GLOBAL BIOGEOCHEMICAL CYCLES, VOL. 12, NO. 4, PAGES 607-620, DECEMBER 1998
Winter fluxesof CO2 and CH 4 from subalpinesoilsin Rocky Mountain National Park, Colorado M. Alisa Mast, Kimberly P. Wickland,RobertT. Striegl,andDavid W. Clow WaterResourcesDivision,U.S. GeologicalSurvey,Denver,Colorado
Abstract. Fluxesof CO2 and CH 4 througha seasonalsnowpackwere measuredin and adjacentto a subalpinewetlandin RockyMountainNationalPark,Colorado. Gasdiffusionthroughthe snow wascontrolledby gasproductionor consumption in the soil andby physicalsnowpackproperties. The snowpackinsulatedsoilsfrom coldmidwinterair temperatures allowingmicrobialactivityto continuethroughthe winter. All soil typesstudiedwere net sourcesof CO2 to the atmosphere throughthe winter,whereassaturatedsoilsin the wetlandcenterwerenet emittersof CH4 andsoils adjacentto the wetlandwere net CH4 consumers.Most sitesshowedsimilartemporalpatternsin winter gasfluxes;the lowestfluxesoccurredin early winter, andmaximumfluxesoccurredat the onsetof snowmelt.Temporalchangesin fluxesprobablywere relatedto changesin soil-moisture conditionsandhydrologybecausesoil temperatures wererelativelyconstantunderthe snowpack.
AveragewinterCO2fluxeswere42.3, 31.2,and14.6mmolm-2d-• overdry,moist,andsaturated soils,respectively, which accounted for 8 to 23% of the grossannualCO2emissions from these soils.AveragewinterCH4 fluxeswere-0.016,0.274,and2.87mmolm-2d-1overdry,moist,and saturated soils,respectively.Microbial activityundersnowcoveraccounted for 12% of the annual CH4 consumption in dry soilsand58 and 12% of the annualCH4 emittedfrom moistand saturated soils,respectively.The observedrangesin CO2andCH4flux throughsnowindicatedthat winterfluxesare an importantpartof the annualcarbonbudgetin seasonallysnow-covered terrains.
seasonalsnowpackscommonly are warm enoughto support microbialactivitythroughthewinter,butrelativelylittle is known abouttheprocesses andratesof gasflux fromsnow-covered soils Landscapesat northernlatitudesand high elevationsaccount [Mellohand Crill, 1996;Sommerfeld et al., 1993]. in addition, for a significantpercentage of theEarth'slandsurfaceandinclude few studies havequantified theimportance of wintergasemissions arctictundra,borealforest,temperateforest,subalpineforest,and fromdifferentsoiltypesin termsof the annualcarbonbudgetsfor alpinetundra[Bouwman,1990]. Soilsin manyof theselandscapes theseecosystems [Brookset al., 1995]. havebeenidentifiedasimportantsourcesandsinksof atmospheric CO2fluxesfromsnow-covered soilsarelargelydueto microbial CO2thatmaybe particularlysensitive to climatechange[Oechelet respiration, althoughrootrespirationmay be importantin some al., 1997]. For example,soilsin arcticecosystems containlarge environments, andratesappearto be relatedto theeffectiveness of poolsof storedcarbonthat may now be actingas net sourcesof the snowpackin insulatingsoilsfrom winter air temperatures. CO2to theatmosphere becauseof increases in surfacesoiltemper- Sommerfeld et al. [1993]observed positiveCO2fluxesfromsubalaturesoverthe past3 decades[Oechelet al., 1997]. Ciais et al. pinesoilsin Wyomingthroughout thewinterwherethesnowpack [1995] recentlyidentified a strongterrestrialbiosphericsink of wasdeepenoughto preventsoilsfromfreezing.In thealpinezone CO2 in the northernhemisphere,althoughthe processof uptake in Colorado, CO2 fluxes were low in early winter becausea hasnotyet beenidentified.Wetlandsarethelargestnaturalsource shallow snowpackcausedsurfacesoils to freeze, then fluxes increased after additional snow accumulation allowed frozen soils of CH 4 to the atmosphere,and northern and high-elevation wetlandsmay accountfor a third of this naturalsource[Moosavi to thaw [Brookset al., 1997]. Winter CO2 emissionsfrom soilsin and Crill, 1996]. a borealforestecosystem in Canadaappeared to be derivedfrom a deepsoilsource because surfacesoilsfrozeowingto thecombinaAs muchas half of northernlatitudeand high-elevationlandtion of shallowsnowpacksand low air temperatures[Winstonet scapes aresnowcoveredfor mostof theyear[Sommerfeld et al., 1993];yet, mostattentionhasbeenfocusedon soil gasemissions al., 1995]. Zimovet al. [1993] andOechelet al. [ 1997] measured winterCO2emissions from soilsin thearcticzone. In theseenviin theseenvironmentsduring the growing season. Soils under ronments,extremelycold air temperaturescausedsoil temperaturesto dropwell belowthe thresholdfor microbialrespiration, Thispaperis not subjectto U.S. copyright.Publishedin 1998 and CO2emissionsprobablywere relatedto physicalprocesses by the AmericanGeophysicalUnion. thatreleased CO2producedduringtheearlypartof thecoldseason 1. Introduction
[Oechelet al., 1997].
Papernumber98GB02313. 607
608
MAST ET AL.: WINTER FLUXES OF CO2 AND CH4
Winter CH4 productionfrom microbialactivityin anoxicenvironmentsmay comprisea significantpart of the annualCH4budget for wetlands. Dise [1992] measuredCH4 fluxesfrom severalmoisture regimesin a Minnesotapeat bog and found that winter fluxes accounted for as much as 21% of the annual flux in this environ-
wetland. Vegetationat the wetlandsite gradesfrom drier grassy areasor subalpine forestaroundtheperimeterto sedge(CarexutriculataandCarexaquatilis)andarcticrush(Juncasarticus)toward the wettercenter.Transitionalzonesalongthe edgeof the wetland are covered by moss (Sphagnum spp.), tufted hairgrass
ment. Winter CH4 fluxes accountedfor 4% of the annual flux from
(Deschampsia cespitosa),andalpinetimothy(Phleumcommu-
a temperatewetlandin New Hampshire[Melloh and Crill, 1996] and between 4 and 43% of the annual flux from different plant communitiesin a subarcticmuskeg[WhalenandReeburgh,1988]. Snow-coveredsoils also may serve as sinksof atmosphericCH4 throughthe winter. CH4 consumption by forestsoilswasobserved to occur through winter in the subalpine zone of the Rocky Mountains [Mast et al., 1994; Sommerfeldet al., 1993] andin a borealecosystem in Canada[Wicklandand Striegl, 1997]. This study investigates soil CO2 and CH4 emissionsduring winter in the subalpinezone of the Loch Vale watershedin Rocky
tatum). Soils in the center of the wetland are anoxic near the surfaceand consistof a thick (at least 175 cm) peat layer that is covered with 5 to 10 cm of water during the growing season.
Mountain National Park, Colorado. Gas concentrations in the
snowpackwere measuredduring the snow-coveredseasonsof 1994, 1995, and 1996 as a meansof estimatinggas flux through snow. Physical properties of the snowpackand the soils were measuredduring the winter. Direct measurementsof flux also were made at the snow surfaceusing a static chambertechnique. Thesedataallowedus to (1) investigatethe effect of the snowpack on the exchangeof gasbetweenthe soil and atmosphere,(2) determine the magnitudeof andtemporalchangein CO2 andCH4 fluxes from different soil types during the snow-coveredseason,(3) calculatethe percentageof the annualflux occurringin winter,and (4) evaluate two methods of estimating gas flux through deep snowpacks. Becausethe study site has topographic,vegetative, and climatic characteristicstypical of many high-elevationand northern ecosystems,the results obtained from this study should have broad applicabilityto snow-coveredterrainsoMosier et al. [1993] notedthatthe globalextentof subalpineecosystems is not well defined but suggestedthat forest-meadow systemswith similarcharacteristics may accountfor asmuchas 12 x 106km2 of the Earth's land surface.
2. Study Site Description 40ø18', longitude105040') locatedin Rocky MountainNational Park, approximately80 km northwestof Denver, Colorado. The site is designatedas a United Nations BiosphereReserve[Baron, 1992] andcurrentlyis one of five sitesfundedby the U.S. Geological SurveyWater,Energy,and BiogeochemicalBudgetsProgram research on the movement
lessthan50 cm in depth [Baron, 1992]. 3. Methods
Wintergasfluxeswereestimatedusingsnowpackproperties andconcentration gradientsof CO• andCH4 in the snowpackin 1994,1995,and1996. In 1994,gasconcentrations weremeasured monthlyto bimonthlyat one wetlandsite andoneforestedsite fromJanuarythroughMay. In 1995,samplingbeganjust priorto snowmelt in earlyMay andcontinued biweeklythroughtheendof June. In addition to the two sites established in 1994, measurements were made at four additional sites in the wetland. In 1996, measurementefforts focused on the wetland. Ten sites were estab-
lishedalonga 120-mtransectthatcrossed a rangeof soilmoisture conditionsfrom drier areasalong the wetlandperimeterto saturated areas in the center of the wetland. Measurements were made
approximately monthlyfrom JanuarythroughApril andbiweekly throughthe snowmeltperiod.
In 1994, snowpackgaseswere sampledthroughfixed-interval samplersinstalled before significantsnow accumulationhad occurred. The samplerconsistedof 25-mm-diametersampling ports,coveredwith stainlesssteelscreens,that weresuspended abovethe groundon metalarmsthat extendedout 35 cm from a centralPVC pole. Pairsof samplingportswereplacedat the soil surface and at 45-cm intervals in the snowpack. The sampling
The studywas conductedin the Loch Vale watershed(latitude
to conduct field-oriented
Forestvegetation is old-growthEnglemann spruce(PiceaengelmanniiPargy)andsubalpinefir (AbieslasioscarpaHook) with a sparse groundcoverof myrteleafblueberry(Vaccimium myrtillus) [Arthur,1992]. Forestsoilsarewell-drainedCryoboralfstypically
of water and
carbonbetweenthe land and atmosphere[Lins, 1994]. Loch Vale is a subalpine/alpine basinthat drainsthe easternsideof the Continental Divide and rangesin elevationfrom 3000 to 4000 m. Land cover types in the basin include exposedbedrock surfaces,talus slopes,alpinetundra,andsubalpineforestsandwetlands.Climate is characterizedby long, cold winters and a short, 2- to 3-month growingseason.Averagedaily air temperaturesrangefrom -6.0øC in winter to 13.7øC in summer [Baron, 1992]. Average annual precipitation is 100 cm, and as much as 70% accumulatesin a seasonal snowpackbetween November and April. Snowmelt generally begins in early May, and snow cover may persist at higherelevationsinto July. Gas flux measurements were made at an elevation of 3200 m in
a small (0.4 ha) subalpinewetlandand in a forestedareanear the
portswereaccessed by nylontubing(2-mm outerdiameter)that ran vertically along the central PVC pole to the snow surface. Snowpackgasesalsowere sampledjust belowthe snowsurface througha thin metaltubeinsertedinto the snow. In spring1994 andin 1995 and 1996, snowpackgaseswere sampledthrougha 10-ram-diameterstainlesssteel probe 4 m in length that was insertedto differentdepthsin the snowpack.The probehad an inner bore diameter of 3 mm and was slotted over a 5-cm interval
near the tip. The probe and fixed-interval samplerproduced similarresults,buttheprobehadtheaddedadvantage of notbeing restrictedto a fixed site locationand samplinginterval. CO• was measuredin the field using a portableinfrared gas
analyzer(IRGA) that was calibratedwith 350 and 2010 ppmv standards carriedinto the field in mylarballoons.The precisionof the field IRGA measurementsdecreasedwith increasingconcen-
trationfrom 5% at 350 ppmvto lessthan 1% for concentrations over2000 ppmv. The IRGA wasequippedwith an internalpump that pulledgasfrom a samplingline attachedto the probeat a maximumrate of 0.3 L rain-•. Duringthe measurements, gasfrom the snowpackwaspulledthroughthe probeby the IRGA pump until the concentrationstabilized(usually15 to 30 s). After taking
MASTET AL.:WINTERFLUXESOFCO2 ANDCH4 the CO2reading,the samplingline wasattachedto a 20-cm3 nylon syringein which a samplewas collectedafter purgingthe syringe three times with air from the samplingline. The syringeswere returnedto the laboratoryfor CH4 analysisby gaschromatography, which was done within
48 hours of collection.
Gas standards
collectedin the syringesshowedno significantchangein concentrationduringa week of storagein the laboratoryor duringtransport from the field site to the laboratory,which was 1370 m lower in elevation. The gas chromatograph had a detectionlimit of 0.1 ppmv CH4 and a precisionof 2.5% for concentrationsbelow 2.0 ppmv and 1% for concentrations abovethis level. Physical snowpackmeasurementswere made at the wetland sitewithin 2 daysof eachgascollectiondate. The watercontentof the snowpackwasmeasuredfrom a verticalsnowpitfaceusinga 1-L stainlesssteelcutter,and snowtemperatureswere measured every 10 cm in the pit wall usinga dial stemthermometer(precision _+0.5øC).Temperatures weremeasuredhourlyat depthsof 1, 5, and 10 cm below the soil surface at one site in the forest from
March 1994 throughAugust 1995 and at two sitesin the wetland (one at the centerand one at the perimeter)from October 1995 through the end of the study,using permanently installed thermisters(thermocoupleswere used at the forest site) connectedto data loggers. Soil water content was measuredhourly at the perimeterof the wetlandfrom October1995 throughJune 1997, using a 15-cm-longwater contentreflectometer(probe CS615, Campbell Scientific, Inc.) insertedvertically into the soil. Air temperatureand snow depth were measuredcontinuouslyat a weatherstationlocatedin the wetland,and barometricpressure was measured at a weather station located 300 m from the wetland at an elevation of 3150 m.
Fluxesof CO2andCH4 from the snowwerecalculatedfrom gas concentrations in the snowpackusing a one-dimensionalform of Fick'slaw describingordinarygaseous diffusionin porousunsaturatedmediaas discussed by Striegl [1993],
dCA
qA= -Delldz
609
tortuosityof snowfrom imagesof cut, polishedsnowpacksections yieldedtortuosityvaluesof about0.9 for snowpacks havingporosities of about 80% (R. Davis, Cold RegionsResearchand EngineeringLaboratory,written communication,1997]. Table 1 lists the effective diffusion coefficientsused to calculateCO2 and CH 4 fluxes on each samplingdate. Calculated diffusivities for CO2 decreased from 1.25 m 2 d -1 in midwinter
to 0.53 m 2 d -1 in late
spring,whichprimarilyresultedfrom a reductionin porosityof the snowpackover the winter. The midwinter values are slightly higher thantherangeof 0.70 to 0.86 m2d4 reportedby Winstonet al. [1995] for CO2 diffusionthrougha midwinter snowpackin a boreal forest. The higher values calculated for Loch Vale are mostly due to the lower atmosphericpressureat this site, which increasesthe diffusivity of CO2 though air by 30% comparedto sealevel. Zimov et al. [1993] reportedexperimentallydetermined diffusivitiesrangingfrom 1.90 to 2.86 m2 d-• for wind-packed snowin Siberia. The valuesreportedby Zimov et al. [1993] are considerably higher than the diffusivity of CO2 through air, suggestingthat their experimentaldesignmeasuredadvectiveas well asdiffusivemovementof CO2throughthe snow. CO2 flux measurements also were made at the snowpack surfaceusinga staticchambertechniqueon selecteddates. The 55 x 65 cm rectangular chamber having a height of 15 cm was inserted 5 cm into the snow surfaceduring measurements. Air inside the chamber was mixed with a small pump, and CO2 concentrationswere measured by circulating the chamber air throughthe IRGA. CO2concentrations were monitoredevery60 s for up to 30 min. The maximumchangein concentration observed in the chamberheadspacewas 152 ppmv over a 16-min interval. Flux was estimatedfrom the linear portion of the concentration versus time curve, which usually included the first 8 min of concentrationdata. For all chamber measurements,the r2 of the
linear regressionwas greaterthan 0.98. Fitting the data with a nonlinear,diffusion-basedmodel, as suggestedby Anthonyet al. [1995], did not significantlychangethe flux estimates.
(1) 4. Results
whereqA is the massflux of gas A (molm-2d-l), Def t is the effective diffusion constant of gas A throughthe snowpack
(m2d-i);anddCA/dzis themeasured concentration gradient of gas A in thesnowpack (mol m-3m-i). The effectivediffusionconstants Deft for CO2andCH4 were estimatedon eachsamplingdateusingthe relation,
=
57•. Zq )
(2)
where 0 is the snowpackporosity,'cis the tortuositycoefficient, and DA is the diffusion constantfor gasA throughair at STP, correctedfor the averagedaily barometricpressure(atmospheres) and snowtemperature(degreesCelsius). Porositywas calculated from the density of ice (Pic½=0.9 I) and t•hewater contentof the snowpackover the gradient interval. Tortuosity is difficult to measureand,in soils,usuallyis describedasa functionof porosity with valuesrangingfrom 0m to 02/3[Striegl,1993]. In this study, the tortuosityof the snowpackwas estimatedfrom the theoretical
relation'c= 0m [Millington,1959],whichyieldedvaluesranging from 0.74 to 0.92. Thesevaluesare similarto the rangeof 0.74 to 0.90 reportedby Sommerfeldet al. [1996] for a subalpinesnowpackin Wyoming. A recentlydevelopedmethodfor measuringthe
4.1. Snowpack and Soil Characteristics
During the 1993-1994 winter, a seasonalsnowpackbegan to accumulateat the studysitein early Novemberthatremainedfairly shallow(< 50 cm) throughthe endof December.Owing to coldair temperaturesand shallow snow depthsin early winter, a layer of temperaturegradient(TG) snowdevelopedat the baseof the snowpack. The snowpack,which reacheda maximumdepthof 190 cm at the weatherstation,becameisothermalbetweentheApril 20 and May 5 samplingdates,and snowcoveredthe studysiteuntil early June. Snowpackcharacteristicsin 1994-1995 were substantially ulil•,•nt than in I '•'• ,•A owing to an ..•,nusu,•.y •.• warm periodin :,=.,-19=-, early spring. Averagedaily air temperatureat the wetland was above0øC at leastonehalf of the daysbetweenMarch 6 andApril 7, 1995, and snowaccumulationduring this period was minimal. Cool, wet weatherfollowed in April and May, and the snowpack nearly doubledin depthduringthis period. Owing to the warm weatherin early spring,a laterally continuous,2- to 5-cm thick ice layer developedat the snowsurfacethat wassubsequently covered by nearly 1 m of late-seasonsnow. Maximum snowdepthwas
610
MASTETAL.:WINTER FLUXES OFCO2ANDCHa
Table1. Average DailyBarometric Pressure, AirTemperature, Snow Depth, Snowpack Temperature, Snowpack Porosity,
Calculated Tortuosities andEffective Diffusivities Def t ofCO2 andCH4 forEachSampling Date Date
Atmospheric Air Snow Temperature, Snow 0 Pressure, Temperature, Depth, arm
øC
øC
q;
Deft CO2,Deft CH4, m2d-1
m2d-1
Jan.10, 1994 Feb.15, 1994 March10, 1994 April 1, 1994 April20, 1994 May 5, 1994 May25, 1994
0.672 0.681 0.679 0.681 0.689 0.689 0.689
-11.8 -4.8 -4.3 -0.7 5.0 7.3 2.9
95 135 155 170 160 160 65
-3.4 -3.2 -2.7 -1.1 -0•3 -0•3 -0.3
0.78 0.71 0.68 0.63 0.54 0.52 0.53
0.92 0.89 0.88 0.86 0.82 0.81 0.81
1.25 1.10 1.04 0.95 0.77 0.74 0.74
1.76 1.55 1.46 1.33 1.09 1.04 1.05
May 10, 1995 June1, 1995 June18, 1995
-0.648 0.689
-0.1 4.4 11.5
205 210 160
-1.0 -0.4 0.0
0.75 0.49 0.47
0.91 0.79 0.78
1.20 0.71 .64
1.69 1.01 .90
June 22,1995'
0.689
7.8
105
0.0
0.47 0.78
.64
.90
Jan.9, 1996 Feb.5, 1996
0.683 0.676
1.00 .95
1.40 1.34
-2.2 -7.9
cm
130 225
-5.1 -3.8
0.67 0.64
0.88 0.86
March 20,1996'
0.663
-10.2
310
-2.0
0.55 0.82
.80
1.13
April,8, 1996 May 7, 1996 May 15,1996 May 21, 1996 May 31, 1996 June4, 1996
0.681 0.684 0.686 0.683 0.684 0ø692
0.8 6.6 9.5 0.8 7.7 8.0
290 280 240 205 200 170
-0.7 -0.2 0.0 0.0 0.0 0.0
0.53 0.49 0.41 0.41 0.42 0.42
.75 .68 .53 .54 .53 .53
1.06 .95 .75 .75 .78 .78
0.0 0.0
0.42 0.75 0.42 0.75
.53 .53
.78 .78
June 11,1996' 0.691 11.4 115 June 18,1996' 0.692 11.4 75 *Snowpack temperature andporosity estimated forthese dates.
.81 0.79 0.74 0.74 0.75 0.75
210 cmonMay 15, 1995;coolairtemperatures delayed theonset record begins in April1996atthebeginning of thesnowmelt
of snowmeltuntiltheendof May, andsnowcoveredthewetland until earlyJuly. Duringthe 1995-1996winter,a seasonal snow-
water content to increase until the site was free of snow. The soil
pack began to accumulate the first week of November and
driedoverthecourseof thesummeruntilseveralheavyrainevents
period(Figure1). Meltingof the overlyingsnowpack causedsoil
increased slowlyuntilthemiddleof Januarywhenseveralstorm
rewet the soil in early September. Seasonalsnow cover was resystems deposited morethan50 cmof snowduringthefollowing establishedat the beginningof November1996, after whichsoil 2-weekperiod.Thisrapidaccumulation of snowtriggered a large watercontentincreasedgraduallythroughtheearlywintermonths.
avalanche on January30, 1996,thatdestroyed theforeston the edge of the wetland and covered the field site with avalanche
As in thepreviousyear,soilwatercontentincreased rapidlyonce snowmeltbegan.
debris.The snowpack, whichreacheda maximumdepthof 300cmonApril23, 1995,wasisothermal ontheMay7 sampling date, and snowpersistedat the wetlanduntil the third week of June.
Soilsin thecenterof thewetlandremained saturated andjust above0øC throughthe snow-covered season (Figure1). In all 3 years,a thickicelayer(10 cm)developed at thesurface of the
4.2. SnowpackCO2 and CH 4 Concentrations
Snowpack CO2andcH4concentrations varied temporally and spatiallyduringthe periodof study.Figure2 showssnowpack
CO2 profiles at a location near the perimeterof the wetland wetlandin earlyNovemberbeforesignificant snowaccumulation. throughthe 1996snow-covered season.Theseprofilesaresimilar Theicelayerremained intactuntiltheonsetof springsnowmelt to thoseobserved at theothersamplingsites,although theslopes whenit beganto meltunderthesnowpack. Soiltemperatures at andshapes of theprofilesvariedfrom siteto sitedepending on sitesin theforestandalongthewetlandperimeteralsoremained snowpackpropertiesandunderlyingsoil characteristics. Concencloseto 0øCthroughthe snow-covered season.The recordof soil trationsof CO2 at all sitesincreasedtowardthe baseof the snowwatercontentat thewetlandperimeter is discontinuous owingto pack,indicating thatsoilswerea sourceof CO2to theatmosphere equipment failureanddestruction of theinstrument installation by throughoutthesnow-covered period. Concentrations at thebaseof the January1996 avalanche.The mostcontinuous part of the the snowpackgenerallyincreasedthroughthe winteruntil the
MASTETAL.:WINTERFLUXES OFCO2ANDCH4 20
j
o,._,15_
contour plotof concentrations in thesnowpack onMarch30, 1996, just prior to peak snowaccumulationfor the year (Figure5). Snowpack depthsrangedfrom230 to 370 cm acrossthewetland owing to redistribution of snow by wind and the January avalanche.CO2andCH4 concentrations in the snowpackvaried significantly acrossthewetland. Althoughsomeof thevariation
snow covered
snow covered
?"
•
.
m10
......
0•
Dec
Feb
June 1996
Sept
Dec
Feb
611
June
1997
Figure 1. Daily soil temperature at 5 cm depth(solidline) in the centerof the wetlandand soil moisturecontentof the upper 15 cm (dashedline) at a sitealongthe wetlandpe•meter.
onset of snowmelt. Maximum CO2 concentrationsat the snow-
pack basein 1996 rangedfrom 2900 ppmv in the centerof the wetlandto 11,550ppmvin a grassyareaadjacentto the wetland. Theseconcentrations are comparableto thosereportedfor snowpacks covering subalpine and montane soils [Solomonand Cerling, 1987; Sommerfeldet al., 1996] but higher than those reportedfor snowpacks in alpineandarcticenvironments [Brooks
may be relatedto snowdepthand density,the patternappearsto correlatemoststronglywith soil moisture. Soil moisturedata acrossthe transectare availableonly duringthe summermonths [Wickland,1997],but the relativedifferencesbetweensitesarenot expectedto vary significantlyamongseasons.The highestCO2 andlowestCH4 concentrations were measuredfrom dry soilson a grass-covered riseat theedgeof thewetland(left sideof Figure5). Soils at this site have an organicmattercontentof 25% and an average volumetric watercontentduringsnow-free periodsof 35% in theupper10 cm [Wickland,1997]. The highestCH4 concentrationsand lowestCO2 concentrations were measuredin the center of the wetlandwhere soilsare coveredby standingwater during snow-freemonths.Vegetationat theselocationsis dominatedby sedgeand arcticrush,and soilshavean averageorganicmatter contentof 77%. CO2andCH4 concentrations at themarginsof the wetland were intermediatebetween the dry and saturatedsoils. Soilsin theseareasare moist,havingan averagevolumetricwater contentof 72%, and organicmattercontentsrangingfrom 71 to 77%. Vegetation at theselocationsis predominantly moss,tufted hairgrass,andalpinetimothy.
et al., 1997; Zimov et al., 1993]. Figure 2 also showsa seasonal
patternin the CO2concentration gradients(dC/dz)thatincreased throughthe midwintermonths(solidlines),reachinga maximum on May 7 at the onsetof snowmelt. These changesin gradient reflecta combinationof changesin snowpackdensityand in gas fluxesfrom the underlyingsoils.
SnowpackCH4 concentrations alsoincreased towardthebottom of the snowpackat samplinglocationsin the wetland,indicating that wetland soils were a net sourceof CH4 to the atmosphere throughthe winter (Figure 3). As observedfor CO2, concentrations of CH4 at the base of the snowpackgenerally increased throughthe winteruntil the onsetof springsnowmelt.Maximum CH4 concentrations at the soil/snowinterfaceat siteswithin the wetland rangedfrom 34 to 360 ppmv,which are comparableto concentrationsreported in a shallow snowpack overlying a temperatepeatlandin New Hampshire[Mellohand Crill, 1996]. The changein CH4 gradientover the seasonwas similarto CO2, with thegradients showinga pronounced increasepriorto theMay 7 samplingdate. CH4 concentrations overdrier sitesin the forest and alongthe wetlandperimetertypicallydecreasedtowardthe snowpackbaseduringthe midwintersamplingdates,indicating that thesesoilswere a sink of atmosphericCH4. Two sitesat the wetlandperimeterin 1995switchedfromnetCH4 consumption to net productionafter soils becamesaturatedby snowmelt,as evidoncodhy the reversalin gradientbetweenthe June1 and 22 samplingdates(Figure4). The reasonfor the apparentmidpack gradientreversalon June18 is unclear.It may havebeencaused by horizontaldiffusionof CH4 alongice layers from highconcentration zonesin the centerof the wetlandor, alternatively,to disturbanceof the concentrationgradientby a transientprocesssuchas wind pumping[Massmanet at., 1997].
Large spatial variationsin snowpackCO2 and CH4 concentrationswere observedacrossthe wetlandtransectas illustratedby a
4.3. Winter CO 2 and CH4 Fluxes
Gasfluxesthroughthe snowpack werecalculatedfor eachsite using(1), wherethe concentration gradient(dC/dz)wasdetermined from a linear fit of the snowpackconcentrationprofiles. Examplesof themostcommonprofileshapes observed duringthe studyperiodare shownin Figure6. Profilessimilarto the curve shown in Figure 6a were most commonly observedduring
300 ß
200
Jan
100
''-..May 7
_
Jun II
i %1•Apr
[]
Mar 20
i
0
1000
2000
3000
CO2 Concentration (ppm) Figure 2. CO2 snowpack concentrations duringthe 1996snowcovered seasonat a site along the wetland perimeter.Solid and dashedlinesrepresentprofilesmeasuredbeforeandafter the onset of springsnowmelt,respectively.
612
MASTET AL.-WINTERFLUXESOFCO2 AND CH4
2501
,
i
,
Wyoming [Sommerfeldet al., 1996], but slightly greater than
200•1••A•.2 150
'r•
100'•,
"•
'"-.[]
501 - •l• E'._•l • 'El May 15
•a'y
ß
I Jan919
.
• .....
Mar 20
fluxesfromalpinesoils(2.5 to 26.4 mmolm-2d-1)onNiwotRidge, approximately20 km southof Loch Vale [Brookset al., 1997]. The flux estimatesreportedby Sommerfeldet al. [1996] may be low becausethey did not correctthe diffusion coefficientfor the lower atmosphericpressureat their study site (3200 m). Winter fluxes measuredin this study generally were greaterthan those observedin high-latitudeecosystems.For example,winter fluxes rangedfrom 5.8 to 21.7 mmol m-2d-1from arctictundrasoilsin Siberia [Zimov et al., 1993] and 0.8 to 16.6 mmol m-2d-1 from arctic soils on the North Slope, Alaska [Oechel et al., 1997]. Winstonet al. [1995] reportedmidwinter CO2 fluxesrangingfrom 0.0 to 36.0 mmol m -2 d -1 from snow-covered
soils in the boreal
forestin Canada. Temporalpatternsin winter CO2fluxesin 1994 and 1996 are presentedin Figure 7. Fluxes at the moistand samrated siteswere lowestin midwinter,peakedat the onsetof spring CH4 Concentration (ppm) snowmelt, then decreasedto near premelt levels through the remainderof the snowmeltperiod. The largestincreasein flux Figure 3. CH4 snowpack concentrations duringthe1996snow- occurredat the moist sitesin 1996, which nearly tripledbetween covered season at a site in the center of the wetland. Solid and the February5 andMay 7 samplingdates. The seasonalpatternat dashedlines representprofilesmeasuredbeforeandafterthe onset dry siteswassimilarexceptthatthe late springfluxesweresignifiof springsnowmelt,respectively. cantly lower than the early winter rates. The magnitudeof and changesin fluxes at the dry and saturated sites were similar between 1994 and 1996 despitedifferencesin sitelocationsandin midwinterconditions whenthesnowpack densityandstratigraphy the timing anddepthof the snowpack. were relatively homogeneous.In thesecases,the concentration Averagewinter CH4 fluxesfrom dry, moist,and saturatedsoils gradientwas calculatedfrom a linear fit of the entire concentration profile. Curvilinearprofilesas shownin Figure6b wereobserved were -0.016 (s.d. = 0.009), 0.274 (s.d. = 0.288), and 2.87 (s.d. = duringthe 1996 snow-coveredperiod as a resultof increasing 2.99) mmolm-2d-1,respectively.Fluxesfromindividualsitesover
0 |•-- ß Jun, 4 •y• •, 0
50
,
100
150
snowdensitywith depth. In thesecases,the concentration gradientswere calculatedfrom the upperhalf of the snowpackwhere theprofileswerecloseto linear.Althoughnotusedin thisstudy,a secondapproachwouldbe to correctthe snowpack concentrations for changesin snowporositywith depthusinga finite difference model [Ishii et al., 1989] beforefitting the linear regression.A midpackice layer,whichdeveloped in spring1995,restricted gas diffusionthroughthe snow,causingthe steppedprofileshownin Figure6c. For theseprofiles,the concentration gradientwasestimatedfromtheportionof thegradientabovetheice layer. For the nonlinear profiles shownin Figures 6b and 6c, inclusionof all pointsin the gradientcalculationor a two-pointcalculation using
the studyperiodrangedfrom-0.035 to 0.016 mmolm-2d4 for dry soils, -0.014 to 2.08 mmol m-2d-1for moist soils, and 0.046 to 13.9
mmolm-2d'• for saturated soils. CH4consumption ratesmeasured over dry soilswere similarto the rangeof-0.004 to -0.016 mmol m-2d4 reportedby Sommerfeld et al. [1993] for subalpinesoilsin Wyomingduringwinter. Fluxesfrom saturatedsoils were similar to average winter fluxes from a temperatepeatland in New Hampshire,which rangedfrom 1.3 to 3.5 mmolm-2d4 [Melloh and Crill, 1996], and from an openbog site in Minnesota,which
base of snow and snow surface concentrations would have resulted
in an overestimation of gasfluxesby as muchas 25% because concentrations towardthe baseof the snowpack wereelevatedby variationsin snowpackstructureanddensity. Gradientestimates of CO2andCH4 fluxesduringthe snow-covered periodsof 1994, 1995,and1996aresummarized in Table2. Samplelocations were groupedaccordingto relative soil moisturecontent[Wickland, 1997]. Saturatedsoilsare in areasof the wetlandcoveredby
25O
+
200
May 10/+
150-
/•.,.P Jun 1 ,¾
standingwater; dry soils include forested and unforestedareas
outsidethewetland,andmoistsoilsaretransitional areasalongthe wetlandperimeter. Fluxes on each date representone to four measurements for eachmoisturegrouping. AveragewinterCO2emissionsfrom dry, moist,andsaturated soilswere 42.3 (standarddeviation(s.d.) = 17.7), 31.2 (s.d. = 11.5), and 14.6 (s.d.= 10.3) mmolm-2d-1,respectively.Fluxes from individualsitesover the studyperiodrangedfrom 16.0 to 76.0 mmolm-2d-• for dry soils,3.0 to 101 mmolm-2d-• for moist soils, and 1.2 to 36.5 mmol m'2 d-• for saturated soils. Fluxes from
dry and moistsiteswere similar to averagewinter fluxesfrom subalpinesoils(27.5 to 73.0 mmolm2 d) in the SnowyRange,
100ß Jun 22 50ß ß
0 0.0
, +/ I
•,J, un18i
1.0
2.0
ß
3.0
CH4 Concentration (ppm) Figure 4. CH4 snowpackconcentration profileson four dates in 1995 at a site alongthe wetlandperimeter.
613
MASTET AL.:WINTERFLUXESOFCO2 AND CH4 375
300 225
øø
150
co,
75
0 375
300 225
-_
30 4O
150
0•
0
dry
moist
saturated
moist- dry
Figure 5. Contourplotsof CO2 andCH4 snowpack concentrations alonga 120m transectthroughthewetlandon March 30, 1996.Tick markson thex axisrepresent relativepositionsof the 10 samplingsitesalongthetransect.
ranged from 0.3 to 3.7 mmol m-2d-• [Dise, 1992]. Temporal
5. Discussion
patterns in winterCH4fluxin LochValedifferedamongsoiltypes. At the dry sitein 1994 (Figure8b), CH4uptakewasgreatestduring
midwinterthendeclinedto nearzeroduringthe snowmelt period. 5.1. Effect of Snow Cover on Winter Gas Fluxes The seasonal variationat thedry sitein 1996waslesspronounced, althougha slightdeclinein flux wasobservedat the onsetof snow-
melt (Figure 8d). Moist and saturatedsoils showeda slight increase in CH4 flUXthroughthemidwintermonths,followedby a dramaticincrease in fluxat theonsetof snowmelt(Figures8a and 8c). The largestchangein flux occurredat the saturatedsite in 1994, which increasedby an order of magnitudebetweenthe February15 and May 25 samplingdates. In 1996, the flux from saturatedsoilsdecreasedthroughthe remainderof the snowmelt period,asidefrom a secondincreasein early June. At the moist sitesin 1996, the flux returnedto premeltlevelsfollowingthe initial releaseandthenincreased slightlythroughtheremainder of the snow-covered period. On severalsamplingdates in 1995, gradient and chamber measurements
were made at the same locations
in order to
comparethesetwo methodsof estimatinggasflux throughsnow. The resultsof eightpairsof measurements madeoverdifferentsoil typesarepresented in Table3. CO2gradientsin the snowpackat the eight sitesrangedfrom -27 to -105 mmol m-4,and the correspondinggradientflux estimatesrangedfrom 17.0 to 62.7 mmol
m'2d'•. Chamber fluxestimates madeat thesamesitesranged from5.5 to 21.1mmolm-2d-1andwereconsistently lowerthanthe gradientflux estimatesby nearlya factorof 3.
Recentstudieshave shownthat gas transportthroughseasonal snowpackoccurslargely by diffusion and that fluxesat the snowpack surfaceprimarily are controlledby biologicalactivityin the underlying soils [Sommerfeldet al., 1996; Brooks et al., 1997; Winstonet al., 1995]. Within-snowprocesses, suchas biological productionor consumption,adsorptionon ice, and dissolutionin water, are thoughtto have a minor effect on gasexchangebetween the soil and atmosphereduring winter [Sommerfeldet al., 1996; Brookset al., 1993]. Pressurefluctuationsat the snowpacksurface caused by wind may enhance gas transport through snow; however,the importanceof this processin termsof the total winter flux is poorly understood[Albert and Hardy, 1995; Massman et ai., i995' Massmanet ai., 1997]. Comparisonof snowpackCO2 profiles and snowpack physical characteristics in Loch Vale showedvariationsone would expectif gas exchangethroughthe snowpackwascontrolledprimarilyby porosityandtortuosityvariationswithin the snowpack(Figure 6). The nearlylinear concentrationgradientin Figure6a is typicalof mostsamplingdatesand reflectsthe lack of structureand relatively narrowrangeof densities in midwinter snowpacks. A more detailed analysisof this profile,however,showeda subtlebreakin slopeat the60-cmlevel.
614
MASTETAL.'WINTERFLUXESOFCO2 ANDCH4 I
I
I
I
f
(a) 150
_
ß ß
& •oo
m
50
I
I
I
I
(b) 150
!
!
!
i
a:: 100
m
50
0 I
f
(c)
2.0
,
E 15 '
•
ice lense •
1.0
0.5
0.0
• 0
1000
I
2000
3000
CO2 (ppm)
0
2oo
f
400
600
Density (kgm-3)
Figure 6. CO2 snowpack concentration profilesin (a) 1994,(b) 1996,and(c) 1995withmeasured snow-density profiles.ET representsa layer of equitemperature snow,andTG a layerof temperature gradientsnow.
(ET) layer. ET metamorphismoccursin responseto a vapor-pressure difference on the individual snow grains and results in a strengtheningof the snowpack due to the formation of necks betweenindividualgrains[Sommerfeldand LaChapelle, 1979]. In contrast,TG metamorphismoccursin responseto a temperature gradient acrossthe snowpackthat results in a loss in snowpack cohesionas the individual grains enlarge at the expenseof the necks[Sommerfeldand LaChapelle, 1979]. This lossof cohesion between grains may have reduced the tortuosity of this layer, which causedthe gradientto be slightlyshallowerin the lower half of thesnowpack.ConwayandAbrahamson[ 1984] founda similar relation between air permeability and snow metamorphic type; increasingpermeabilityappearedto be associatedwith TG metamorphismand decreasingpermeability with ET metamorphism. On severalsamplingdatesin 1996, profiles in the wetland had a distinctcurvilinearshapeas shownby Figure 6b. The snowpack characteristicthat appearedto influencethe shapeof the gradient was a distinctincreasein snowdensitywith depth,which probably was causedby the unusuallyrapid accumulationof snow during the month of January. For a few profiles,the concentrationswere adjustedfor changesin porosityand tortuositywith depth which causedthe profilesto straighten,indicatingthat snow densitywas the primary factor controllingthe profile shape.
Severalstudieshavesuggested that ice layersand melt features within the snowpackmay havea significanteffecton gastransport from snow-covered soils[Hardyet al., 1995; Winstonet al., 1995; Melloh and Crill, 1995, 1996]. An ice layer was observedto have a dramaticeffecton gasconcentration profilesin LochVale during the springof 1995. The profilein Figure6c showsa largeincrease in CO2concentration at the samelevel as a thickice layerlocated midway throughthe snowpack. Concentrationgradientsabove and below the ice layer were linear, althoughthe gradientwas slightly steeperbelow than above the ice layer. This profile demonstrates thattheice layerhada considerably lower diffusivity than the surroundingsnowbut did not block gas flux from the snowpacksurface. The basal ice layer that developedat the wetland surfacealso was sufficientlypermeableto allow continuousgasexchange betweenthe soil andatmosphere as evidenced by theCO2andCH4 gradients measured in the snowpack overthe wetlandthroughoutthe snow-covered period. Melloh and Crill [1996] also observedthe buildup of CH 4 below a midpack ice layer; however,near-ambientsnowpackCH4 concentrations above the layer indicatedthat the ice inhibitedgasflux at the snowpack surface. At a few sitesalong the wetland perimeter,high CH4 concentration layerswereobservedbetweenice layersthat developedin thespring.Becausea similarpatternwasnotobserved for CO2, we believethat thesehigh-concentrationlayers may have resultedfrom horizontaldiffusion of CH 4 between the ice layers from highconcentration zonesin the centerof the wetland.These resultsshowthat snowpackstructurescan, at least temporarily, affect the exchange of gas between the soil and atmosphere; however,oncethesnowpackbeginsto melt, gastrappedbeneathor betweenice layerswill eventuallybe released[Mellohand Crill,
If separate regression linesarefit to thetwo regions,theconcentra- 1996]. tiongradientabove60 cm (-25.5 ppmvcm-],p < 0.001)is slightly Perhapsthe mostimportantrole the snowpackplaysin controlgreaterthan the gradient below 60 cm (-21.0 ppmv cm-], ling wintergasproductionor consumption is providinginsulation p < 0.001). The changein slopeappearsto reflecta changein for soilsfrom cold midwinterair temperatures.Despiteaverage snowmetamorphic type ratherthansnowdensity.The snowpack air temperaturesin Loch Vale of-9øC from November through stratigraphyindicateda layer of a coarse-grained temperature February,soil temperatures remainedslightlyabove0øC and well gradient(TG) snowoverlainby a fine-grainedequitemperature abovethe minimumtemperaturerangeof-5 ø to -7øC reportedfor
MASTETAL.'WINTERFLUXES OFCO2ANDCH4
615
Table2. AverageWinterCO2 andCH4 FluxesFromDry, Moist,andSaturated Soils
Date
Average DailyCO2,Flux,mmol m'2d'1
Dry
n
Moist
n
Saturated
Dry
n
Moist
Saturated
n
Jan. 10, 1994
34.8
1
--
0.9
1
Feb. 15, 1994
29.0
I
--
4.1
1
-0.035
1
--
0.60
1
1
-0.029
1
--
1.64
1
1
-0.026
1
--
1.39
1
1
0.000
1
--
2.92
1
1
0.000
1
--
3.63
1
1
-0.006
1
--
2
-0.026
2
2
0.53
March 10, 1994
April 1, 1994
April 20, 1994 May 5, 1994 May 25, 1994 May 10, 1995
45.6
53.3
56.7 71.8 16o0 57.3
1
n
Average DailyCH4Flux,mmol m'2d-1
--
1
4.6
--
1
10.4
--
1
14.8
--
1
38.2
--
3
33.6
--
20.4
June 1, 1995
47.1
2
29.1
1
12.2
3
-0.011
June 18, 1995
4700
3
38.6
1
32.4
2
-0.015
3
0.16
2
--
2
June 22, 1995
52.8
3
26.7
1
17.0
1
-0.021
Jan. 9, 1996
50.9
1
28.4
4
10.1
4
-0.019
4
--
4
-0.012
Feb. 5, 1996 March 20, 1996
April 8, 1996 May 7, 1996
-62.2
76.0 37.0
12.4 1
1 1
16.5
33.7 63.0
4 4
4 4
13.4 8.7
9.6 26.0
.060
13.9
1
2.84
1
2
4.56
2
2
6.10
2
1.77
2
4
0.64
4
0.18
4
0.92
4
2
0.22
4
0.68
4
4
-0.011
1
0.17
4
1.55
4
3
-0.013
1
1.2
4
5.52
3
I
0.18
4
1.88
4
May 15, 1996
17.7
1
33.5
4
23.0
4
-0.021
May 21, 1996
30.2
1
29.8
4
10.4
4
-0.014
I
0.13
4
1.68
4
4
-0.018
1
0.23
4
1.45
4
3
-0.010
I
0.15
4
3.53
3
1
0.15
4
0.73
4
1
0.25
1
2.27
2
May 31, 1996 June 4, 1996
39.1 17.4
I I
35.5 33.5
4 4
10.6 9.6
June 11, 1996
21.0
1
29.2
4
5.3
4
-0.023
June 18, 1996
25ø8
1
27.5
1
5.1
2
-0.019
n denotes number of measurements.
microbialrespiration[Flanaganand Bunnell,1980; Coxsonand Parkinson, 1987; Brooks et al., 1996]. Without a snow cover, most
surfacesoilsin the studyareacertainlywouldfreezeandmicrobial activitywould be greatlydiminishedin winter. The effectiveness of the snowpackin insulating soils in different environments appearsto be controlledby snowpack depthandaveragewinterair temperature. In the subalpinezone in Loch Vale, snowpacks generallyaredeepenoughto preventsoilsfrom freezingduringthe early winter when averagedaily air temperatures are aslow as -9.0øC. Similarresultswerereportedby Sommerfeldet al. [ 1996] for subalpinesoilsin Wyoming,wheresoiltemperatures remained closeto 0øC once a consistentsnowpackhad been establishedin early winter. In contrast,Brookset al. [1996] found that alpine soilson Niwot Ridge froze early in the winter owingto a shallow
averageair temperatures werefoundto havehigherCH4emissions than thosehaving higher air temperatures[Melloh and Crill, 1996]. The authorssuggested that a deeperand more continuous snow cover, maintainedby lower air temperatures,limited frost
penetration, whichresulted in a longerwinterperiodof microbial activity in yearshaving lower averageair temperatures.Dise [ 1992] observeda similar relationbetweenfrost depth and snow coverin a Minnesotapeatland.In a yearwith heavysnowfall,the peat surfaceneverfully froze,but in a subsequent dry year, the peatfrozeto a depthof 30 cm at somelocations.In high-latitude ecosystems, wintersoiltemperatures generallyarebelowthe-5øto
-7øCthreshold forbiological activity owingto thecombination of relativelyshallowsnowpacks andextremelylow air temperatures. In the boreal regionof Canada,Winstonet al. [1995] observed
soiltemperatures aslowas-14øCin midwinter, which snowcover, andsurface soiltemperatures fellaslowas-14øC. surface
As snowbeganto accumulatelater in winter,the underlyingsoils warmed to above-5øC and microbial activity was observed. '•'•'""" •'""" Rock y •'• .... •1" s•tes ' ...... ;,-h,,, ;,. •aa m o •, elevation ............ ............ ...... andprobablyreceivecomparableamountsof snowfall. However, the Niwot siteis on an alpineridgethatis subjectto wind scouring, whereas the Loch Vale and Wyoming sites are in areas that are more conducive to snow accumulation.
At lower elevations,the periodof snowcover tendsto be less consistentthroughthe winter and the effectivenessof the snowpackasaninsulatormayvaryfromyearto year. For example,in a temperatepeatland in New Hampshire, winters having lower
causedCO2 productionto ceasein the shallowsoil zone. The frozensurfacesoils,however,remainedgaspermeablebecauseof low soil moisture [Wicklandand .•trD,g!, !997], which allowed CO2fromdeepersoilsources to diffuseto thesoilsurface[Winston et al., 1995]. In arcticecosystems in Siberiaand the North Slope, Alaska, soil temperatures as low as -20øC were observedat soil depthsas deepas 100 cm duringthe midwintermonths[Zimovet al., 1993; Oechelet al., 1997]. During this period, soil temperatureswere too low for biologicalactivityand CO2 flux was attributedto physicalprocesses thatreleasedCO2producedduringthe earlierpartof thecoldseason[Oechelet al., 1997]. On theNorth
616
MAST ET AL.' WINTER FLUXES OF CO2 AND CH4
80
(a)
1994 i
dry
, ,
,
cq20 ¸ saturated I
I
80
_---,
1996
,
'
t
n:•
(b) '
• 60- dry_
o _ moist ß •
•
ß
4.0-
0
carbon substrateswere released into the soil, which stimulated the
20
. saturated
Jan
n []
Feb
Mar
Apr
May
Jun Jul
Figure 7. Temporalpatternsin winter CO2 fluxesfrom dry, moist, and saturatedsoils in (a) 1994 and (b) 1996.
Slope, Alaska, thereis also a deepergeologicsourceof CO2 that reachesthe surfacethroughfracturesin the bedrock[R. G. $triegl and R. W. Healy, U.S. GeologicalSurvey,written communication,
growthof microbialpopulations.This mechanismprobablydoes not apply to the soilsin this studybecausesoil temperatureswere observedto stay above0øC throughthe winter and there was no thawingeventto releasesubstrates into the soil. An early winter minimumin CO2 fluxeswas also observedby Sommerfeldet al. [1996], who suggestedthat temporaltrendsin winter gas fluxes might be driven by changesin soil moisture. Data from the soil moistureprobeat the edgeof the wetlandindicate that soil moistureduring the winter was at a minimum at the time a permanentsnowcoverwas establishedin early November, then increasedgraduallythroughthe midwinter until the onsetof snowmelt in early May (Figure 1). This is consistentwith the early-winter minimum and gradualincreasein CO2 flux prior to snowmeltat the dry andmoistsites(Figure7). The mechanismfor the midwinter
1997].
5.2. Controls on Spatial and Temporal Variations in Winter CO2 and CH 4 Fluxes The maintenanceof above-freezingsoil temperaturesby the insulatingsnowpackis the primary reasonthat microbialactivity persistedthroughthe winter in Loch Vale soils. However,spatial patternswere observedin winter gasfluxesthat couldnot be attributed to the effects of soil temperature,which remainednearly constantthroughthe snow-covered season.The spatialpatternsin winter CO2 and CH 4 fluxes in Loch Vale were similar to those observedat this site during snow-freemonthsand appearto be controlledprimarily by soil-moistureconditions[Wicklandet al., 1996; Wickland,1997]. The highestCH4 flux was foundin saturated areas of the wetland
perimeter. The dominant sourceof CO2 in winter from aerobic soils at high elevationsis the decompositionof organicmatterby heterotrophic microbial communities [Brooks et al., 1997]. Heterotrophicrespirationprobablyis alsothe sourceof CO2 in the center of the wetland, but the fluxes may be lower becauseless oxygen is available to surfacesoils in the wetland environment. CO2 from root respirationin the wetlandwas assumedto be negligible in the winter. In additionto spatialtrends,CH4 and CO2 fluxesin Loch Vale varied temporally through the winter. Most sites had similar patterns for CO2 and CH4; fluxes were lowest in early winter, increasedto maximumvaluesat the onsetof springsnowmelt,then droppedagainthroughthe remainderof the snowmeltperiod. The exception was CH 4 consumption at the dry sites, which was greatestin early winter and declinedduring snowmelt. Because soil temperaturesremainedwithin a very narrow range over the measurementperiod,otherprocesses are probablycontrollingthe observed temporal trends in emissions. Brooks et al. [1997] proposedthat increasesin winter CO2 fluxesfrom alpine soilson Niwot Ridge were drivenlargelyby substrateavailability. In this alpine environment,soils froze early in the winter then subsequently thawedwith continuedsnowpackaccumulation.Brooks et al. [1997] hypothesizedthat, followingthe thawingevent,labile
where conditions
were favorable
for
methanogenesis. At dry sites outside the wetland, CH4 was consumed through most of the winter, indicating soils were aerobicand favorablefor methanotrophs.The patternof CO2 flux acrossthe wetlandtransectwas oppositethe patternof CH4, with the lowestfluxesin the wetlandcenterandthe highestfluxesat the
increase in soil moisture is unclear since free water
typicallyis not presentin the snowpackuntil late April or early May. Sommerfeldet al. [1996] hypothesized that changesin soil moisturein early winter mightbe drivenby temperaturegradients in the snowpackthat causevaporto be transportedfrom the soil towardthe snowpack.Zimovet al. [1993] observedan increasein soil moisturein the upperlayer of arcticsoilsin early winterthat was attributedto the upwardmigrationof water from deepersoil horizons toward the frozen surface.
Once the snowpackbecomesisothermalin spring, there is a large influx of meltwaterinto the soil environment. At dry sites, CO2 fluxes decreasedsignificantly after the onset of snowmelt, perhapsbecausediffusionof oxygenand CO2into and out of the saturatedsoils was limited. The onsetof snowmeltalso appeared to causeCH4 consumptionto slowin aerobicsoils(Figure8), and, in somecases,soils becamenet CH4 producersafter being saturatedby snowmelt(Figure4). Mosier et al. [ 1993] notedthe opposite patternin subalpinemeadowsoils during snow-freemonths where drying after snowmeltcausedsoilsto switch from net CH4 productionto CH4 consumption.In the centerof the wetland,we observedlarge releasesof CH4 that were coincidentwith the onset of snowmelt(Figures8a and 8c). We believethis episodicrelease
MAST ET AL.: WINTER FLUXESOF CO2 AND CH4
617
15.0 [2
1994
(a)
1996
/ / / /
10.0
/ / / / / /
5.0
/
saturated
saturated I
0.0
•
mø1stl •
•
0.00
'-
1994
(b)
'.
1996
(d)
i
"•
-0.01
!
o
-0.02
•
-0.03
dry
dry •
-0.04 -0.05
Jan
Feb Mar
Apr May
Jun Jul
Jan
Feb Mar
Apr
May
Jun Jul
Figure 8. Temporalpatternsin winterCH4 fluxesfrom(a) saturated and(b) dry soilsin 1994,andfrom(c) moist and saturated,and (d) dry soils in 1996.
of CH4 may be relatedto hydrologicchangeswithin the wetland ratherthan increasesin microbialactivity. One possiblemechanismfor episodicreleaseis meltingof the ice layeron the wetland surfaceand subsequentreleaseof gastrappedbeneathit. This is consistentwith our snowpitobservations that the basalice layer began melting shortly after the snowpackbecameisothermal. Anotherhydrologicprocessthat might causeepisodicreleaseis theinfluxof meltwaterintotheshallowgroundwater system.Infiltratingsnowmeltfrom talusslopesabovethe wetlandmightpush older,CH4-richwatersto the wetlandsurface,causingsubsequent degassingand releaseof CH4 into the overlyingsnowpack.This hypothesisis supportedby isotopicand geochemicalstudiesin LochVale, whichshowthatstreamflowduringthe initial stagesof snowmeltis derivedprimarilyfrom premeltgroundwaters pushed into the stream channel by infiltrating snowmelt[Mast et al., 1995' Campbellet al., 1995]. Followingthe episodicreleaseof CH4,fluxesdeclinedat boththe moistandsaturated sitesthrough the remainderof the snowmeltperiod. This declinemay be the result of the influx of oxygenatedsnowmeltinto the anaerobic wetland soils that shouldcauseCH4 productionto decreaseand CH4 consumptionto increase.
5.3. Winter GasFluxesas a Percentageof Annual Budgets Several recent studies have determined that winter fluxes of
CH4 andCO2fromsnow-covered soilscomprise animportantpart
of the annual budgetsof these gases [Dise, 1992; Sommerfeld et al., 1993; Brooks et al., 1997; Oechel et al., 1997]. In Loch
Vale, total winter CO2 emissionsin 1995-1996 (November 1 to June18) for dry, moist,andsaturatedsoilswere 11.4, 6.5, and 2.7 mol m-2,which accounted for 23, 12, and8%, respectively, of the grossannualCO2 emissionsfrom thesesoils [Striegl et al., 1996; K. P. Wickland, unpublished data, 1997]. Total winter CO2
Table 3. Comparison of CO2 FluxEstimates Based
onConcentration Gradients intheSnowpack (Fg) andChamberMeasurements (Fc) Made on the
SnowpackSurface
Date
Fg
Fc
Fg/F c
May 10
59.8
19.6
May 10
55.6
18.2
3.1 3.1
June 1
20.6
8.6
2.4 3.1
June 22
31.3
10.0
June 22
62.7
18.8
3.3
June 22
57.2
21.1
2.7
June 22
17.0
5.5
3.1
June 22
26.7
10.2
2.6
FcandFgin unitsof mmoles m-2d-•.
618
MAST ET AL.: WINTER FLUXES OF CO2 AND CH4
emissionsat otherhigh-elevationsitesin the Rocky Mountains rangefrom 1.1molm-2for alpinesoils[Brooks etal., 1997]to 10.9 mol m-2 for subalpinesoils [Sommerfeldet al., 1993], which accountsfor as much as 25% of the carbon fixed by primary
producersduring the growing season[Brookset al., 1997; Sommerfeld et al., 1993].Total winterCH4 fluxesin LochVale from dry, moist,and saturatedsoilswere-3.7, 50.3, and315 mmolm-2,respectively. Usingmeandailyflux valuesfromWickland et al. [ 1998 ] for snow-freemonths,winter fluxeswere calculatedto accountfor 12% of the annualCH4 consumption by
dry soils,and58 and 12% of the annualCH4produced by moist
Severalstudies havereported thatchamber measurements made at thesoilsurfacesystematically underestimate truefluxesbecause increasingconcentrations in the chamberheadspace distortthe concentration gradientnearthe soil surface[Hutchinson and Mosier,1984;Juryet al., 1982;Mosier,1989;Livingston and Hutchinson; 1994]. Usinga 3-D simulation, Healy et al. [1996] demonstrated that chamber measurements made at the surface of
porous mediaconsistently underestimated truefluxesandthatthe biasincreasedwith increasingair-filledporosity. For example, theirmodelpredictedratiosof field-based chamberfluxesto true fluxesrangingfrom 0.955 for a porosityof 0.1 to 0.435 for a
of0.5. Theaverage chamber fluxFctogradient fluxFg andsaturated soils,respectively.The differences in thepercentage porosity ratio of 0.372 (s.d.= 0.095) calculatedfrom our field dataagrees of winter CH4 flux amongthesethree soil typesappearto be byHealyet al. [1996] relatedto the responseof microbialactivity to soil temperature fairlywellwiththeratioof 0.435predicted duringsnow-free months[Wickland,1997]. The resultsfor saturated soils in Loch Vale are similar to those at a Minnesota peat-
land, where winter fluxes accountedfor as much as 21% of the annual CH4 flux [Dise, 1992], but slightly higher than at a
temperate peatlandin New Hampshire, whereonly4% of annual CH4emissions occurredunderthe snowpack [Mellohand Crill, 1996].
5.4. Comparison of Gradient and Chamber Flux Estimates
There is no widely acceptedmethod for estimatinggas flux through seasonal snowpacks, and most previous studies have utilized one of the following two techniques:(1) measurementof snowpackconcentrationgradientsand applicationof a one-dimensionalflux model [Solomonand Ceding, 1987;Sommerfeld et al., 1993] or (2) direct measurementof flux at the snowsurfaceusinga closed-chamber technique[Whalenand Reeburgh,1988; Winston et al., 1995]. Fluxes in this studywere determinedby measuring snowpack concentration gradients; however, several closedchambermeasurementswere madein 1995 in an effort to compare thesetwo methodsof measurement.The datapresentedin Table3, whichcomparegradientandchamberflux estimatesat eightlocations, show that the chamber fluxes were consistentlylower than the gradientfluxesby nearlya factorof 3. Becauseuncertainties in 0, 'r, andgasconcentrations generallydo not introducemorethana 15% errorin gradientflux estimates[Sommerfeldet al., 1996], the discrepancybetween these two methodsindicatesthat chamber measurements at the snowsurfaceprobablyunderestimate gasflux throughsnow. One possibleexplanationfor lower chamberfluxes is that an adequatesealwas not achievedbetweenthe chamberand the snowsurface[Livingstonand Hutchinson,1994]. On a few occasions, we observed that the concentration in the chamber
headspacewas reset to ambientlevels by stronggustsof wind, which indicated that the seal was not adequateon windy days. Dise [1992] reportedthat an adequateseal could not be made betweenthe chamberand snow surfaceduring winter CH4 flux measurements at a Minnesotapeatland. Whalenand Reeburgh [ 1988] usedchambersto measurewinter CH 4 fluxesfrom subarctic muskeg communities. To establish an airtight seal, permanent chamber bases were installed in the soil to which vertical sections
were addedthroughthe winterto stayevenwith the snowsurface. While this techniqueappearsto work well in areaswith shallow snowpacks,it would be impractical in the Rocky Mountains, whereseasonalsnowpackstypicallyreachseveralmetersin depth.
for a porosityof 0.5. Becausethe snowpack porosityaveraged 0.52 duringour measurements, the modelappearsto providea reasonableexplanationfor the discrepancybetweenthe two methodsof measurement. Winstonet al. [1995] also compared
chamberandgradientestimatesof CO2flux throughsnowbut
observed a widerange offluxratios (Fc/Fg=0.2to5.5),some of which indicatedthe chamberfluxeswere significantlyhigherthan
thegradient fluxes.The higherchamber fluxeswereattributed to thepresence of meltchannels andtreewellsthatmayhavepreferentiallychanneled gasfromthe soilto thesnowsurface[Winston et al., 1995; Hardy et al., 1995]. These resultssuggestthat gradient measurements maybe inaccurate undersomeconditions, particularly in areaswithshallow orheterogeneous snowpacks.
6. Summary and Conclusions CO2andCH4fluxesweremeasured fromsnow-covered soilsin and adjacentto a subalpinewetlandto investigatethe processes controllingtheexchange of gasbetweenthe soil andatmosphere duringwinter.Gasconcentration profilesin thesnowpack varied with thephysicalproperties of the snow. Ice layersappeared to slowthediffusionof gasthroughthesnowpack butdidnotinhibit gasflux at the snowsurface.The mostimportantrolethe snow-
packplayedin controllingwinterCO2andCH4 fluxeswasin providinginsulationfor soilsfrom coldmidwinterair temperatures, which allowed microbial activity to continuethroughthe snow-covered season.Soil moistureandhydrologyappearedto be
theprincipalfactorscontrollingCO2andCH4 fluxesduringthe winterin LochVale. Spatialpatternsin CO2andCH4 emissions alongthewetlandtransectwereconsistent with the soil-moisture gradient across thewetland.Temporalchanges in CO2fluxfrom aerobicsoilsmayhavebeendrivenby gradualincreases in soil moisturepriorto snowmelt.Episodicreleaseof CH4 fromthe wetland wasprobably related tohydrologic changes in thewetland ratherthanan increasein microbialactivity. Winterfluxesof CO2 andCH4werefoundto accountfor a significant percentage of the
annualbudgets of thesegases.WinterCO2emissions in 19951996 accounted for 23, 12, and 8% of the grossannualCO2emis-
sionsfromdry,moist,andwet soils,respectively. Wintermicrobial activityaccounted for 12%of theannualCH4 consumption in dry soilsand58 and 12%of the annualCH4 emittedfrommoist and saturatedsoils,respectively.Owing to the inherenthigh porosities of snow,estimates of gasfluxfromsnowpacks arelikely to be sensitiveto the methodof measurement.Snowpackconcen-
trationgradients appearto providea moreaccurate technique for
MASTET AL.: WINTERFLUXESOF CO2 AND CH4 estimatinggas flux than staticchambermeasurements madeat the
snowsurface,particularlyin areaswhereseasonalsnowpacks reachmetersin depth.
619
Flanagan, P. W., and E L. Bunnell, Microflora activities and decomposition, in An ArcticEcosystem: The CoastalTundraat Barrow, Alaska,editedby J. Brown,et al., pp. 291-334,Van NostrandReinhold, New York, 1980.
Hardy,J. P.,R. E. Davis,andG. C. Winston,Evolutionof factorsaffecting Acknowledgments. This study was funded by the U.S. gastransmissivity of snowin the borealforest,in Biogeochemistry of GeologicalSurveyundertheWater,Energy,andBiogeochemical SeasonallySnow-Covered Catchments, editedby K. A. Tonnessen, M.
Budgets Program. Theauthors thankD. H. Campbellfor providing snow stratigraphy and density data. Previous versions of this manuscript were greatly improved by the comments of G. C. Winston, T. R. Moore, D. T. Chafin, H. H. Strickland, and two anonymousreviewers.
W. Williams, and M. Tranter,IAHS Publ. 228, 51-60, 1995. Healy, R. W., R. G. Striegl, T. F. Russell,G. L. Hutchinson,and G. P. Livingston, Numerical evaluationof static-chambermeasurementsof soil-atmosphere gasexchange,Soil Sci.Soc.Am. J., 60, 740-747, 1996. Hutchinson,G. L., andA. R. Mosier,Improvedsoil covermethodfor field measurementof nitrousoxide fluxes,Soil Sci. Soc.Am. J., 45, 311-316, 1984.
References
Albert,M. A. andJ.P. Hardy,Ventilation experiments in a seasonal snow cover,in Biogeochemistry of SeasonallySnow-CoveredCatchments, edited by K. A. Tonnessen,M. W. Williams, and M. Tranter, IAHS Publ., 228, 41-50, 1995.
Anthony,W. H., G. L. Hutchinson,and G. P. Livingston,Chamber measurement of soil-atmosphere gas exchange:Linear vs. diffusion-
Ishii, A. L., R. W. Healy, andR. G. Striegl,A numericalsolutionfor the diffusionequationin hydrogeologic systems,U.S. Geol. Surv., Water Resour.Invest.Rep., WR189-4027, 86, 1989. Jury,W. A., J. Letey, and T. Collins, Analysisof chambermethodsfor measuringnitrousoxideproductionin the field, Soil Sci. Soc.Am J., 46, 250-256, 1982.
Lins, H. F., Recentdirectionstakenin water,energy,andbiogeochemical budgetsresearch,Eos Trans.AGU, 75(33), 433,438-439, 1994.
Livingston,G. P., and G. L. Hutchinson,Enclosure-based measurementof tracegas exchange:Applicationsand sourcesof error,in Methodsin Ecology: Trace Gases,editedby P. Matson and R. Harriss, Blackwell Arthur,M. A., Vegetation, in Biogeochemistry of a SubalpineEcosystem, Sci., Cambridge,Mass., 1994. Ecol. Stud.,vol. 90, editedby J. Baron,pp. 76-92, Springer-Verlag, basedflux models,Soil Sci. Soc.Am. J., 59, 1308-1310, 1995.
New York, 1992.
Baron,J., Biogeochemistry of an Alpine Ecosystem, EcoL Stud.,vol. 90, Springer-Verlag, New York, 1992.
Bouwman,A. F., Global distributionof the major soils and land cover types,inSoilsandtheGreenhouse Effect,editedby A. FoBouwman, pp. 61-127, JohnWiley, New York, 1990. Brooks, P. D., S. K. Schmidt, R. A. Sommerfeld, and R. Musselman,
Massman, W., R. A. Sommerfield, A. R. Mosier, K. F. Zeller, T. J. Hehn,
andS. G. Rochelle,A modelinvestigation of turbulence-driven pressure pumpingeffecton therateof diffusionof CO2, N20, andCH4 through layeredsnowpacks, J. Geophys.Res.,102, 18,851-18,863,1997. Massman, W., R. A. Sommerfeld, K. Zeller, T. Hehn, L. Hudnell, and S.
Rochelle,CO2 fluxthrougha Wyomingseasonal snowpack: Diffusional andpressurepumpingeffects,in Biogeochemistry of SeasonallySnowCoveredCatchments,editedby K. A. Tonnessen,M. W. Williams, and
Distributionandabundance of microbialbiomassin RockyMountain M. Tranter, IAHS Publ., 228, 71-79, 1995. springsnowpacks, paperpresented at 61st WesternSnowConference, Mast, M. A., D. W. Clow, and R. G. Striegl, Flux of carbon dioxide and 1993 AnnualMeeting,QuebecCity,Que., 1993. Brooks, P. D., S. K. Schmidt, D. A. Walker, and M. W. Williams, The
Niwot Ridge snowfenceexperiment: Biogeochemical responses to changes in the seasonal snowpack, in Biogeochemistry of Seasonally Snow-CoveredCatchments,edited by K. A. Tonnessen,M. W. Williams, and M. Tranter,IAHS Publ., 228, 293-302, 1995.
Brooks,P.D., M. W. Williams,andS. K. Schmidt,Microbialactivityunder alpinesnowpacks,Niwot Ridge,Colorado,Biogeochemistry, 32, 93113, 1996.
Brooks,P. D., S. K Schmidt,andM. W. Williams,Winterproduction of CO2 and NO2 from alpine tundra: Environmentalcontrols and relationship to inter-system C andN fluxes,Oecologia,110, 403-411,
methanefrom a seasonalsnowpack,EOS Trans., AGU 75(44), Fall Meet. Suppl.,205, 1994. Mast, M. A., C. Kendall, D. W. Clow, D. H. Campbell, and J. Back, Determinationof hydrologicpathwaysin an alpine-subalpinebasin usingisotopicand chemicaltracers,in Biogeochemistry of Seasonally Snow-Covered Catchments,edited by K. A. Tonnessen,M. W. Williams, and M. Tranter, IAHS Publ., 228, 263-270, 1995.
Melloh, R. A., and P.M. Crill, Winter methanedynamicsbeneathice and in snowin a temperatepoorfen, Hydrol. Processes,9, 947-956, 1995. Melloh, R. A., and P.M. Crill, Winter methanedynamicsin a temperate peatland,GlobalBiogeochem. Cycles,10(2), 247-254, 1996. Millington, R. J., Gas diffusionin porousmedia, Science,130, 100-120, 1959.
1997.
Moosavi,S.C., and P.M. Crill, Controlson CH4 flux from an Alaskan borealwetland,Global Biogeochem.Cycles,10(2), 287-296, 1996. Campbell, D. H., D. W. Clow,G. P.Ingersoll, M. A. Mast,andNoE. Spahr, Processescontrolling the chemistry of two snowmelt-dominated Mosier,A. R., Chamberand isotopetechniques, in Exchangeof Trace streamsin the Rocky Mountains,WaterResour.Res., 31(11), 2811GasesbetweenTerrestrialEcosystems and the Atmosphere, editedby 2821, 1995. M. AndreaandD. Schimel,pp. 175-188,JohnWiley, New York, 1989.
Ciais,P.,P.P. Tans,M. Trolier,J. W. C. White,andR. J. Trancy,A large
Northern Hemisphere terrestrial CO2 sinkindicated by the!3C/12C ratioof atmospheric CO2,Science, 269, 1098-1102,1995. Conway,H., andJ. Abrahamson, Air permeability asa texturalindicatorof snow,J. Glaciol., 30(106), 328-333, 1984.
Coxson,D. S., and D. Parkinson, Winter respiratoryactivityin aspen woodlandforestfloor litter and soils,Soil Biol. Biochem.,19, 49-59, 1987.
Dise, N. B., Winter fluxes of methanefrom Minnesotapeatlands, Biogeochemistry, 17, 71-83, 1992.
Mosier, A. R., L. K. Klemedtsson, R. A. Sommerfeld, and R. C.
meadow,GlobalBiogeochem. Cycles,7(4), 771-784, 1993.
Oechel,W. C., (]. Vourlitis, andS. J. Hastings, Coldseason CO2 emission fromarcticsoils,GlobalBiogeochem. Cycles,11(2), 163-172, 1997. Solomon,D. K., andT. E. Ceding,The annualcarbondioxidecyclein a montanesoil:Observations, modeling,andimplicationsfor weathering, WaterResour.Res., 23(12), 2257-2265, 1987.
Sommerfeld,R. A., and E. R. LaChapelle,The classificationof snow metamorphism,J. Glaciol., 9(55), 3-17, 1979.
Sommerfeld, R. A., A. R. Mosier,andR. C. Musselman, CO2, CH4,and
620
MAST ET AL.: WINTER FLUXESOF CO2 AND CH4
N20 flux througha Wyomingsnowpackand implicationsfor global budgets,Nature, 361, 140-142, 1993. Sommerfeld, R. A., W. J. Massman, R. C. Musselman, and A. R. Mosier,
Diffusionalflux of CO2 throughsnow:Spatialandtemporalvariability amongalpine-subalpinesites,Global Biogeochem.Cycles,10(3), 473482, 1996.
Striegl, R. G., Diffusional limits to the consumptionof atmospheric methaneby soils,Chemosphere, 26, 715-720, 1993. Striegl, R. G., K. P. Wickland, M. A. Mast, and D. W. Clow, Carbon dioxideandmethaneexchangeat a subalpinewetland,Rocky Mountain National Park, Colorado, USA, paper presentedat INTERCOL V
Wickland, K. P., R. G. Striegl, S. K. Schmidt, and M. A. Mast, Methane flux in subalpinewetlandand unsaturatedsoilsin the southernRocky
Mountains,GlobalBiogeochern. Cycles,in press,1998 Winston,G. C., B. B. Stephens,E. T. Sundquist,J.P. Hardy, and R. Eo Davis, Seasonalvariabilityin CO2 transportthroughsnowin a boreal forest, in Biogeochernistry of SeasonallySnow-CoveredCatchments, edited by K. A. Tonnessen,M. W. Williams, and M. Tranter, IAHS Publ. 228, 61-70, 1995.
Zimov, S. A., I. P. Semiletov, S. P. Daviodov, Y. V. Voropaev, S. E Prosyannikov, C. S. Wong,andY. H. Chan,WintertimeCO2 emission from soils of northeasternSiberia, Arctic, 46(3), 197-204, 1993.
International Wetlands Conference, Perth, Australia, 1996.
Whalen, S.C., andW. S. Reeburgh,A methaneflux time seriesfor tundra environments,Global Biogeochem.Cycles,2, 399-409, 1988. Wickland, K. P., Controlson spatialand temporalvariabilityof methane exchangein subalpinesoils, Master's thesis, 53 pp., Univ. of Colo, Boulder, 1997.
Wickland, K. P., and R. G. Striegl, Measurementsof soil carbondioxide and methaneconcentrations and fluxes,and soil propertiesat four ages of jack pine forestin the southernstudyarea of the BorealEcosystem AtmosphereStudy,Canada, 1993-95, U.S. Geol. Surv. Open File Rep. 97-49, 73 pp., 1997.
D. W. Clow, M. A. Mast, R. G. Striegl, K. P. Wickland, Water Resources Division, U.S. Geological Survey, Denver Federal Center, Denver, CO 80225. (e-mail:
[email protected];
[email protected];
[email protected];
[email protected])
Wickland, K. P., M. A. Mast, R. G. Striegl, and D. W. Clow, Seasonal
comparisonof CO2 and CH4 fluxesfrom a subalpinewetland,EOS Trans.AGU, 76(46), Fall Meet. Suppl.,F231, 1996.
(ReceivedNovember 14, 1997; revisedJune 11, 1998; acceptedJuly 10, 1998.)
