Nov 20, 1998 - Abstract. Conversion of humid tropical forest to agriculture significantly alters trace gas emissions from soils. We report nitrous oxide (N20), ...
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 103, NO. D21, PAGES 28,047-28,058, NOVEMBER
20, 1998
Nitrous oxide, nitric oxide, and methane fluxes from soils
following clearing and burning of tropical secondary forest
A.M. Weitz,• E. Veldkamp, 2M. Keller,TMJ.Neff,4andP.M. Crill• Abstract. Conversionof humidtropicalforestto agriculturesignificantlyalterstracegas emissionsfrom soils.We reportnitrousoxide (N20), nitric oxide (NO), andmethane(CH4) fluxes from secondary forestsoilsprior to and duringdeforestation, andthroughoutthe first agricultural cropping.Annualaveragenitrogenoxideemissions from forestsoilswere 1.5 ng N cm'2h-1for N20 and0.9 ng N cm'2h-1for NO. Forestclearingincreased the levelof extractable nitratein soils andaveragenitrogenoxidesfluxes(2.7 ng N cm-2h-I for N20, and8.1 ng N cm-2h-I for NO). Immediatelyafterbiomassburning,short-term peaksof N20 andNO (123 ng N cm-2h-I for N20,
and41 ngN cm-2h-1for NO) weresuperimposed on generallyincreased fluxes.Peakemissions declinedwithin 3 daysafterburning.Postbumfluxesstayedhigherthanmeasuredon adjacent forestsitesfor 3-4 months(averagesfor postbumfluxeswere 17.5ng N cm-2h-1for N20 , and 19.2 ngN cm-2h-1for NO). Increased N20 andNO emissions afterclearinganduntilcroppingwere probablydueto a combinationof increasedratesof nitrogencyclingandhighergaseousdiffusion in dryingsoils.Comparedto emissions from youngpasturesin the region,fluxesof nitrogen oxidesfromunfertilizedagricultural areaswerelow (3.9 ngN cm-2h-1for N20 and3.4 ngN cm-2 h-• for NO), probablydueto nitrogenuptakeby fastgrowingcornplantsandlosses by leaching with drainingsoilwaterin the wet season.Variationin CH4 fluxeswashigh for all landuse periods.Forestsoilsconsumed an averageof 1.0mg CH4m-2d-l, whichslightlyincreased in drier soilsafterclearing(1.2 mg CH4m-2d-i).PostbumCH4consumption by soilswasslightlyreduced (0.8 mg CH4m-2d-l) compared to forestsoils.Unfertilizedagricultural soilsconsumed lessCH4 than forest soils.
1. Introduction
Soil-atmospherefluxes of nitrous oxide (N20), nitric oxide (NO), and methane (CH4) affect processesrelevant to global warming and environmentalpollution [Ramanathanet ai., 1987; Intergovernmental Panel on Climate Change(IPCC), 1996]. N20 is an importantgreenhouse gasbecauseof the spectralpropertiesof the moleculeand its relativelylong lifetime.N20 alsois involved in stratospheric ozone chemistry[Cicerone, 1987]. NO is highly reactive;henceit is importantfor local and regionalatmospheric photochemistry[Williams et al., 1992]. In the troposphere,NO is involvedin chemicalproductionprocesses of differentenvironmental pollutants, such as ozone and nitric acid. CH4, currently increasingin the atmosphere,is an important greenhousegas [Ramanathanet al., 1987; Graedel and Crutzen, 1993]. Soils are significantsourcesand sinksof N20, NO, andCH4. The biological• chemical,and physical conditionsin soils determinesoil-atmospheretracegasexchange. qnstitute fortheStudyof Earth,OceansandSpace,Universityof New Hampshire,Durham. 2Institute of SoilScienceandForestNutrition,Universityof GOttingen, GOttingen,Germany.
3International InstituteforTropicalForestry, USDA ForestService,Rio Piedras,Puerto Rico.
4Department of BiologicalSciences,StanfordUniversity,Stanford. Californiaø
Copyright1998 by the AmericanGeophysicalUnion. Papernumber98JD02144• 0148-0227/98/98JD-02144509.00
Nitrogen oxides are producedin soils by biotic and abiotic processes, but soil microbialactivity is the mostimportantsource and sink for NO and N20. Severalreviews [e.g. Firestoneand Davidson,1989; Galbally, 1989; Robertson,1989; Williamset al., 1992;Davidsonet al., 1993] discusssoil microbialproductionand consumption of nitrogenoxidesandthe controllingfactorsfor soilatmosphere nitrogenoxideexchange. Briefly,decomposition of soil organicnitrogencompounds releasesammonium(NH4+);nitrogen mineralizationratesdeterminethe furthernitrogenflow in soils. Aerobicmicrobialmetabolism converts NH4+ intonitrite(NO2-)and nitrate(NO3-)(nitrification).Both oxidesare reducedto gaseous forms during anaerobicmicrobial oxidation of organic matter (denitrification).Both microbialprocessesare sourcesfor NO and N20. On the microscaleaerobic and anaerobicsitescan occur in close proximity [van Cleemput and Samater, 1996]; therefore nitrificationand denitrificationmay occursimultaneously in soils. Ultimately,netNO andN20 soil-atmosphere exchangedependson the balancebetweengasproductionand consumptionandthe rate of nitrogencyclingin the ecosystem[Galbally, 1989]. Soilsareimportantsources andsinksof CH4. Soil watercontent stronglyaffectsCH4 dynamics.Microbial consumption of CH4 occursin mostwell-aerated soils.Anaerobicsoilconditions support microbialCH4production. Commonlyunsaturated soilsarenetCH4 sinks[Steudleret al., 1989;Crill, 1991]. Uptakeof CH4 decreases with increasing soilwatercontentandwith additionof nitrogenas ammonium-N in fertilizers or from atmosphericdeposition [Steudleret al., 1989;Mosieret al., 1991]. In moistsoils,anaerobic and aerobic micrositescan exist in close proximity; thus CH4 consumptionand productioncan be measuredin adjacentareas. Born et al. [1990] suggestedthat soil CH4 uptakeratesare con-
28,047
28,048
WEITZ ET AL.: SOIL TRACE GAS FLUXES DURING DEFORESTATION
strainedprimarilyby gastransportprocesses andonly secondarily Table 1. Soil Characteristics for Soils Under Forest Cover by microbialactivity. BulkDensity Depth, Humid tropicalecosystems are characterized by high ratesof nitrogenmineralizationand decomposition of organicmatter.On DystropeptForest the globalscale,tropicalforestsoilsare the largestnaturalnitrous
In
oxidesource, emittingabout3 Tg N20-N annually(1 Tg = 1012 g), which accountsfor 21% of the total annual N:O production [Matsonand Vitousek;1990].NO fluxesfrom humidtropicalforest soils are of similar rangeas N:O emission[Williamset al., 1992; Davidsonand Kingerlee, 1997]. Each year 20 to 60 million hectaresof tropicalland, mainly forest and savanna,are cleared and burned in preparationfor shifting cultivationor permanentagriculturaluse [Crutzenand Andreae,1990].Forestconversionaffectssoil biogeochemical and physicalproperties. Biomassburninginstantaneously increases NO and N20 emissionsthroughvolatilization.On longertimescales, nitrogenmineralization,nitrification,and soil inorganicnitrogen contentincrease[Ewel et al., 1981; Matson et al., 1987; Anderson et al., 1988].Correspondingly elevatednitrogengasemissions are measuredfor periodsof severalmonths [Andersonet al., 1988; Neffet al., 1995] up to yearsafterforestconversion[Luizaoet al., 1989; Keller et al., 1993]. Levine et al. [1996] found that both burningandwettingincreaseNO fluxesfrom savannasoils,while N20 tluxes Irom those Ory soils were t>elow Oetectlonlimit. Nitrificationwasthe dominantsourcefor NO emissions duringthe conversionof tropical lowland forest in CostaRica [Neff et al., 1995].Tropicalforestsoilscommonlyare net CH4sinks,but under agriculture andpastures,soilsmay becomenetCH4 sourceswhere compaction reducesgaseous diffusionratesfavoringanaerobic CH4 production[Keller et al., 1990, 1993]. We reporton tracegasfluxes(NO, N20 , andCH4)fromtwo soil types (clay and loam) measuredunder secondaryforest,during forestconversion, andfromthefirstcropcycleafterforestclearing
andburning.We includeNO emissiondatafor theclearingperiod originallypresented by Neffet al. [ 1995].Simultaneously withtrace gasemissions we measuredphysicalandchemicalsoilproperties.
2. Study Area Thestudywasperformed at theLa SelvaBiologicalStation(10ø 26'N x 84ø 0' W) in the Atlanticlowlandsof the provinceof Heredia,CostaRica.The climateishumidtropical,withanaverage temperatureof 25.8øC and annual precipitationof 3962 mm [Sanfordet al., 1994]. Precipitationis distributed throughout the year with a maximummonthlymeanof 481 mm in July.A weak dry seasonoccursbetweenJanuaryand April with a minimum monthlymeanprecipitationof 152 mm in March. In 1993we selected four l-ha sitesof secondary forestat the La
SelvaBiologicalStation.Siteswerelocatedontwo soiltypes,both developedon alluvial depositsof andesiticvolcanic material [Sollinset al., 1994]. Two siteswere on a fertile, loamy soil developed on low riverterraces (fluventicEutropept), andtwowere on lessfertile clay soils developedon high river terraces(andic Dystropept). Physicalandchemicalsoilcharacteristics aregivenin Table 1. Largepartsof the originalforestin the areaswerecleared for agricultureabout 1953. Both Eutropeptsiteswere usedfor cacaoplantations(Theobromacacao) with shadetreesof various species,includingCordia alliodora(laurel)andBactrisgasipaes (peachpalm,or pejibaye).Most shadetreeswereremnantsof the partiallyclearedforest.OneDystropeptsitewasusedfor a peach palmplantation,andthe otherwasusedfor pasture.The plantations wereabandoned about1968, andthe pastures wereabandoned10
Average s.d. Corg, % ø•o' (1•2I•)
0.05 0.15 0.30 0.50 0.70 0.90
0.77 0.88 0.88 0.83 0.82 0.83
0.05 0.15 0.30 0.50 0.70 0.90
0.66 0.74 0.69 0.71 0.76 0.93
0.06 0.11 0.13 0.08 0.06 0.05
7.3 4.5 2.9 1.6 1.3 1.3
0.4 0.3 0.2 0.1 0.1 0.1
4.7 4.7 4.8 4.9 4.9 4.9
0.4 0.3 0.2 0.1 0.1 0.0
6.0 5.9 5.9 6.0 5.9 6.2
EutropeptForest 0.05 0.09 0.04 0.05 0.04 0.02
7.6 4.0 3.1 1.5 0.9 1.0
yearslater.Secondaryforestdevelopedon all sites.This forestis differentfrom primaryforestin speciescomposition[Hartshorn andHamreel,1994];also,secondary forestshowsa well-developed understoryvegetation.
3. Site Preparation and Management In January1994 (Dystropept)and March 1994 (Eutropept), secondary forestwasclearedby manuallaborusingchainsawsand machetes.After removalof the understory,treeswere felled and sometimberwasremoved.The remainingslashwaschoppedand cutintoaboutl-m-longpiecesandallowedto dry on site.Clearing tookabout2 weeksfor eachsite.In March 1994(Dystropept) and April 1994(Eutropept),dry biomasswasburned.For severaldays after the initial fire, unburnedwood was piled by hand and reburned.With startof the wet seasonin the secondhalf of May 1994,corn(Zea mays)wasplantedby useof plantingsticks,the traditionalseedingtechniqueappliedby localfarmers.Plotswere weededby macheteandthinnedabout6 weeksafterplanting.In orderto follow traditionallow-inputfarmingtechniques of the region,we did not fertilize,till the soil,or applypesticides. Corn washarvested about4 monthsafterseeding. The sequenceof forest,clearingoperationsand agriculture providedusefullandmanagement periodsfor the analysisof our
experiment: (1) forest,coveringall measurements fromsecondary forestspriorto clearingandfrom adjacentforestsitesfollowing clearing;(2) cleared,startingwith the day of the cuttingand coveringthe time of full canopyclearanceuntil the biomasswas burned;(3) postburn,includingthe few daysof rebumingslash until planting;and (4) agriculture, signifyingthe first cropping cycleestablished afterforestconversion andincludingthefallow phase aftercornharvest(September 20-22, 1994)untilplantingof the secondcorncrop(November15-17, 1994). The time covered by eachlandmanagement periodis givenin Table2. 4. Methods
In July 1993we startedmeasurements of NO, N20, andCH4 fluxesfromsecondary forestsoils.For gascollectioneachsitewas equipped witheightchamber bases (polyvinylchloride (PVC) rings of 0.25 m diameterand0.12 m height,insertedabout0.02 m into thesoilsurface). Sampling spotswereselected randomlyfollowing the procedure described by Keller and Reiners[1994]; chamber basesweremaintained throughout theforestandclearedperiods.
WEITZ ET AL.: SOIL TRACE GAS FLUXES DURING DEFORESTATION
28,049
Table 2. Time Scheduleof Land ManagementCategorieson Both Soil Types LandUse Forest
Dystropept(Clay)
Eutropept(Loam)
Jul. 1, 1993 to Jan. 10, 1994
Jul 1, 1993 to Feb. 24, 1994
Cleamed Jan.11, 1994 to Mar. 13, 1994 Postbum Mar. 14, 1994 to May 19, 1994 Agriculture May 20, 1994 to Nov. 15, 1994
We removedchamberbasesimmediatelybeforeburningof slash, andreinstailed themfor permanent useon randomlyselectedpoints the day after the fire. Concurrentwith clearing,we installed chamberbasesat randomlyselectedpointsin adjacentforestareas to completea full year of flux measurements under secondary forest.Forestsiteswere sampledmonthly.During the clearedand postburnperiodswe sampledmorefrequently.Underagricultural use, gas fluxes were sampledmonthly. During agriculturewe startedsamplingN20 fluxes in high time resolution(4.2-hour. interval)usingan automatedchambersamplingandmeasurement system(P.M. Crill, unpublished data,1998).
Techniquesusedto measureN20, CH4 and NO fluxes are presented in detailby Keller andReiners[ 1994];a slightrevision for N20 wasdonein October1993 [Veldkampand Keller, 1997]. Briefly, gasfor the analysisof N20 and CH4 was sampledfrom static,ventedfieldchambers (headspace, 0.01 m3)at 1, 7, 14,21, and 28
min
after a chamber
was closed with
a removable
acrylonitrile-butadiene-styrene top, which exactly fits the PVC chamberbase.Chambertopshad two portsto allow gassampling usingnylon syringeswhile maintainingequilibriumwith atmosphericpressure. Gassampleswereanalyzedwithin 36 hoursafter collection.N20 was measuredwith an electroncapturedetector
Feb.25, 1994 to Apr. 5, 1994 Apr. 6, 1994 to May 24, 1994 May 25, 1994 to Nov. 17, 1994
depth)fromeightrandomlyselectedlocations,gatheringsoil from four spotsinto one bulk sample.The resultingtwo soil samples (each about 400 g fresh weight) were used to determine soil moisturecontentand extractableinorganicnitrogen [Keller and
Reiners, 1994].Thegravimetric soilwatercontent0g(g g-i)was derivedafterovendryingof soilto constantweightwithin 48 hours at 105øCandwasconvenedto proportionof water-filledporespace (% wfps) usingthe formula given by Linn and Doran [1984]
% wfps= (0gx bdx 100%)/ ( 1 - (bd/ 2.65)) with bulk density(bd (Mg m-3))datameasured fromforestsoils (Table1), andparticledensityestimated as2.65 Mg m'3. For the determinationof soil extractableinorganicnitrogen (nitrate (NO3') and ammonium(NH4+)) we prepared2 M KCI extractions[Keeneyand Nelson, 1982] (mixing ratio, 7.5 KCI :1 field moistsoil).Soil extractswere refrigerateduntil sampleswere analyzedusingstandardcolorimetricmethodswith an ALPKEM Analyzer [ALPKEM Corporation, 1990]. All chemicalanalyses weredoneat the laboratoryof the U.S. ForestService,Rio Piedras, Puerto Rico.
For soil carbonand total nitrogenanalysesbulk soil samples
(ECD)(Shimadzu GC-8Aequipped witha 63Nidetector) andCH4 weretakenfrom onesoil pit per siteat 0.05-, 0.15-, 0.3-, 0.5-, 0.7-, with a flame ionizationdetector(FID) (ShimadzuGC-mini FID) gaschromatography. Sampleconcentrations were determinedby comparisonto commerciallypreparedstandardgaseswhich had beencalibrated againstNOAA ClimateMonitoringandDiagnostics Laboratory(CMDL) and National Institute of Standardsand Technology(NIST) standards.Gas fluxes were calculatedfrom linearregressions of concentration versustime. NO wasdeterminedindirectlyafter conversionto NO 2.Air was sampledin a continuous air flow from dynamicchambersthat also preciselyfit the permanentPVC chamberbases.Chamberhead-
space(0.01-m3 headspace volume)wassampledat a flow rateof 300mL min-• andmixedwithpurifiedair (1200mL min-•)priorto
and 0.9-m depth, air dried, passedthrough a 2-mm sieve, and shippedto the laboratory.Soil organiccarbonwasdeterminedusing the modified Walkley-Black procedure [Nelson and Sommers, 1982], and convertedto soil organicmatterusingthe factor2.16. Soil total nitrogenwas determinedusingthe modifiedKjeldahl method[Bremnerand Mulvaney, 1982]. Undisturbedsoil samples
of 3x10-4 m3 volumewere taken at the sameprofile depthsto determinebulk densityand soil water retentionpropertiesusinga hangingwater columndevice [Klute, 1986]. Field capacitywas estimated from water retention
curves as the soil water content at
6-kPa suction.
We analyzedthevariationbetweensoil typesand land manageanalysis;ozonewas not scrubbedfrom air enteringthe chamber. mentperiodsusinganalysisof variation(ANOVA) techniques with NO was convertedto NO 2 on a CrO3 catalyst;then NO 2 was a split-plotdesign.Geographicallythe siteswere locatedin two quantified by luminol chemiluminescencedetection (Scintrex blocks,spatiallyseparatedby about 1.3 km. Each block had two LMA-3). A standard additionof approximately 3 mL min'• of mainplots,oneon a Eutropeptand one on a Dystropept.Each soil 1-ppmNO in oxygen-freenitrogenmaintainedthe signalin the type wassubdividedinto four subplottreatments(forest,cleared, instrument's linearrange.Betweenchambermeasurements we used postburn,and agriculture).We choosea conservativetest to largerstandardadditionsto calibratethe instrumentresponse.NO evaluatea potentialeffectof the physicallocationof the plotsnext fluxes were calculated from the linear increase in concentration to an effectof the soiltypesandtreatments. The analyzedstatistical with time after chamber closure.
Measurementswere performedbetween 0730 and 1430 LT. During that period,averagechangesin air temperatureare about 7.5øC.Frompreviousstudieswe expectedonly minor influenceof diurnal temperaturevariation on soil gas fluxes (M. Keller, unpublisheddata, 1993). Simultaneouslywith gas samplingwe measuredshadedair temperaturenear the soil surfaceand soil temperatureat 0.02-, 0.05-, and 0.1-m depth.Air temperaturedata wereusedin gasflux calculations. We reporttracegasfluxesfrom the soil to the atmosphereas positive values; negative fluxes indicategasconsumption by soils. On eachmeasurement day we collectedsoil samples(0- to 0.1-m
model is
y = It + block+ soil + (block*soil)+ treatment + (soil*treatment)+ E
wherey is the considered response variable;it is the overallmean for the variable;block, soil (main plots),andtreatment(subplots) are effectsin the model;(block*soil) is the main plot errorterm; and (soil*treatment)is an interactionterm in the ANOVA model; and E is the subploterror.The F value for the block and the soil effect is calculatedusingthe mean of the sum of squaresfor the interactionterm (block x soil) as the "main plot," while the mean
28,050
WEITZ ET AL.: SOIL TRACE GAS FLUXES DURING DEFORESTATION
of thesumof squares for E is usedwith thetreatmenteffectandthe (soilx treatment)interactionterm.For eachmeasured variablewe converteddaily averagesinto averageresponses for eachtreatment, resultingin a balanceddesigncomparing16 treatmentmeans.For statisticalanalysis,NO andN20 fluxeswere log transformed. We analyzed data using the JMP IN software,developedby SAS Institute[SallandLehman,1996].The Tukeycriterionwasusedin multiple comparisonteststo distinguishsignificantdifferences (p=0.05)for treatments andsoil-treatmentinteractionterms.These calculations weredoneoff-line from JMP IN. This analyticaldesign representsthe best approximationof the experiment,ignoring inconsistencies in thetemporalandspatiallayoutwhichresultfrom theappliedslash-and-bum practiceandof restrictions in thenumber of measurementsfeasible. For analysis of CH4 dynamics we compareda linearregressionmodelderivedfor forestto CH4flux data measuredduring subsequentland managementperiods,and describedstatistics of the residualsusingSigmaPlotversion2.0 for Windows.
MonthlyrainfallbetweenJuly 1993andNovember1994always exceeded 50 mm (Figure 1). The periodfrom mid-Januaryto midMay 1994 was moderatelydry (subsequently referredto as "dry season").For both soil types,soil moisturedatawere abovefield capacitymostof the wet seasonanddroppedbelow field capacity during the dry period (Figure 1). Forest soilsbecamedrier than clearedsoils.Average soil water contentunderforest(78% wfps Dystropept, and 75% wfps Eutropept) differed significantly betweensoil types. For both soil typeswe showchangesin soil inorganicnitrogen (KCl-extractable NO3-andNH4+) contentfrom forestto agricultural landmanagement in Figure2. Averagevaluesfor eachvariable,soil and land managementperiodare listedin Table 3. In forestsoils, inorganicnitrogencontentwas comparablefor both soil types(6.9
ñ 6.2 ggNO3'-Ngd[• (meanñ standard deviation;n= 41), 7.3 ñ 3.8 gg NH4+-Ngds -l (n= 41)) with littlevariationthroughout theyear (4.9 ñ 2.0 gg NO3--Ngd,-• (n= 25) wet season, and10.9ñ 5.9 gg NO3--Ngds '• (n= 15) dryseason; 8.9ñ 3.6 ggNH4*-Ngas -• (n= 25) wet season, 5.2 + 2.7 gg NH4*-Ngas 'l (n= 15) dry season). After
forest clearing soil extractableinorganicNO3' increasedrapidly (Table 3, Figure 2) until cropping, which coincided with the Bulk densitieswere low in bothsoiltypes(Table 1), with higher beginningof the wet season.PostburnNO3- concentrations were densitiesmeasuredin the Dystropept(clay) profilecomparedto the significantlydifferent from all other land managementperiods. Eutropept(loam).In bothsoilsthe topsoilstructureis dominatedby Clearinghad little effect on soil extractableNH4+ concentrations. well developed, loosely packed crumb aggregates.Generally, Biomassburningincreased NH4+ levelin bothsoils.PostburnNH4+ throughoutthe experiment,soilsremain highly porousand show concentrations differedsignificantlyfrom all otherland useperiods high valuesfor total porosity. (Table3). Duringthe agricultureperiod,amountsof extractable soil Statistical analysis showed no block effect. There was no inorganicnitrogendecreasedrapidly (Figure 2). significanteffect due to soil type, but eachvariabletreatmenthad NO, N20 , and CH4 fluxes measuredbetweenJuly 1993 and a significanteffect,andCH4 alsohada significant(soilx treatment) November 1994 for subsequentland managementperiods are interaction.We discussresultsfor each variable in the following shownfor the exampleof the Dystropept(Figure3). Averagenitric paragraphs.For details on differencesbetweenmean values, see oxidefluxesfrom forestsoilswere comparablefor both soiltypes 5. Results
Table 3.
(Figure4, Table3); annualemissions average0.9 ñ 1.2ngN cm-2 Table 3. Statisticsof MeasuredVariables,Distinguished per Soil Type andLand ManagementCategory Water-Filled
NH4+-N, ug gd[I
NO&N, ug ga., -•
average
7.09a
5.66a
s.d.
3.72
4.08
Period
NO, ng-Ncm-2h-•
N20, ng-Ncm-2h-•
CH4, mgm-2d-I
Pore Space,
0.86•
1.62•
-0.94•
0.80
1.38
0.78
76.79 • 12.01
%
DystropeptSoil Forest
Cleared
Postburn Agriculture
n
(19)
(19)
(20)
(20)
(20)
(20)
average
7.04•
23.78•
4.50b
2.81•
-1.04•
74.44 b
s.d.
2.35
16.65
5.39
3.94
0.46
n
(16)
(16)
(16)
(16)
(16)
(16)
17.57b
69.87b
16.53•
9.30b
-0.58b•
72.32 •b
average
6.65
s.d.
6.15
25.21
12.55
10.10
0.51
n
(17)
(17)
(16)
(17)
(17)
8.82
(17)
average
8.16•
15.43•
5.50•
5.76•
-0.49•
76.70 •b
s.d.
12.75
14.24
13.41
6.18
0.54
8.10
n
(21)
(21)
(21)
(21)
(21)
(21)
average
7.45•
7.93•
0.94•
1.46•
-0.98•
73.62 •
s.d.
4.27
7.36
1.06
1.21
0.53
10.20
n average
(22) 5.59•
(22) 15.94•
(22) 12.92b
(22) 2.64•
s.d.
2.15
! 1.78
10.21
2.80
0.48
n
(12)
(12)
(12)
(12)
(12)
(12)
average
18.13b
38.00b
22.76•
28.04•
-0.99•
65.90 •b
EutropeptSoil Forest
Cleared Postburn
Agriculture
(22) -1.450
(24) 63.08 b 6.31
s.d.
14.76
13.13
17.01
35.63
0.55
n
(17)
(17)
(17)
(18)
(18)
(18)
average
5.15•
7.43•
1.33•
2.11•
-0.36•
69.18 •b
s.d.
3.05
3.71
1.07
1.38
0.67
n
(21)
(21)
(21)
(21)
(21)
3.11
6.04 ,
(21)
Differentlettersuperscripts per variableandcategoryindicatesignificantlydifferentmeans;s.d.is standard deviation,andn is numberof daily averagedmeasurements.
WEITZ ET AL.' SOIL TRACE GAS FLUXES DURING DEFORESTATION
variation among daily averageswas high on both soil types. Postburn fluxes differed significantly from the other land managementperiods(Table 3).
lOOO
monthly precipitation 800
Forestsoilsconsumed CH4 (1.0 ñ 0.7 mg CH4 m2 d'l; n= 40).
600
AverageCH4consumption waslessduringthewet season(0.7 ñ 0.6
400
o
28,051
mgCH4m2d'l; n= 27) andgreaterin thedryseason (1.5 ñ 0.4 mg CH4m2d'l;n= 13).Variationbetween measurement dayswashigh.
200
CH4 emissionswere measuredoccasionallyfrom burned sites. PostburnCH4 fluxes differedsignificantlyfrom fluxes measured duringthe clearedperiod(Table 3).
,
lOO
90
6. Discussion
80 70
Dystropept:
60
• I' ,
', '
50
I
I
I
I
I
I
'l
I
I
'l
I
I
I
I
I
I
I
I
For both soil typesthe rangeof bulk densities(Table 1) compares to datafor forestsoilsfromthesameregion[Reiners et al., 1994].No heavymachinerywas usedfor forestclearing; traditional clearing techniquescaused very little surface disturbance.
100 , ; [ 80FC _ • 1' • I I
I
•
Biomass burningmayhavestimulated mineralization resulting insignificantly increased postbum NH4+concentrations onbothsoil types(Table3).Forest wascleared andslashwasburned earlierin
thedry seasonon theDystropept thanon theEutropept(Table2). We observeddifferenttracesof NH4+ accumulationin soil types with time after biomassburning(Figure2) primarilybecauseof precipitation. DuringthepostbumperiodtheDystropept received 50 , , !........ ' littlerain(4.3 mmthefirsttwo weeks,and27 mm thethirdweek); soil NH4+ concentrations increasedgradually.In contrast,on the -•
