedsimplyastheproductofSOCconcentration,layerthicknessandbulkdensity,asdescribedpreviously.Inpractice,however,bulkdensitiesvaryamongmicrocosms,andespeciallybetweensoilsamplingtimes.Oftensoilmanagementofenvironmentalconditionswillbedissimilaratinitialandsubsequentsamplingtimes.TheanalyticalconcentrationsofCandNdeterminedforthesoilmustbeplacedbackintothefield

9

contexttoassesswhetherappreciablechangesinSOCstockshaveoccurred.ManyinvestigatorsfailtorecognizetheinterdependenceofsoilmassandSOCmass,butgreaterbulkdensitiesforthesamesoilthicknessmeansagreatersoilmass,andthiswilltendtoinflateSOCstock.Insomesettings,theremaybegoodreasons(e.g.basedonsoilprofilemorphology)tovarythethicknessesofthesoillayerssampled,butthiswillfurtheraccentuatedifferencesinthemassesofsoilbeingcompared.

Toavoiderrorsassociatedwithcomparingunequalsoilmassesinsettingswheresoilredistributionisnegligible,thethicknessesofsoilbeingcomparedareadjustedtoattainanequivalentsoilmass.Theapproachdoesnothavetobecomplicated,butsomeinvestigatorsinitiallyfinditunsettlingtocompareSOCstocksinunequalsoilvolumesorthicknesses.Theequivalentsoilmassapproachsimplyadherestothepremiseofmassconservation,andadjustssoilthicknesssothatsoilmassisequivalent.Someassumptionsareinherentinthisapproach,becausethethicknessesofthesoillayersbeingcomparedmaynotexactlycoincidewiththosesampled,butinpracticetherequiredinterpolationsusuallyarepalatable,andtheuncertaintiesaboutSOCconcentrationsdiminishasdeeperlayersareconsideredwherevariationswithdepthtendtobesmall.

10

11

Modelofadissolvedorganicmattermoleculeorhumicsubstance.Colorcodesforatomtypesareasfollows:carbon(cyan),hydrogen(white),oxygen(pink),nitrogen(blue),andsulfur(yellow).from:SchultenH-R(1999)Analyticalpyrolysisandcomputationalchemistryofaquatichumicsubstancesanddissolvedorganicmatter.J.Anal.Appl.Pyrolysis49(1):385–415.

12

13

References

Bemmelen,J.W.Van.1890.DieZusammensetzungdesMeeresschlicksindenneuenAlluviendesZuidersee(Niederlande),In:DieLandwirtschaftlichenVersuchs-Stationen.Chemnitz,Bd.37,S.239

Chatterjee,A.,R.Lal,L.Wielopolski,M.Z.MartinandM.H.Ebinger.2009.Evaluationofdifferentsoilcarbondeterminationmethods.CriticalReviewsinPlantScience28:164–178.

Heanes,D.L.1984.Determinationoftotalorganic-Cinsoilsbyanimprovedchromicaciddigestionandspectrophotometricprocedure.Commun.SoilSci.PlantAnal.15:1191–1213.

Walkley,A.;Black,I.A.1934.AnexaminationoftheDegtjareffmethodfordeterminingsoilorganicmatter,andaproposedmodificationofthechromicacidtitrationmethod.SoilScience37:29-38.