The Regents of the University of Colorado, a body corporate, contracting on behalf of the
University of Colorado at Boulder for the benefit of INSTAAR
Influence of Vegetation, Temperature, and Water Content on Soil Carbon Distribution and
Mineralization in Four High Arctic Soils
Author(s): Bo Elberling, Bjarne H. Jakobsen, Peter Berg, Jens Søndergaard, Charlotte
Sigsgaard
Reviewed work(s):
Source: Arctic, Antarctic, and Alpine Research, Vol. 36, No. 4 (Nov., 2004), pp. 528-538
Published by: INSTAAR, University of Colorado
Stable URL: http://www.jstor.org/stable/1552307 .
Accessed: 02/11/2011 08:55
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
INSTAAR, University of Colorado and The Regents of the University of Colorado, a body corporate,
contracting on behalf of the University of Colorado at Boulder for the benefit of INSTAAR are collaborating
with JSTOR to digitize, preserve and extend access to Arctic, Antarctic, and Alpine Research.
http://www.jstor.org
Arctic,Antarctic,andAlpineResearch,Vol.36, No. 4, 2004, pp. 528-538
of Vegetation,
Influence
andWaterContent
onSoilCarbon
Temperature,
Distribution
andMineralization
in FourHighArctic
Soils
Bo Elberling*:
BjarneH. Jakobsen*
Peter Bergt
Jens S0ndergaard*and
CharlotteSigsgaard*
of Geography,
of
*Institute
University
0sterVoldgade10,DK-1350
Copenhagen,
K., Denmark
Copenhagen
of Environmental
Sciences,
tDepartment
of Virginia,291 McCormick
University
Road,P.O.Box 400123,Charlottesville,
VA 22904-4123,U.S.A.
author.
t Corresponding
Abstract
Soil organic matterdistributions,reservoirs,and mineralizationrates in tundrasoils are
importantfactorsfor understandingbiogeochemicalcarboncycling. This studyfocuses on
spatial trends and environmentalcontrols of soil carbondistributionand microbialsoil
respirationin 4 tundravegetationcommunitiesin an arcticvalley in NE-Greenland(74?N),
includingDryas and Cassiopeheaths,Salix snow bed, and fen vegetation.Measuredtotal
soil organiccarbonin the upper50 cm averaged(+SD) 11.0 ? 1.5 kg C m-2 with spatial
variationsstronglyaffectedby vegetation,hydrology,andburiedorganiclayers.Observed
soil CO2 concentrationsand effluxes were simulatedwith a steady-statediffusion model
using laboratorymeasuredCO2 productionsas input. SimulatedCO2 profiles and CO2
effluxes (up to 3 tmol CO2 m-2 s-1) agreed with field observationsand revealed the
importanceof both vegetation-and depth-specificCO2productionand CO2diffusionfor
understandingthe spatialvariationin near-surfacesoil CO2 gas dynamics.These results
confirmthatmoleculardiffusiondominatesgas transportin the studiedsoils; but also that
the complexityof CO2production/transport
coupledto soil heterogeneity(in particularthe
litterlayer)complicatesthe applicationof soil-diffusionmodelsto estimateseasonaltrends
of soil gas effluxes.
Strikingchangesin arcticplantcommunitiesandconcurrentsoil
CO2 effluxes have been observed over small distances due to
Soil carbonandnutrientcyclingwithintheterrestrial
arcticregions
microtopography,which seemed related primarilyto hydrology
is an important
et al.,
et al., 1996). Consequently,changesin plantcommunity
(Oberbauer
topic (e.g., CoyneandKelley, 1974;Oberbauer
1996;Burkinset al., 2001) as theselandscapeshold about14%of the
structure,composition,and distributionin the landscapewill be an
world'sorganiccarbon(Postet al., 1982)andmaybe someof themost
additionalfeedbackmechanismfor long-termclimatechanges(Zimov
sensitiveecosystemswith respectto climaticchanges(Oechelet al.,
et al., 1996). This feedbackis not limited to the plants, but also
of soil organiccarbon involves the complex interactionsbetweenplants,the soil organic
1993, 1998;Maxwell,1997).Thedecomposition
resultsin a releaseof severalgreenhousegassesto the atmosphereof
matterpools, andthe soil organicmattermineralization,
which,again,
which the release of carbondioxide (CO2)is consideredthe most
is closely related to the availabilityof nutrientsfor plants and
important(Oechelet al., 1993). CO2is being producedprimarilyby
microorganismscompeting in nutrient-limitedarctic ecosystems
near-surface
et al., 1991).
(Oberbauer
respirationin living roots (Billingset al., 1977) and by
soil microorganisms
(Buchmann,2000). Consequently,
heterotrophic
Reporteddata for the arcticshow that less attentionhas been
10 to 100
gas-filledporesin soils typicallyhave CO2concentrations
given to the well-drainedand drieruplandtundraareas where the
timeshigherthantheatmosphere
(Welleset al., 2001),resultingin a net
vegetationcoveris limitedanddominatedby heathvegetation(Welker
flux
from
the
soil
to
the
The
of
near-surface
et
al., 1999;2000).Thisis despitethefactsthatthedriertundrais more
CO2
atmosphere. transport
such
as
has
been
considered
as
a
result
of
CO2
gases
long
mainly
widespreadtowardthe mostnortherndry arcticregions,e.g., mostof
diffusion(e.g.,Lundegardh,
northernGreenlandnorthof 74?N (Bay, 1992), and thatthese soils
1927;De JongandSchappert,
1972)driven
show the greatestthawdepths(Oberbauer
et al., 1996).
by concentration
gradientsandlimitedby thedecreasein continuousairfilledsoilporeswithincreasingwatercontent.Thedecomposition
of soil
For these reasons,the assessmentof presentcarboncycling and
organic matter is sensitive to several site-specificenvironmental long-termtrendsof net carbonbalancein arcticregionsrequiresnot
conditionsincludingthe quality, abundance,and redistributionof
of specific controlson below-ground
only improvedunderstanding
carbonsubstrates(VanCleve, 1974;Nadelhofferet al., 1991, Fahne- CO2dynamicsat appropriate
spatialandtemporalscales(Groganand
stocket al., 2000) andclimaticfactors,in particularsoil temperature Chapin,1999), but also improvedunderstanding
of the interacting
(FangandMoncrieff,2001),watercontent(HowardandHoward,1993;
processes at a landscapescale including soil types, soil water
Kirschbaum,1995), and snow (Welkeret al., 2000). Availabilityof
distribution,vegetationtypes, soil organicmatterquality,and soil
present,as well as faunalabundance gas transport.
oxygen,typesof microorganisms
andactivityareadditionalfactorsinfluencingsoil microbialrespiration
Whileit is still debatableto whatextentthe arcticregionwill act
et al., 1998).
as a sourceor a sinkfor atmospheric
carbonin the future,the net loss
processes(Lomander
Thespatialandtemporalvariabilityof suchconditionsin the field
or gainin carbonis expectedto resultin significantchangesin thetotal
results in substantialvariationsin rates of soil organic carbon ecosystem pools. However, future estimatesof changes in pools
on totalanddepthdistribution
of existing
mineralization,soil CO2 production,and effluxes between tundra requiredetailedinformation
et al., 1991, 1992;
soil carbon.So far,very littlehas beenreportedon plantcommunityvegetationtypes(e.g., Evanset al., 1989;Oberbauer
controlledvariationsin the amountof soil organiccarbonandon the
Hobbie,1996;GroganandChapin,1999).
Introduction
528 / ARCTIC, ANTARCTIC, AND ALPINE RESEARCH
? 2004 Regents of the University of Colorado
1523-0430/04 $7.00
TABLE 1
Soil physicaland biogeochemicalpropertiesof majorsoillvegetationtypesin Zackenberg,Greenland(+ one standarddeviation)
Vegetation type
Dry Dryas heath
Moist Cassiope heath
Salix snow bed
Wet Eriophorumfen
% Watersaturationrange (0-5 cm)
Area below 200 m a. s. 1. (km2)
'Approximate% of total area
Total soil C (0-20 cm) (kg C m-2)
Total soil C (0-50 cm) (kg C m-2)
Vegetation C:N
Vegetation-C(g C m-2)
'Plant cover (%)
Litter-C(g C m-2)
Root-C (g C m-2)
Belowgrounddebris-C (g C m-2)
Above-below ground biomass ratio
2Soil Temperature(5 cm) (?C)
2Soil CO2 efflux (gmol m-2 s-1)
Maximum thaw in August (cm)
40-60
4.36 ? 0.5
26
4.1 + 0.4
6.1 + 1.1
35
327 + 73
65
13 + 7
155 + 25
265 + 150
0.81
9.13 + 3.8
0.836 + 0.2
80
60-80
4.09 + 0.6
25
6.3 + 3.0
8.5 + 2.6
43
360 + 73
65
21 + 6
123 + 10
423 + 121
0.70
5.49 + 1.8
0.52 + 0.1
65
65-90
3.17 + 0.3
19
10.5 +1 .8
21.2 + 4.1
28
80 + 39
80
36 + 10
nd3
877 ? 156
nd3
7.59 + 3.8
1.22 ? 0.3
45
90-100
4.87 + 0.6
30
5.6 + 3.2
11.3 + 3.8
31
1134 + 19
88
nd3
nd3
nd3
nd3
7.59 + 3.8
2.48 + 0.5
40
Dry Dryas heath
A
C
B/C
Horizon
Depth (cm)
4pH-H20
4Bulk density (g cm-3)
4C (%)
4N (%)
4C/N
4p (total) (mg kg-1)
4P (uorg) (mg kg 1)
0-2
2-35
>35
5.9
6.4 7.0-7.2
0.8
1.5
1.6
3.9
0.6
0.5
0.2
<0.1
0.1
nd3
16.3
11.5
341
357
300
72
50
99
Moist Cassiope heath
A
5-17
0-5
5.7
6.0
0.8
1.3
6.7
2.1
0.4
0.1
16.2
17.3
510
299
188
156
Salix snow bed
Ab
C
17-22
5.5
1.5
4.9
0.2
20.4
466
109
>22
6.6-7.1
1.6
0.4
<0.1
nd3
250
240
B/C
A
B1
B2
0-2
2-6
6-20
5.1
5.2
5.3
0.6
1.5
1.2
15.4
3.7
3.2
0.9
0.2
0.2
17.3
17.6
16.8
1240
612
751
803
202
334
Ab
Wet Eriophorumfen
B/C
A
20-22 >22
0-10
5.0
5.9
5.4
0.8
1.6
0.6
14.2
1.8
3.8
0.9
0.2
0.4
15.6
11.3
10.9
1220
460
680
778
93
371
B1
B2
Ab
B/C
10-50
6.7
0.8
1.8
0.2
9.0
660
587
50-80
7.6
1.1
1.0
0.1
10.0
nd3
nd3
80-87
6.0
1.3
3.3
<0.1
nd3
nd3
nd3
>87
7.2
1.3
1.4
<0.1
nd3
nd3
nd3
Modified from Bay (1998) and Soegaardet al. (2000).
2 Average of hourly readingsin July August 2001.
3 Not determined.
4 Values
given for each horizon representthe average of samples taken with 5 cm incrementswithin each horizon (3 replicates).
potentials for future shifts in the distribution of plant communities as
a result of climate changes.
The present study was designed to (1) evaluate spatial trends and
environmental controls of soil carbon reservoirs and mineralization
rates during a growing season of 4 characteristic vegetation types
found within an arctic valley in NE-Greenland, where all plant
communities were subject to similar above-ground climatic conditions;
(2) quantify the importance of soil horizons in terms of storing organic
carbon and releasing CO2 produced by organic matter mineralization;
and (3) discuss the effects of climatic changes on vegetation patterns
and corresponding soil CO2 dynamics.
andMethodology
Material
STUDY SITES
The study sites are situated in the Zackenberg Valley near the
Zackenberg Research Station in NE-Greenland (74?28'N, 20?34'W).
The sites were selected to represent major plant communities which
again reflected the hydrological regimes present in the valley.
Vegetation types were identified in the field by topographic position,
vegetation, and soils. This classification was confirmed by an aerial
photo that previously had been used to quantify the area distribution
(Table 1) of all major vegetation zones (Bay, 1998; S0gaard et al.,
2000). The abrasionplateaus representthe most exposed and driest areas
in the valley. Here, the snow cover is thin or absent throughout winter,
resulting in early and fast development of the active layer during spring.
The sparse vegetation cover is dominated by Dryas sp. (including D.
integrifolia). Flat low-lying areas receiving the average amount of snow
(0.5-1 m yr-) were typically free of snow by late June and remained
moist but aerated during most of the growing seasons (1998-2001).
Cassiope tetragona dominates these areas, although other heath species
such as Vaccinium uliginosum are found in patches. The variety of heath
vegetation covers more than 30% of the valley below 200 m above sea
level (asl.) and is the most dominant vegetation type found in the valley.
Areas receiving water from persistent snow drifts (snow beds) remain
moist or wet throughoutthe growing season and are dominated by Salix
arctica. Fen vegetation dominated by graminoids is found in undrained
landscape depressions and other low lying areas receiving substantial
amounts of water from snow melt throughout the summer. Arctagrostis
latifolia and Eriophorum triste are key species found in less-wet parts,
whereas Eriophorum scheuchzeri and Dupontia pilosantha dominate
the wetter parts. The four studied vegetation types, referred to as the
Dryas, Cassiope, Salix, and Eriophorum sites, were selected within an
area of 3 km2. These four types represent a simplification of vegetation
types previously reported for the Zackenberg Valley (Bay, 1998),
including also grassland, saltmarsh, vaccinium, abrasion plateau, and
fell-fields. In this study, grassland areas were split into Eriophorumand
Salix sites, depending on the dominance of Salix sp., while abrasion
plateaus dominated by Dryas sp. were included in the Dryas group.
Small areas consisting of salt marsh, fell-fields, lakes, and rivers
representing less than 10% were not included.
SOILS
The Zackenberg Valley is a generally flat valley dominated by
noncalcareous sandy fluvial sediments. The present soil development is
weak and has been classified as Typic Psammoturbels (Elberling and
ELBERLING ET AL. / 529
Jakobsen,2000).A relictA-horizon(buriedold surfacelayer,Ab)of 15 cm thicknesswas foundin most soil profilesat depthsbetween15
and30 cm andwas in placesassociatedwitha well-developedPodzol.
Previous datings of these buried horizons have shown that they
representa soil developmentduringthe HoloceneClimateOptimum
et al., 2002).
startingat least 5000 yr ago (Christiansen
CLIMATE
A climatestationlocatedwithinthe studyareawas establishedin
1995.Thestationis operatedby theDanishPolarCentre(DPC)as part
of the ZackenbergZERO(ZackenbergEcologicalResearchOperaBasedon datafromthatstation,themean
tion),GeoBasisProgramme.
is about-10?C, andthe annualprecipitation
annualtemperature
varies
from 150 to 200 mm (Caningand Rasch,2000). Minimumair tem2.5 cm
peraturesduringwinterare close to -40?C, andtemperatures
below the soil surfacearebelow-18?C for about4 monthsper year.
at 2.5 cm depthswere
Duringthe period 1997-2001, temperatures
for
an
of
124
average
days per year (on average4.3?C).
positive
air temperatures
Climatedataused in this studyincludeprecipitation,
(at 2 m above ground),and snow thickness(using Campbellmodel
SR 50 sonic distancemeasuringprobes).
and automaticlogging instrumentsrespondingto changes in the
apparentdielectricalconstant(ThetaProbe, Soil Moisturesensors,
ML2x, Delta-T Devices Ltd., Cambridge,UK). Water content
measurements
were logged at seven depthsat the Cassiopesite on
an hourlybasis throughoutthe summerof 2001. Additionalmanual
watercontents(0-5 cm)
readings(at least3 replicates)of near-surface
weremadeat the sametime as the soil CO2effluxmeasurements.
Air
andgroundtemperatures
at 2 cm depthwereloggedat all sites on an
hourly basis using TinyTag loggers (Gemini Data Loggers Ltd.,
Chichester,UK). Manualtemperature
readingswere taken by preinstalledthermistorsat a minimumof five depths. The depth of
thawingwas determinedmanuallythroughoutthe growing season
using a graduatedstainlesssteel rod insertedverticallyinto the soil
untilfrozensoil was encountered.
CHEMICAL
ANALYSES
Intactdepth-specific
andvolume-specific(100 cm3)soil samples
(3-5 replicates)were collectedat 5 cm depthintervalswithineach
horizonin lateJulyor Augustandstoredin the darkat maximum4?C
untilanalyzedin thelaboratory
(a totalof 6 replicateprofiles).Samples
storedformorethana few dayswerestoredat temperatures
as close to
0?Cas possible.Sampleswerecollectedfrom 1998 to 2001.
Soil pH was measuredin distilledwaterat a soil-solutionratioof
SOILC02 EFFLUXES
1:2.5. Total inorganicC (TIC) and total organic C (TOC) were
measure- measuredusing a Total OrganicCarbonAnalyzer(DohrmannDCEcosystemCO2efflux (microbialand rootrespiration)
of decarbonated
mentsweremadeusinga LiCorinfrared(nondispersive)
gas analyzer 190). Measurements
samplesrevealedthatcarbonate
minerals
were
absent
from
all
Soil
Flux
6400-09/6262
Chamber,
LiCor,
Lincoln,
CO2
(LiCor
samples.Totalnitrogenwas measured
U.S.A.). The CO2 analyzerwas attachedto a portablechamber, by the Kjeldahlmethod(Bremnerand Mulvaney,1982);Al and Fe
were determinedby atomic absorptionafter acid-oxalateextraction
functioningas a darkandclosedsoil-fluxchamberwhenplacedon top
of pre-installed
opencollards(9 cm long and 10 cm insidediameter). (McKeagueandDay, 1966);exchangeablebases(Ca,Mg, Na, andK)
Themeasuringprocedure
is describedin Elberling(2003)andmodified were analyzedby AAS (atomicabsorptionspectrophotometry)
after
from Welles et al. (2001) and Buchmann(2000). The final flux
extractionwithNaOOCCH3
andNH400CCH3(Jakobsen,1991),and
reportedis the flux observedat ambientchamberCO2concentrations exchangeableH and Al were determinedby titrationafterextraction
with 1.OMKCI (Sims, 1996). InorganicP and total P (after 1 h at
andis the averageflux basedon at least 10 replicatecollardsper site
were madeduring 550?C)weredetermined
(in most cases 17 replicates).Flux measurements
by colorimetrywithascorbicacidafter6.0M
2 weeksin August1999 as well as throughout
sulfuricacidextractions(Kuo,1996).Air-filledporositywascalculated
JulyandAugust2001.
were madeat least every secondday at the Cassiope fromthe grain-sizedistribution(determinedby sieving and hydromMeasurements
site, weekly at the Dryas and Salix sites, and three times at the
eter),totalvolume,weight,porosity,andsolid bulkdensity.
weremadeat the Dryas
site. Additionally,measurements
Above-groundbiomasswas quantifiedby harvestingall vegetaEriophorum
andCassiopesites withoutvegetation(morethan20 cm to the nearest tionin 0.25 m2plots(5 replicates)Littermaterialwascollectedfromthe
sameplots.Below-ground
biomasswasquantified
of
by wet separation
plant). Such barrenareas could not be found at the Salix and
In
measurements
were
out
on
sites
dead
and
live
roots
of
color
and
and
sites.
carried
addition,
(visual
inspection
Eriophorum
appearance)
where the litter layer (0-2 cm) and the A-horizon(5-7 cm) were
organicdebrisfromsoil samples.Samplesweresubsequently
ground,
madeon the first2 daysafterthe removalof
removed.Measurements
homogenized,andanalyzedfor soil organicC andN.
effectsof disturbance.
materialwerenot includeddue to short-term
Soil gas was withdrawnfrompre-installed
passivegas samplers
EXPERIMENTS
C02 PRODUCTION
of 02 andCO2weremeasuredonholding250 mL,andconcentrations
fromsoil
siteusinga portabledualwavelengthprecisionCO2analyzer(GasData
CO2productionratesweremeasuredin the laboratory
PCO2,Gas DataLtd.,England).Soil gas was extractedat leastevery
samplesthat had been storedin polyethylenebags at 0?C to 4?C
for a maximumof 1 week. Prior to measurements,soil samples
secondweekin thesummersof 2000 and2001fromdepthsof 5, 10, 15,
and
to
60
cm.
The
and
concentrations
were
measured
were carefullysplit to removeroots and stones.Afteran initialpre20,25,40,
02
CO2
an accuracyof 500 and 20 1tLL-1, respectively.To evaluatespatial incubationof 48 h, weighedsoil samples(equivalentto 2-3 g drysoil)
weretransferred
to 12 mL Venojecttubes,left for 2 h pre-incubation,
variationsin soil gas composition,soil gas was also extractedfrom
in depth-specific
variousdepthsby connectingthe gas analyzerto a soil probe(Drager andout-gassedwith airfreeof CO2.Soil respiration
soil sampleswas subsequently
measuredby monitoringthelinear(r2>
Sicherheits-tecnik
GmbH,Germany),whichwas pushedverticallyinto
were
the soil at intervalsof 5 cm. Thesedetailedprofilemeasurements
0.95) increaseof headspaceCO2concentrations
duringa 6-8 h period
of the temperature
and soil
followed immediatelyby measurements
by gas chromatography,
usingeithera HewlettPackard(HP-6890)or
a MicroLab(ML-GC8212)equippedwith a PorapackQ-columnkept
watercontentsatthesedepths.On 30 Julyand3 August1999,a survey
andwatercontentin
at 30?Canda TCDdetectoroperatedat 200?Canda flowrateof 25 mL
of the spatialvariationin soil CO2concentrations
He min-'. These measurements
are referredto as BSR (basal soil
10 and 25 cm depthswas performedin a 4 x 4 m grid (n = 16) at
werenotpossibleat
measurements
a Cassiopeanda Dryassite.Soil gasmeasurements
below.
respiration)
all at Eriophorum
sitesdueto highwatercontent.
Additionalexperiments
madewithsamplesaftervariousdegreesof
The volumetricsoil watercontentwas monitoredusinghandheld reworkingindicatedthatsamplesplittinghad only little influenceon
AND
530 / ARCTIC,ANTARCTIC,
ALPINE RESEARCH
rates
CO2productionrates,whereasgentlemixingincreasedshort-term
Aon
from
material
rate
measurements
by up to 20%.CO2production
horizons from the four vegetation types were made at constant
( 1?C)at 0 ?C,7?C,21?C,and26?C.Additionalsamples
temperatures
were partlyair-driedand subsequentlyrewetted(about5% and 10%
relativeto in-situmoisturecontent)beforebeing incubatedat 26?Cto
test whetherthe availabilityof waterinfluencedthe measuredrates.
Substratelimitationwas tested by gently mixing soil materialwith
a powdermixturecontainingglucose (1.2 mg glucose-Cg-1 soil) and
talc (AndersonandDomsch,1978).Resultsarereferredto as substrateinduced soil respiration(SIR-measurements)and compared with
were
samplesmixedwithtalc only. All CO2productionmeasurements
from
soil sampleswereallprepared
madewith3 replicates.Manipulated
soil samplesairdriedfor a few hours,followedby sieving(2 mm) and
subsequentlyrewettedto specificwatercontentsandthoroughlymixed
to ensure homogenous samples. Control incubationswith empty
VenojecttubesshowedthatCO2diffusionintothetubeswasnegligible.
Repeated incubation experimentsrevealed that the results were
independentof the experimentaltime.
MODELLING
Assuming steady-stateconditionsand neglectingthe effect of
watermovement,the relationshipbetweenone-dimensionaldiffusion
andproductionof CO2in soils can be describedby Fick's secondlaw:
(De
+)
R= 0
(1)
x is the depth,De is the effective
whereC is the CO2concentration,
diffusioncoefficient,andR is the productionrateper unit volumeof
soil. Depthvariationsof De andR were observedin the studiedsoils,
whichpreventedany analyticalsolutionsto Equation1. Consequently,
Equation1 was solvednumericallyusinga controlvolumeapproachas
describedby Berget al. (1998). In short,the soil columnsweredivided
into N horizontallayers, so-calledcontrolvolumes,each with a grid
point located in the center. The numericalsolution allowed us to
for eachgridpointandalso fluxesacross
calculateCO2concentrations
each boundarybetween the control volumes based on known or
estimatedvalues of De and R. A value of N = 100 was used, which
precisenumericalsolutionsto Equation1, whichhavebeen
guaranteed
successfullytestedagainstanalyticalsolutions(Berget al., 1998).The
effectivediffusioncoefficient(De) in Equation1 was estimatedas
De
Daa10/3
2
neutral (pH range 5.0 to 7.2) with pH generally increasingwith
depth.Most sites were more or less cryoturbatedwhich were noted
as discontinuoussoil horizons.The maximumactive layerdepthwas
observedby the end of August(1999-2001) and variedfrom 40 cm
at Eriophorumsites to 80 cm at Dryas sites. The highestproportion
of fine materialand less-sortedsedimentswas found at Salix sites
(25% finer that 2 im; d5o= 15 gm); largergrain sizes and greater
degree of sortingwas found in Eriophorum(20% finer than 2 ,um;
d5o= 75 jlm), Cassiope(10% finerthan 2 jm; d50= 150 pm), and
Dryas (5% finer than 2 pm; d5o = 360 ,um). The grain size
distribution,drainageconditions,and the distributionof snow drifts
determineda variationin near-surfacewater saturationfrom about
40% at Dryas sites to 100%at Eriophorumsites. Litterlayerswere
notedto be very irregularlydistributedat Cassiopeand Salix sites (in
the rangeof 0-2 cm, on average0.8 cm, n = 17) and almostabsent
at Dryas and Eriophorumsites. However, layers of peat were
sites. These layerscontainsubstantialmineral
detectedat Eriophorum
materialand are consequentlyconsideredto be part of the mineral
soil profile.
Soil profiles had maximumconcentrationof carbon (C) and
withdepth
nitrogen(N) nearthe surfaceanddecreasingconcentrations
of
buried
1
and
with
the
2)
(Table 1, Figs.
organic-rich
exception
layers, which were found in most profiles.These layers occurredat
differentdepthsandrevealedvaryingthicknessesandcarboncontents
both between replicatesites at each vegetationtype (Fig. 1B) and
betweenvegetationtypes.The greatestdifferencein buriedlayerswas
observed between Cassiope and Eriophorumsites (Fig. 2). At
Eriophorumsites the buriedorganiclayer was observedwithin the
permafrost.Single plant remains (birch leaves) were found in all
buriedorganiclayers.
Soil bulkdensitiesincreasedwith depthfromabout0.8 g cm-3 to
values near 1.6 g cm-3. Variationsin soil bulk densitieswithineach
vegetationtype were on the same orderof magnitudeas observed
betweenvegetationtypes, except for the consistentlylower densities
observedat the Eriophorumsite.
For each soil layer,the total amountof soil organiccarbonwas
of the elementby
calculatedby multiplyingthe averageconcentration
the averagesoil densityandthicknessof the layer.The totalamountof
soil organicmatterwithin the upper20 cm, 50 cm (Table 1), and
100 cm was obtainedby addingthe sumof theselayersandis referred
to as thesoil carbonreservoirto a certaindepth.Onestandarddeviation
for bulkdensity,percentsoil organiccarbon,andcarbonreservoirsare
(2)
whereDa is the diffusivityin free air [equalto 0.139 X (T/273)2cm2
s-1; where T is the temperature(?K) reportedby De Jong and
Schappert(1972)], a is the air-filledporosity,and ) is the porosity.
Equation2 was originallysuggestedby Millington(1959) for relating
the effectivediffusionin soil to the fractionof "continuous"air-filled
porevolume.The equationwas latershownto be suitablefor a variety
of sandy soils (Jin and Juri, 1996). Values of R were derivedfrom
measuredCO2 productionsrates and used with calculatedeffective
diffusioncoefficientsas inputto the model afterbeing temperaturecorrected.The model was calibratedusing observedsoil CO2efflux
and subsequentlyused to computethe correspondingCO2concentration profilefor model verification.
Bulk Density (g cm-3)
1.0
1.5
0.0
0
0.5
% Soil organic carbon
2.0
0
2
4
6
8
10
A10
10
\
E 20" 30-
Average
O H| =6.4 kg C m-2
O H,=8.3kgCm-2
40-
A
H = 7.0kg Cm-2
0
H,= 12.3kg Cm-2
50
Results
SOIL CHARACTERISTICSAND RESERVOIRS
of replicateprofilesfrom the four vegetation
Soil characteristics
types are summarizedin Table 1. All sites were slightly acidic to
FIGURE 1. Profiles (replicates) of soil bulk density and soil organic
content at four Cassiope sites showing the maximum spatial variation
observed in Zackenberg within a single vegetation type. One standard
deviation (?) of 3 replicates per depth is shown as horizontal lines. The
total amount of soil organic carbon is calculated to a depth of 50 cm.
ELBERLING ET AL. / 531
givenin Table1, indicatingthatalthoughlargevariationsareobserved
betweenprofilesof eachvegetationtype,variationsbetweendifferent
vegetationswere larger.Maximumvariationswithina single vegetation type wereobservedfor Cassiopesites (Fig. 1), consistentwithan
of the buriedAb-horizon.
irregulardistribution
The depth used for these stock estimatesis critical for both
absolutestockvaluesreportedandforcomparisons
betweenvegetation
sites. It can be seen (Fig. 2B) thatusing50 cm as the depthfor stock
estimatesincludesonly 55%of the soil organiccarbonfoundin the
sites, whereas92%is includedat Cassiope
upper1 m at Eriophorum
sites. In contrast,the 50 cm depthstockvaluerepresents100%of the
carbonfoundwithinthe activelayerat Eriophorum
sites (maximum
thawdepthabout40 cm) but 95%at Cassiopesites (maximumthaw
depthabout65 cm). If buriedsurfacelayersare foundnearthe total
depthfor estimates,the standarddeviationsbecome relativelylarge
(e.g., Cassiopesites, Fig. 1).
The resultingC-reservoirs
estimatedto depthsof 20 and 50 cm
(Table1) indicatethatDryashad the lowest carbonstock withinthe
upper50 cm (6.1 kg soil-Cm-2),followedby Cassiope(8.5 kg soil-C
m-2), Eriophorum(11.3 kg soil-C m-2), and Salix (21.2 kg soil-C
m-2); the last contains 2-3 times as much as Dryas sites. The
differencesin C stock were only slightlyaffected
vegetation-specific
(less than 10%)when calculatedwithouttakingburiedAb-horizons
into account.Only the reservoirestimatefor 20 cm at Cassiopesites
wasreducedup to 25%.Takingtheareadistribution
of eachvegetation
type in the valley into account(see Table 1), the average(?SD)
amountof soil organicC to a depthof 50 cm is 11.0 ? 2.2 kg soil-C
m-2 or 110 tonnesoil-Cha1, of which37%was foundat Salix sites,
30% at Eriophorum
sites, 19%at Cassiopesites, and 14%at Dryas
sites. Comparedto other fractions of ecosystem carbon pools
(vegetation,litter,debris,and roots),the soil organicpools represent
morethan94%of the totalecosystemorganiccarbon(Table1).
8
A
I
__.,.
20 -
20
40
60
I()
8)
I
I-I
_'-
40-p
60-,
80 -
,---.-----'
-
heath
Cassiope
----- Eriophorum
fen
:
100
FIGURE 2. Concentrations and distribution of soil organic carbon
in Cassiope and Eriophorum sites; given as (A) percent C at various
depths and (B) accumulated organic C to a certain depth in percent of
the total amount of organic carbon within the upper 1 m.
qualityto decomposition,includingVan Cleve (1974), who showed
that C:N ratios were negatively correlatedwith organic matter
(measuredas percentweightloss over2 yr) fora range
decomposition
of circumpolar
tundrasites. C:Nratiosin A-horizonmaterialcollected
at Dryas,Cassiope,and Salix sites (Table 1) were similar(16.2 to
17.3),withthe lowestvaluebeingobservedat Cassiopesites (Fig. 4).
the C:Nratiosalonecannotexplaintheloweractivityof
Consequently,
carbonsourcesat Cassiopesites. In contrast,C:N ratiosof vegetation
andlitterat Cassiopesites werethe highestdetectedof all vegetations
types investigated(Fig. 4). The control of plant species on litter
decompositionhas previouslybeen reportedfrom Alaskantundra
(Hobbie, 1996), who showed that decompositionrates were more
relatedto carbonqualitythanto nitrogenconcentrations.
The temperaturedependencyof microbialsoil respirationat
Cassiope sites in Zackenberghas been investigatedpreviously
(ElberlingandBrandt,2002).In thesestudies,an exponentialincrease
in soil respirationwas observedwith increasingtemperature
andwas
CONTROLSON SOIL C02 PRODUCTION OBSERVED
532 / ARCTIC,ANTARTInc,AND ALPINERESEARCH
0
0
IN THELABORATORY
The amountsof CO2 being releasedfrom depth-specificsoil
samplesper unit weight revealedthat the releasefrom the two Ahorizonsat Cassiopesites (A andAb)was not significantlydifferent(p
< 0.05) andthattheseA-horizonsproducedabout3-5 timesas much
CO2as underlyingB- and C-horizons.The same was noted for the
othervegetationtypes.Calculating
CO2production
perdryweightsoilC at optimumconditionsprovideda measureof therelativedifferences
in reactivityof the soil organicmatteror organicmatterquality(Fig.
3A). Thesereactivitynumbersrevealedthatthe reactivityof Ab was
A-horizonat the Cassiopesites, thatthe
higherthanthe near-surface
oppositewas notedat Salix sites, thatCO2productionratesunit soil
organic-Cin near-surface
layers(A-horizons)werethe samefor Dryas
and Salix but 40% higherthan for A-horizonmaterialat Cassiope
sites and that reactivityincreasesfrom A-horizonsto underlying
B/C-horizons(Fig. 3A).
CO2productionratesin samplesafterair dryingand rewetting
revealedno significantchangesin rates,exceptfortheDryasA-horizon
andthe DryasB/C-horizon,whichincreasedby factorsof 1.6 and 1.2,
respectively,aftera 5%increasein the watercontent(datanot shown).
ThisindicatesthatDryassitesaretheonly sitesbeingpartlylimitedby
the availabilityof water.
The productionof CO2increasedsignificantly(p < 0.05) in all
samplesafterglucoseaddition,suggestingthatsoil respirationis also
limitedby the availabilityof easily-degradable
carbonsubstrates(Fig.
3B). A maximumincreasein CO2productionafterglucose addition
was observedin A materialfrom the Cassiopesite increasingby
a factorof 2 (Fig. 3B).
C:N ratioshave previouslybeen used to relatecarbonsubstrate
% Depth-integrated organic soil-C relative to I m
% Carbon
g
3
B
.o
mID
L.
o
3
.i
._
40
A
s
o
A
30-
8
:D
4
.: ^
u
u
20-
|
I
K -4-,
.,-*
o
CD
3a
1-
Uf
10-
0
I
0
A
I
B/C
Dryas
I
C
A
I
B/C
I
Ab
Cassiope
I
C
A
I
B,
I
B.
Salix
I
Ab
C
FIGURE 3. (A) C02 production rates per unit organic soil-C (at
26EC) as observed in individual soil horizons at Dryas, Cassiope, and
Salix sites, and (B) the relative increase in C02 production rate after
glucose addition. One standard deviation (?) of 3 replicates per
horizon is shown as vertical lines.
and 7.5 times higher than measured rates in A-horizon material from
:assiope and Salix, respectively.
50
40
I
30
U
20
OIL RESPIRATIONAND C02 EFFLUXES DURING THE 2001
JROWINGSEASON
Temporal trends in observed soil CO2 effluxes, active layer
epths, water contents, and soil temperaturesduring the 2001 growing
eason (Fig. 5) indicate that efflux variations during the 2-month period
orrelated relatively well with near-surface soil temperature (Fig. 6).
o10estimates based on field observed soil temperatures at 2 cm and
bserved CO2 effluxes show Qlo values equal to 3.3 (r2 = 0.8) for
'assiope sites, 1.9 (r2= 0.7) for Dryas sites, and 2.7 (r2= 0.7) for Salix
ites. The few measurements at Eriophorum sites did not allow
10
Dryas
V
R
Salix
Cassiope
L
A
V
R
L
A
R = Roots
V = Vegetation
R
V
L = Litter
Eriophorum
L
A
V
R
L
A
A = A,-horizon
FIGURE 4. C:N ratio for vegetation, litter, roots, and near-surface
organic horizons (A-horizons) at the four vegetation sites.
described using an Arrhenius-type equation. The increase in reaction
rate per 10?C (Qlo) equaled 2.1 (r2 = 0.99) for the A-horizon for the
entire temperaturerange studied (-5?C to 27?C). For the temperature
range most relevant for this study (0-10?C), the Qlo value equals 2.5.
Results from the present study reveal that Qlo values estimated from
rates observed at 0?C, 7?C, 21?C, and 26?C could not prove any
significant deviations from the previously reported Qlo values.
Consequently, it is assumed that CO2 production rates observed in
the laboratory for different vegetation types and depth intervals can be
converted to in-situ rates by reducing the rate according to the observed
o10estimates.
Significant variations in soil CO2 effluxes were observed between
lant communities (Table 1), with minimum effluxes observed at
'assiope sites (on average 0.52 + 0.1 lmol CO2 m-2 s-1) and
naximumeffluxes observed at Eriophorumsites (on average 2.48 ? 0.5
,mol CO2 m-2 s-1). The range of these CO2 fluxes is comparable with
reviously reported soil CO2 effluxes in High Arctic Greenland
Christensen et al., 1998; Jones et al., 2000). Observed effluxes
ormalized to a temperatureof, e.g., 10?C (Fig. 6) indicate that overall
ffluxes from Cassiope and Dryas are approximately the same, but
factor of 2 less than effluxes at Salix sites. In turn,effluxes at Salix sites
re a factor of 2 less than effluxes from Eriophorum sites (Table 1).
Soil CO2 efflux measurements on barrenground and manipulated
ites after the removal of litter and the A-horizon were only possible for
Qlo value.
Additional CO2 production rates were obtained from incubated
litter material from Cassiope and Salix sites. Variations in production
rates per volume sample were on the order of 100%; average rates were
4.0
A
E 3.0E
2.0-
1.0-
0.0
I
l
....................................................
0
200
h
B
40
640
A
A
A
A
660
o0
< 800 Dryas
oo
-- Cassiope
-
[
A Salix
. .
0 Eriophorum
100.
60
'
50-
C
>.
E
_I
"
--n
n10-
A
'4
II II
D
c^ r3>-.-
A
A
-,0 _ ^ -:-_-
rl--.i...r ,
July
-8
A
tet
......Air
.........
4 i..,
2001
August
I
Il
r (
'
a
.
.
FIGURE 5. Field observations
during the 2001 growing season
(July and August) including (A)
average soil C02 effluxes shown as
marks for each of the four vegetation types and one standard deviation (+) shown as vertical lines,
(B) active layer depths, (C) precipitation and near-surface water
contents with one standard deviation (+) shown for replicates at
Cassiope sites (n = 17), and (D) air
and soil temperatures.
ELBERLING ET AL. / 533
Buchmann(2000),theratiobetweenthetwo respiration
componentsis
and
between
1:9
and
9:1.
site-specific may vary
Comparingsoil CO2
1.20,
effluxesfromnonvegetation
o
patcheswithambientfluxesmayindicateto
A
"'~
whatextentrootrespiration
influencesobservedsoil respiration
rates.
'E 1.00
Fornonvegetated
siteswithintheCassiopeheath,effluxeswerereduced
I</~ "?
o"
~?
[
to 70% comparedto sites nearbywith an averagevegetationcover.
oA*
,
-?
E 0.80'
E
?
However,when nonvegetatedsites were comparedto vegetatedsites
-^^ A
aftertheremovalof littermaterialatbothtypesof sites,thedifferencesin
0.60E
.- :
a
I^^^
rateswerenotsignificant(n= 5). LitterwasalmostabsentatDryassites
andeffluxesfromnonvegetated
patcheswereabout95%comparedto
= 0.4000
ambientcondition.This suggeststhatmicrobialrespiration
dominated
the
total
efflux
as
observed
dark
soil
chamber
CO2
(>95%)
using
0.20
measurements
at CassiopeandDryassites.However,thediscussionon
microbialversusplant-associated
I
I
I
I{ II1
respirationis controversialbecause
4
14
16
18 20
12
0
2
6
8
10
estimatesof theimportance
of microbialrespiration
arepartlyrelatedto
Temperature(?C)
the methodologyused. It remainsuncertainto what extent root
FIGURE6. Observedsoil CO2effluxesversussoil temperatures
respiration(andplantrespirationin the chamber)is actuallyreduced
(at
a depthof 2 cm).Exponentialregressionlinesare shownas solid lines
during short-termdark chamber measurements.Using different
methods,Billingset al. (1977) foundroot respirationto accountfor
for Cassiope(solidsquares,0.2203e 1196x;r2= 9.8;Qlo= 3.3, Rl =
0.73), Dryas (open squares, 0.3553e0.0662x,r2 = C).7; Qio= 1.9; Rlo=
66-90%of thetotalsoil respiration
in arcticsoils dominatedby Carex
01'3X;r2:= 0.7; Qlo= 2.7; Rlo
0.69)andSalix(solidtriangles,0.4648e?
and
whereas
Oberbauer
et al. (1992)concludedthat
sp.
Eriophorum
sp.,
= 1.28). The dashed line is based on laborator
y studiespreviously microbialrespiration
accountedfora majorportionof soil respiration
in
reportedfor Cassiope(Elberlingand Brandt,20(
a similarsite.Thedominanceof microbialrespiration
atthetwodrysites
02~2).~
in Zackenbergadds to a more generaldebateon the importanceof
someof thevegetationtypes.Patchesof barrengro>und
largeenoughfor
microbialrespiration
versusplant-associated
in
respiration,
particularly
measurements
werenotfoundatSalixandEriopho
rumsites,anda litter
the contextof a warmingclimate.Fieldexperimentshave shownthat
layerin the propersense was not detectedat Dryas and Eriophorum artificialwarmingenhancesnet ecosystemCO2efflux, primarilyby
sites.Soil CO2effluxmeasurements
on nonvegetal
tedsitesat Dryasand
stimulatingmicrobial-driven
decompositionof organicmatterrather
Cassiopesiteswere94%and70%of ambientrates,respectively(Fig.7).
thanby increasingplant-associated
respiration(Welkeret al., 1999;
Removingthe litterlayerat CassiopeandSalix ssitesreducedthe soil
Groganand Chapin,2000). Thus,the high correlationbetweennearrespirationrates to 70-75% comparedto rates from ambientsoils.
surfacetemperatures
andeffluxesobservedin this studyconfirmsthe
alreducedtheratesto
RemovingthelitterlayerandtheA-horizonmaterii
dominanceof microbialrespirationduringshort-termdarkchamber
about50%atEriophorum
sitesand18-20%atCassiopeandDryassites
butcannotbe usedto concludethatmicrobialrespiration
measurements,
y much lower for
(Fig. 7). Standarddeviations were generally
dominatesin general.
sites, especiallyafterthe removalo1f littermaterialfrom
manipulated
The apparentdominanceof microbialrespiration
at Cassiopeand
thesurface.Thissuggeststhatmostof thevariationobservedin ambient
Dryas sites makes it possible to compareactualdepth-specificsoil
al amountof carbon
replicatesrepresentsspatialvariationin the actu;
resultsobtainedin the laboratory
with field observationof
respiration
availablefor decomposition
in the litterlayer.
totalCO2effluxesfromthe soil surface.Salixsites areincludedin this
comparisoneven though the assumed dominance of microbial
respiration
may not be valid. The relativeimportanceof soil layers
SOILC02 AND02 CONCENTRATIONS
to producingCO2is evaluatedassumingsteady-state
conditions,which
Subsurfaceconcentrations
of CO2as observed fromgas samplers impliesthatthe flux at the permafrost
tableis zeroand,consequently,
wereabout10-20 timesthatof atmospheric
concerntrations
throughout thatthe totalproductionof CO2in the profileequalsthe effluxof CO2
the growingseason,andconcentrations
to the atmosphere
for a certainperiodof time.Forthatpurpose,CO2
generallyincreasedwithdepth
rost boundary,CO2
(Figs. 8 and 9). However,towardthe permafi
concentrations
tendedto decline.Thisdeclinethr(
oughoutthe growing
125
seasonsof 2000 and2001 (Fig.8) indicatesthatttheconcentrations
are
not representingtrue steady-stateconditions.H(owever,steady state
seemsto be a reasonableassumptionin the upper
rpartof the profiles.
100
Concurrentoxygen profiles (data not shown) indicatedthat nearatmospheric02 concentrations
always existed tlIroughoutthe entire
o
active layer (19-21%). However,it is worthnm
_ 75oting that these gas
o
0
}?
concentrations
in the largerair-filled
representthe bulkconcentrations
._
pores, whereas smaller pores may have locally much higher or
E 50lowerconcentrations.
1.40
,o
C
Discussion
LABORATORY
VERSUSFIELDOBSERVED
C02 PRODUCTION
Duringsteady-stateconditions,soil CO2effluxesobservedin the
field representthe combined effects of soil microbialand root
respiration,whereas CO2 productionobserved during laboratory
incubationsrepresentonly soil microbialrespiration.As reportedby
534 / ARCTIC, ANTARCTIC, AND ALPINE RESEARCH
X
25
25 -
4
* Ambient
O Barrenground
0 Afterlitterremoval
A Afterlitterand A-horizonremoval
I
0
Dryas
Cassiope
Salix
Eriophrium
FIGURE7. Observedsoil CO2effluxesat Dryasand Cassiopesites
withoutvegetation(morethan20 cmto thenearestplant)andafterthe
removalof the litterlayer(0-2 cm) and the A-horizon(5-7 cm).
0
0.1 _
-0.2
--0.13'
20
'
~.
........
..........
...........
. . . . .
.
.
.
.
.
.
as
.
. ... ...
. . ..??.....o.............
... ... . ... ???
:;
...
.z
.... . .... .... .. .... ........
............:.
.
. ........
. .... . . ... . ... . ... . . . .... . . . . . . . . . . . ....
!:!:!:!::i:!:!:!:!:!:i:i:
.ii
. . . . ii
. . . . .. . . . . . . . . ..0.5:::?:
. . .i. . . . ....
. ..... . . . . . . . . . .......
.iii...........
. . . . . . . . . . . . . . . . . . . . . . . . ......
............. ..............
........................,
.
.. .
....................
40
.........................
. . . . .,:.....
. . . . . ..,.. . . ,,.. . . . . . . . . . . . .
?
..........................
.........................
....?''
.... .
.....
......................
...
.-. -.-... ..... . . . . .... .. . .... .. .... . ..... .... ..... . ... . ... . ... .
.....
..............
...... ....
.
... .. ... ... .
.....
..... .. .
........
.........................
,.........
...............
................
s.................
... .... .
... . . ost....
sra... .. .......
.
nafr1ot .-.-.-.-.-.
.........
. 11.I.Pen
......................
.........
.
.. . ... ...
~~~~~~~~~
...
,-,-*. . .
.
.
*.*.*. . . .. .. .. . . . . ......
....................
. . . . . . . . . . . . . . . I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......................
...................................I................
.............................
I
.......................
.. .
..
. .. . .
. . .. .. .. .. .. .. .. .. ..
.
?:?:?:?:
??:-:
? ??:-:?:?::?:?:?:?::?:?:
? :?
..?it
-.
....---
*.*.* * .**
. ... .. .. ... .. .. . .... .. . . . . . ... .
li....... .. l t.........................................
. .. . . .. . ....... ...
60
July
I
August
2000
...
July
....
...
r.rt...l
rt
i
I
I
August
2001
FIGURE8. Observedsoilpore gas CO2concentrations(%)in a Cassiopesoilprofileduringthe2000 and2001 growingseasons(measurements
are shownas dots).
productionratesobservedfrom soil sampleshave been expressedper
volumesoil (Fig. 9E) andadjustedto field-observedtemperatures
(Fig.
9D). Assumingsteady-stateconditions,the totalCO2productionfrom
the surface to the permafrostboundarycan be estimated as the
integratedCO2productionprofile.Takinginto accountthe fractionof
soil respirationin the litterlayer to the total soil respiration,0% for
Dryas sites and 30% for Cassiopeand Salix sites, CO2 production
profilescan be constructed.However,the CO2distributionwithinthe
soil profile is not only a result of CO2 productionbut also of soil
diffusivitybeinga functionof the watercontent(Fig.9C) andporosity,
among other parameters(Eqation 2). In order to evaluate the
relationshipbetweenobservedCO2 concentrationprofiles,measured
% C02
0.0
0 - 'l
'
0.2
10
II
20
o
0.4
% CO2
0.6 0.0
Av
1.0
2.0
3.0
4.0
CO2productionprofiles,and soil CO2effluxes, the diffusionmodel
given by Equations1 and 2 was used.
SIMULATING
SOILCO2CONCENTRATION
PROFILESAND EFFLUXES
Simulatedresultsshown in Figs. 9A and 9B are based on data
collected on 2 August 1999. Only steady-stateCO2 concentration
profilesare simulated,which explainswhy attemptsto simulatethe
entireCO2 concentration
profile,includingobservationsbelow 30 cm,
failed for Cassiope and Salix sites. It is assumed that low CO2
concentrations
nearthe permafrostboundaryreflecta non-steady-state
% Water (per vol)
Temperature (?C)
0
20
40
60 0
5
10
CO2 production
(nmol C cm'3 s')
0
0.002 0.006
IA
\
I
I
30
J
L
ao
I
f
50'
60
I
q
o
B
o Dryas
C
* Cassiope
A
Salix
-
- --
Model
FIGURE9. Field measurements
at Dryas,Cassiope,
of (A and B) soil CO2concentration
profiles,(C) watercontents,and (D) soil temperatures
and Salix sites as observedon 2 August1999. CO2productionprofiles(E) as observedin the laboratoryhave been used as inputto a diffusion
model,and theresultingmodeloutputsare shownas solid linesfor Dryasand Cassiope(A)andSalix(B). Onestandarddeviation(?) basedon 16
replicatesof CO2concentrationsand watercontentsobservedat two depthsat Dryas and Cassiopesites is shownas horizontallines (A).
ELBERLING ET AL. / 535
situationwhereCO2diffusesthroughout
the growingseasonfromsoil
layers above as the soil profilethaws. But predictedconcentrations
withintheupper30 cmagreewelltheobservedconcentrations
(r2=0.88
forCassiopesites,0.98 forDryassites,and0.72 forSalixsites).Thus,it
is concludedthat CO2productionprofiles(Fig. 9E) providereliable
estimatesof the relativelydepth-dependent
soil CO2 production.It
shouldbe notedfurther(Fig.9E) thatCO2productionin the buriedAb
horizonrepresents
upto 20%of thetotalactivelayerof CO2production
in August.However,most of the time the decompositionwithinthe
buriedlayeris probablyfurtherlimitedby temperature
andhigherwater
to thesurfaceA-horizon.AlsotheAbhorizonmust
contentas compared
have been subjectto very slow decompositionratessince deposited
a few thousandyearsago (Christiansen
et al., 2002), andthe present
situationis thereforeunlikely to representlong-termsteady-state
conditions.
ThesimulatedtotalCO2production
to a depthof 40 cm
integrated
was within1 standarddeviation(SD) of observedeffluxesduringthe
firstweekof August1999.However,thedefinitionof the litterlayeris
becausethethicknessandreactivitywerehighlyvariable
quiteuncertain
in the field. In the modeling,the litterlayer was includedwith the
averagethicknessobserved,a reactivityseven times that of the Ahorizonand a high diffusioncoefficientonly limitedby the actual
volumeof air-filledporosity.Allowingthelitterthicknessandreactivity
to vary within the 1 SD of observations,a perfectmatchbetween
simulatedand observedeffluxescould be obtained.Excludinglitter
layersin themodelinghadverylittleeffecton soil CO2concentrations
but caused reducedCO2 effluxes (to about 70%). This important
ecologicalsignificanceof litteris consistentwithotherstudiesfromthe
et al., 2000).Thisreductionwasconsistentwith
arctic(e.g.,Fahnestock
withan
atmanipulated
sites(Fig.7) andin agreement
fieldobservations
of
litter
material.
This
further
coefficient
expectedhigh diffusivity
explainswhy the litterlayerin some vegetationtypes (Cassiopeand
Salix) is importantfor the overallreleaseof CO2to the atmosphere
forsoil CO2concentration
withouthavinga similarimportance
profiles.
Thisalsoexplainsgeneralproblemsassociatedwithapplying1-Dsoildiffusion-modelsto estimatesoil gas effluxes and that attemptsto
simulatetemporaltrendsin soil CO2effluxesin thepresentstudyfailed
(simulationsnot shown).As shown in Fig. 8, the CO2concentration
profileremainedfairlyconstantin the upper20 cm throughoutthe
growingseason despiteof pronouncedseasonaltrendsin soil CO2
effluxes(Figs.5 and6).
andhydrological
conditions.Thisstudyonly
specificplantcommunities
allowsa descriptiveevaluationbutis important
to showtheinteracting
parametersthat may determinewhetheran ecosystem (soil-plant
system)will becomea futurenet sourceor sink for organiccarbon.
Furthermore,it illustratesthat changes due to climate changes
are unlikely to be uniform across the arctic tundra landscape
(Welkeret al., 2000).
Dryassitescontaina smallamountof thetotalecosystemorganicC
(Table 1), and respirationrates are limited by the C-substrate,
andprobablyalso waterin some periods.Comparedto
temperatures,
ratesat Cassiopesitesaremorelimitedby thequality
Dryas,respiration
of C-substrate
(Fig.3) andmorelimitedby temperature
(Fig.6) dueto
a higherwatercontent.But becauseof the largerquantitiesof organic
matteratCassiopesitescompared
to Dryassites(thickerA-horizon),the
overallsoil CO2effluxatthesetwositesata giventemperature
is almost
the same (Fig. 6). Decreasingwatercontentsand increasingtemperatureswill likelyleadto a positivefeedbackin termsof near-surface
soil
respirationat Cassiopesites, whereasthe net effect on Dryassites is
more uncertain.Increasingtemperatureswill provide a positive
feedback,but the lack of watercan be a limitingfactoras discussed
by Welkeret al. (2000).Cassiopesites storelargequantitiesof organic
carbonin potentialpools of easilydegradableC below the presentAhorizon(Fig. 1), whichis presentlyonly contributing
less than20%of
thesoileffluxes.Theeffectsof globalwarminganda thickeractivelayer
on soil CO2effluxes,e.g., fromCassiopesites, can be expectedto be
even higher than predictedusing only the temperaturesensitivity
(Fig. 6), becauseCO2productionfromthe decomposition
relationship
of carbonstoredin buriedA-horizonswill be an additionalsourceof
andholdverylargequantities
CO2.Salixsitesarealsosubstrate-limited
of organicmatter(Table1), giving rise to high effluxes(Fig. 6) and
maximumsoil CO2 concentrations(Fig. 9B). Decompositionof
potentialpools is sensitiveto the watercontent,as decreasingwater
contentsarelikelyto increaseboththe availabilityof oxygenandsoil
temperaturesand therebyincreasedecomposition.In that respect,
sitesaremoreextreme,holdinganactivecarbonpoolof the
Eriophorum
most reactiveC-sourceand presentlygiving rise to maximumsoil
effluxesthroughout
the growingseason(Fig. 5). As Eriophorum
sites
arepartlywatersaturated
the
season,
throughout growing
increasingair
will
soil
but
enhance
respirationdirectly
may also
temperatures
withchangesin waterbalance)influence
indirectly(andin combination
the depthto the watertableandtherebychangesoil redoxconditions.
A shift fromanaerobicto aerobicconditionsstops the productionof
IN THEARCTICREGION-ACTIVE
CARBON
RESERVOIRS
methanebut generallyincreasesCO2productionrates(Chapmanand
ANDINACTIVE
POOLS
Thurlow,1996; Blodau and More, 2003). Finally,changes in the
are closely relatedto changes in winter preIn mostarcticecosystems,easilydecomposable
organicmaterials hydrologicalregime
snow
and the overallwind regime(Zimov
redistribution,
cipitation,
horizons(Van Cleve, 1974), and the high
are foundin near-surface
et
et
Fahnestock
al.,
1998,
al.,
1996;
1999).
of organiccarbonstoredin theupperpartof the soil profile
proportion
From the discussionabove it appearsthat temperature,
water
has been noted for a rangeof high-latitudetundrasoils (Post et al.,
all crucialparameters
and
are
substrate
the
content,
controlling
quality
in
both
Aand
the
of
carbon
However,
Ab-horizons
1982).
importance
in
in
observed
variation
CO2dynamics Zackenberg.
spatial
found in this study is probablya more site-specificcharacteristic, presently
in theseparameters
will ratherquicklyanddirectlyaffectthe
althoughit may also applyto othersoils. A comparisonof substrate Changes
activesoil carbonpool andthereforethesoil carbondynamics.Indirect
qualityand CO2productionratesin A- and Ab-horizonsis complex.
andmoreslowlyrespondingeffectsmayalsohavea significantimpact
The sourceof carbonin buriedlayerswas most likely differentfrom
the planttypes foundat specificsites todayas indicatedby abundant onthesoilcarbonpool.Sucheffectsarelikelyto be theresultof changes
on a landscapescale, e.g., changingwind pattern,snow distribution,
birchleaves foundin the Ab-horizonof previouslydescribedPodzols
draining,and formationof thermocasts (Evanset al., 1989;Hobbie
(Christiansenet al., 2002). It is worth noting that Ab-materialat
et al., 2000).A finalindirectandmoreslowlyresponding
climaticeffect
than
in
seems
more
reactive
material
sites
significantly
Cassiope
in
is
the
structure
and
area
distribution
shift/transition
the
to
whereas
seems
be
near-surface
community
plant
A-horizon,
opposite
present-day
of vegetationtypes(Oberbauer
et al., 1996;GroganandChapin,2000).
the case for Salix sites (Fig. 3).
It is worthnotingthatthe most importanteffects may be the slowly
withinthe
Forevaluatingtheorganiccarbonreservoirsdistributed
andindirecteffectswhicharepoorlydescribedforthearctic
to
evaluate
the
controls
on
it
is
responding
organic
Zackenbergvalley
important
and which presentlymake it difficultto providereliablelong-term
matterdecompositionfor severalorganiccarbonreservoirs:active,
estimatesfor globalchangefeedbackin the arctic.
potential,and inactivepools of which some seem closely linkedto
536 / ARCTIC, ANTARCTIC, AND ALPINE RESEARCH
CONCLUSIONS
Eurasianand Greenlandicarctic.Journalof GeophysicalResearch,
103:29,015-29,021.
Severalbiogeochemicalandphysicalfactorsandtheirinteractions
Christiansen,H. H., Bennike,O., B6cher,J., Elberling,B., Humlum,
affectsoil organiccarbonreservoirsandmineralization.
This studyhas
0., andJakobsen,B. H., 2002: The Zackenbergdelta-a Holocene
focused on the interactionsbetween depth-specificdistributionand
environmentalarchive in NE Greenland.Journal of Quaternary
and
reactivityof carbonsubstrateandvegetationtypes.Fromlaboratory
Sciences, 17: 145-160.
fieldobservationsas well as modeling,thefollowingconclusionscanbe
Coyne, P. I., and Kelley, J. J., 1974: Releaseof carbondioxidefrom
frozensoil to the Arcticatmosphere.Nature,234: 407-405.
made.Morethan95% of total ecosystem-Cwas foundas potentially
De Jong, E., and Schappert,H. J. V., 1972: Calculationof soil resreactivesoil organiccarbonpools. Samplingin orderto quantifytotalC
pirationandactivityfromCO2profilesin the soil. Soil Sciences,113:
reservoirsrequiresa procedurethat takes into account site-specific
328-333.
activelayerdevelopment,e.g., the occurrenceof buriedsurfacelayers.
B., 2003: Seasonaltrendsof soil CO2dynamicsin a soil
Elberling,
The simulatedCO2 activityprofiles suggest that buriedold surface
to freezing.Journalof Hydrology,276: 159-175.
subject
layers(Ab)arepresentlyimportantfor the overallsoil CO2dynamics.
Elberling,B., andBrandt,K. K., 2003: Uncouplingof microbialCO2
Decompositionwithinthe litterlayerinfluencesstronglyon total soil
productionand CO2releasein frozen soil and its implicationsfor
CO2fluxesobserved,contributing
up to 30%of the totalCO2fluxesto
field studiesof arcticC cycling.Soil Biologyand Biochemistry,35:
the atmospherebut withouthavingthe equivalentimportanceon the
263-272.
the spatialdistributionof
subsurfaceCO2 distribution.Furthermore,
Elberling,B., and Jakobsen,B. H., 2000: Soil solutionpH measurements using in-line chamberswith tension lysimeters.Canadian
littermaterialseemsto be responsiblefor mostof the spatialvariations
Journalof Soil Science,80: 283-288.
of observedsoil CO2effluxes.
Evans,B. M., Walker,D. A., Benson, C. S., Nordstrand,E. A., and
between terrain,
Petersen,G. W., 1989: Spatialinterrelationships
snow distribution
andvegetationpatternsat an arcticfoothillssite in
Acknowledgments
Alaska.HolarcticEcology, 12(3):270-278.
This study was fundedby the DanishNaturalScience Research Fahnestock,J. T., Jones, M. H., Brooks,P. D.,
Walker,D. A., and
CouncilStenogrant(reg.no. 1235;B. Elberling).Thanksareextended
Welker,J. M., 1998:WinterandearlyspringCO2effluxfromtundra
to the DanishPolarCentre(M. Rasch)for access to climatedataand
communitiesof northernAlaska.Journalof GeophysicalResearch,
to the laboratory(M. Andersenand S.N. Rasmussen)at the Institute
103: 29,023-29,027.
of Geography,Universityof Copenhagen,for making most of the
Fahnestock,J. T., Jones,M. H., andWelker,J. M., 1999:Wintertime
soil analyses.
CO2effluxfromarcticsoils:implicationsfor annualcarbonbudgets.
GlobalBiogeochemistryCycles, 13(3):775-779.
J. T., Povirk,K. L., and Welker,J. M., 2000: Ecological
Fahnestock,
Cited
References
significanceof litter redistributionby wind and snow in arctic
landscapes.Ecography,23(5): 623-631.
Anderson,J. P. E., and Dosch, K. H., 1978: A physiologicalmethod
of microbialbiomassin soils. Soil
for the quantitative
measurements
Fang, C., and Moncrieff,J. B., 2001: The dependenceof soil CO2
effluxon temperature.
Soil Biologyand Biochemistry,33: 155-165.
Biology and Biochemistry,10: 215-221.
and
F. S., 2000: Initialeffects of experimental
P.,
of
A
the
vascular
on
1992:
Grogan,
Chapin,
C.,
plants
Bay,
phytogeographicstudy
northernGreenland-northof 74 degreeslatitude.Meddelelserom
warmingon above-andbelowgroundcomponentsof net ecosystem
CO2exchangein arctictundra.Oecologia, 125(4):512-520.
Gr0nland,Bioscience,36: 3-102.
effects of
Bay, C., 1998: Vegetationmappingof Zackenbergvalley, Northeast Grogan,P., andChapin,F. S., 1999:Arcticsoil respiration:
climateand vegetationdependon season.Ecosystems,2: 451-59.
Greenland.DanishPolarCenterandBotanicalMuseum,University
andplantspeciescontrolover litter
Hobbie,S. E., 1996:Temperature
of Copenhagen.29 pp. (availablefromwww.dpc.dk)
decompositionin Alaskan tundra.Ecological Monographs,66:
N., andRysgaard,S., 1998:Interpretation
Berg,P., Risgaard-Petersen,
503-522.
of measured concentrationprofiles in sediment pore water.
Hobbie,S. E., Schimel,J. P., Trumbore,S. E., and Randerson,J. R.,
Limnologyand Oceanography,43: 1500-1510.
2000: Controlsover carbonstorageand turnoverin high-latitude
Billings, W. D., Peterson,K. M., Shaver,G. R., and Trent,A. W.,
soils. GlobalChangeBiology,6: 196-210.
1977: Root growth,respirationand carbondioxideevolutionin an
arcticsoil. Arcticand AlpineResearch,9: 129-137.
Howard,D. M., and Howard,P. J. A., 1993: Relationshipsbetween
for a rangeof soil
CO2evolution,moisturecontentandtemperature
Blodau, C., and Moore, T. R., 2003: Micro-scaleCO2 and CH4
types.Soil Biology and Biochemistry,25: 1537-1546.
dynamicsin a peat soil duringa waterfluctuationandsulfatepulse.
Soil Biology and Biochemistry,35: 535-547.
Jakobsen,B. H., 1991:Multipleprocessesin theformationof subarctic
In Page,A.
podzolsin Greenland.Soil Sciences, 152: 414-426.
Bremner,J. M. andMulvaney,C. S., 1982:Nitrogen-total.
the dependenceof gas
L. et al. (eds.), Methodsof soil analysis. Part 2. Chemicaland
Jin, Y., and Juri,W. A., 1996: Characterizing
diffusion coefficient on soil properties.Soil Science Society of
microbiologicalproperties. Madison, WI, American Society of
AmericaJournal,60: 66-71.
Agronomy,595-624.
Jones,M. H., Fahnestock,J. T., Stahl,P. D., andWelker,J. M., 2000:
Buchmann,N., 2000: Biotic and abiotic factors controlling soil
A note on summerCO2 flux, soil organicmatter,and microbial
respirationrates in Picea abies stands. Soil Biology and BiobiomassfromdifferentHighArcticecosystemtypesin Northwestern
chemistry,32: 1625-1635.
Greenland.Arctic,Antarctic,and AlpineResearch,32: 104-106.
Burkins,M. B., Virginia,R. A., andWall,D. H., 2001:Organiccarbon
Kirschbaum,M. U. F., 1995: The temperaturedependenceof soil
cycling in TaylorValley,Antarctica:quantifyingsoil reservoirsand
soil respiration.GlobalChangeBiolology,7: 113-125.
organicmatterdecomposition,andthe effect of global warmingon
soil organicC storage.Soil BiologyandBiochemistry,27: 753-760.
Caning,K., and Rasch, M., 2000: ZackenbergEcological Research
In Page,A. L., et al. (eds.),Methodsof soil
Kuo,S., 1996:Phosphorus.
Operations,5th AnnualReport,1999. Copenhagen,Danish Polar
Center,Ministryof Researchand InformationTechnology.86 pp.
analysis. Part 3. Chemical methods. Soil Science Society of
Americaand AmericanSociety of Agronomy,Madison,WI, pp.
(availablefromwww.dpc.dk)
869-919.
Chapman,S. J., and Thurlow,M., 1996: The influenceof climateon
Lomander,A., Katterer,T., and Andr6n,0., 1998: Carbondioxide
CO2andCH4emissionsfromorganicsoil. Agriculturaland Forest
evolutionfromtop- andsubsoilas affectedby moistureandconstant
Meteorology,79: 205-217.
and fluctuatingtemperature.Soil Biology and Biochemistry,14:
Christensen,T. R., Jonasson,S., Michelsen,A., Callaghan,T. V., and
controlson soil respiration
in the
2017-2022.
Havstr6m,M., 1998:Environmental
ELBERLING ET AL. / 537
H., 1927: Carbondioxide evolutionof soil and crop
Lundegiirdh,
growth.Soil Science,23: 417-453.
Maxwell,B., 1997:Recentclimatepatternsin thearctic.In Oechel,W.
C., Callaghan,T., GilmanovT., Holten,J. I., Maxwell,B., Molau,
U., and Sveinbj6msson,B. (eds.), Global change and arctic
terrestrialecosystems.New York,Springer,21-46.
McKeague,J. A., and Day, J. H., 1966: Dithionite-and oxalateextractable
Fe andAl as aidsin differentiating
variousclassesof soil.
CanadianJournalof Soil Sciences,46: 13-22.
Millington,R. J. 1959:Gas diffusionin porousmedia.Science, 130:
100-102.
Nadelhoffer,K. J., Giblin,A. E., Shaver,G. R., and Laundre,J. R.,
1991: Effects of temperatureand substratequality on element
mineralization
in six arcticsoils. Ecology,72: 242-253.
Oberbauer,S. F., Tenhunen,J. D., and Reynolds, J. F., 1991:
Environmental
effectson CO2effluxfromwatertrackandtussock
tundrain the arcticAlaska,U.S.A.Arcticand AlpineResearch,23:
162-169.
S. F., Gillespie,C. T., Cheng,W., Gebauer,R., SalaSerra,
Oberbauer,
effectson CO2efflux
A., andTenhunen,J. D., 1992:Environmental
fromripariantundrain the northernfoothillsof the BrooksRange,
Alaska,U.S.A. Oecologia,92: 568-577.
S. F., Gillespie,C. T., Cheng,W., Sala,A., Gebauer,R., and
Oberbauer,
Tenhumen,J. D., 1996:Diurnalandseasonalpatternsof ecosystem
CO2effluxfromuplandtundrain the foothillsof the BrooksRange,
Alaska,U.S.A.ArcticandAlpineResearch,28: 328-338.
Oechel, W. C., Vourlitis,G. L., Brooks, S., Crawford,T. L., and
Dumas, E., 1998: Intercomparison
among chamber,tower, and
aircraftnet CO2 and energy fluxes measuredduringthe Arctic
Science Land-Atmosphere-Ice
Interactions(ARCSS-LAII)flux
study.Journalof GeophysicalResearch,103:28,993-29,003.
Oechel,W. C., Hastings,S. J., Vourlitis,G. L., Jenkins,M., Riechers,
G., andGrulke,N., 1993:Recentchangeof arctictundraecosystems
froma net carbondioxidesink to a source.Nature,361: 520-523.
538 / ARCTIC, ANTARCTIC, AND ALPINE RESEARCH
A. G.,
Post,W. M., Emanual,W. R., Zinke,P. J., andStangenberger,
1982:Soil carbonpoolsandworldlife zones.Nature,298: 156-159.
In Sparks,D. (ed.), Methodsof
Sims,J. T. 1996:Limerequirements.
soil analysis. Part 3. Chemicalmethods.Madison,Wisconsin,
AmericanSocietyof Agronomy,SSSA Book seriesno. 5, 491-515.
T.
C., Friborg,T., Hansen,B., Christensen,
S0gaard,H., Nordstroem,
R., andBay, C., 2000:Tracegas exchangein a high-arctic
valley.3.
andscalingCO2fluxesfromcanopyto landscapeusing
Integrating
flux data, footprintmodeling,and remote sensing. Global BiogeochemicalCycles,14: 725-744.
Van Cleve. K., 1974: Organicmatterquality in relationto decomposition.In Holding,A. J., Heal,O. W., MacLean,S. F., Jr.,and
Flanagan,P. W. (eds.), Soil organismsand decompositionin
tundra.Stockholm,Sweden, TundraBiome SteeringCommittee,
311-324.
Welker,J. M., Brown,K. B., andFahnestock,J. T., 1999:CO2fluxin
Arctic and alpine dry tundra:comparativefield responsesunder
ambientand experimentally
warmedconditions.Arctic,Antarctic,
and AlpineResearch,31(3):272-277.
Welker,J. M., Fahnestock,J. T., andJones,M. H., 2000:AnnualCO2
flux in dry andmoist arctictundra:Field responsesto increasesin
summertemperaturesand winter snow depth. ClimateChange,
44(1-2): 139-150.
T. H., andMcDermitt,D. K., 2001:
Welles,J. M., Demetriades-Shah,
Considerations
for measuringgroundCO2effluxeswith chambers.
ChemicalGeology,177: 3-13.
S. F.,
Zimov,S. A., Davidov,S. P., Voropaev,Y. V., Prosiannikov,
Semiletov,I. P., Chapin,M. C., and Chapin,F. S. 1996:Siberian
CO2efflux in winteras a CO2sourceand cause of seasonalityin
CO2.ClimaticChange,33(1): 111-120.
atmospheric
Revisedms submittedFebruary2004