GLOBAL BIOGEOCHEMICAL
CYCLES, VOL. 11, NO. 2, PAGES 173-189, JUNE 1997
Equilibrium responsesof global net primary production
and carbon storageto doubled atmospheric carbon dioxide:
Sensitivityto changesin vegetationnitrogen concentration
A. David McGuire
u.s. GeologicalSurvey,AlaskaCooperativeFish andWildlife ResearchUnit, Universityof Alaska,Fairbanks
JerryM. Melillo, DavidW. Kicklighter,Yude Pan,XiangmingXiao, andJohnHelfrich
The Ecosystems
Center,Marine BiologicalLaboratory,WoodsHole, Massachusetts
BerrienMoore III, CharlesJ. Vorosmarty,andAnnetteL. Schloss
ComplexSystemsResearchCenter,Institutefor the Studyof Earth,Oceans,andSpace
Universityof New Hampshire,Durham
Abstract.We rantheterrestrial
ecosystem
model(TEM) fortheglobeat0.5øresolution
for
atmospheric
CO2concentrations
of 340 and680 partsper millionby volume(ppmv)to evaluate
globalandregionalresponses
of netprimaryproduction(NPP) andcarbonstorageto elevated
CO2for their sensitivityto changes
in vegetation
nitrogenconcentration.At 340 ppmv,TEM
15
1
estimatedglobalNPP of 49.0 10 g (Pg) C yr- andglobaltotalcarbonstorageof 1701.8Pg C;
theestimateof totalcarbonstoragedoesnot includethe carboncontentof inertsoilorganic
matter. For the referencesimulationin whichdoubledatmospheric
CO2wasacco.mpanied
with
nochange
invegetation
nitrogen
concentration,
global
NPPincreased
4.1PgC yr-1(8.3%),
and
globaltotalcarbonstorageincreased114.2Pg C. To examinesensitivityin the globalresponses
of NPP andcarbonstorageto decreases
in the nitrogenconcentration
of vegetation,we compareddoubledCO2responses
of thereferenceTEM to simulations
in whichthe vegetationnitrogenconcentration
wasreducedwithoutinfluencingdecomposition
dynamics("lower N"
simulations)andto simulationsin whichreductionsin vegetationnitrogenconcentration
influencedecomposition
dynamics("lowerN+D" simulations).We conductedthreelowerN simulationsandthreelowerN+D simulations
in whichwe reducedthenitrogenconcentration
of
vegetationby 7.5, 15.0, and22.5%. In the lowerN simulations,
the responseof globalNPP to
doubled
atmospheric
CO2increased
approximately
2 PgC yr'1foreach
incremental
7.5%reductionin vegetationnitrogenconcentration,
andvegetationcarbonincreasedapproximately
an
additional40 Pg C, andsoilcarbonincreasedan additional30 Pg C, for a totalcarbonstorage
increaseof approximately
70 Pg C. In the lowerN+D simulations,
theresponses
of NPP and
vegetationcarbonstoragewererelativelyinsensitiveto differencesin the reductionof nitrogen
concentration,
but soilcarbonstorageshoweda largechange.The insensitivityof NPP in the
N+D simulationsoccurredbecausepotentialenhancements
in NPP associated
with reduced
vegetation
nitrogenconcentration
wereapproximately
offsetby lowernitrogenavailabilityassociatedwith the decomposition
dynamicsof reducedlitternitrogenconcentration.
For each7.5%
reductionin vegetationnitrogenconcentration,
soilcarbonincreasedapproximatelyan additional60 Pg C, while vegetationcarbonstorageincreasedby onlyapproximately
5 Pg C. As
the reductionin vegetationnitrogenconcentration
getsgreaterin the lower N+D simulations,
moreof the additionalcarbonstoragetendsto becomeconcentrated
in the northtemperateborealregionin comparison
to thetropics.Otherstudieswith TEM showthatelevatedCO2
morethanoffsetsthe effectsof climatechangeto causeincreasedcarbonstorage.The resultsof
thisstudyindicatethatcarbonstoragewouldbe enhancedby the influenceof changesin plant
nitrogenconcentration
on carbonassimilationanddecomposition
rates. Thus changesin vegetationnitrogenconcentration
mayhaveimportantimplications
for the abilityof theterrestrial
biosphere
to mitigateincreases
in the atmospheric
concentration
of CO2andclimatechangesassociated with the increases.
Copyright1997 by theAmericanGeophysical
Union.
Introduction
Papernumber97GB00059.
The atmosphericconcentration
of CO2 continuesto increase
becauseof human activity [Watsonet al., 1990, 1992]. The
0886-6236/97/97GB-00059512.00
173
174
MCGUIRE ET AL.: CO2,VEGETATION [N], AND GLOBAL CARBON STORAGE
1995 assessment
of the IntergovemmentalPanel on Climate
Change(IPCC) indicatesthat atmosphericlevelsof CO2 could
be greaterthan 800 partsper million by volume(ppmv)by the
year 2100 if there is no changein the trend of fossil fuel use
[IPCC WorkingGroupI (WGI), 1996]. The buildupof CO2 in
the atmosphere
has the potentialto affect net primaryproduction (NPP) and carbonstorageof terrestrialecosystems.Net
primaryproductionis the rate at which vegetationin an ecosystem fixes carbonfrom the atmosphere(grossprimaryproduction) minusthe rate at which it returnscarbonto the atmosphere
(plantrespiration).The responses
of NPP andcarbonstorageto
elevatedCO2 are importantto understandbecausethey may
have substantialeffects on humans. BecauseNPP represents
food, fuel, and fiber for human consumption[Vitouseket al.,
1986], the responseof NPP may affect the availabilityof these
resources.Risinglevelsof atmospheric
CO2 havethe potential
to increaseglobalsurfaceair temperatureand changeprecipitation and solarradiationpatternsover the next century[Mitchell
et al., 1990]. Becausethe response
of carbonstoragewill affect
the rate of CO2 accumulationin the atmosphere,
it may influencethe rate at which climatemight change. Thus it is important to identify the potentialrange and spatialdistributionof
NPP andcarbonstorageresponses
to elevatedatmospheric
CO2.
A major uncertaintyaboutthe responses
of NPP and carbon
storageto elevatedatmosphericCO2 concernsthe role of the
nitrogencycle in theseresponses[McGuire et al., 1995a]. At
the tissuelevel, an approximatedoublingof atmosphericCO2
reducesleaf nitrogenconcentration
an averageof 21% but hasa
smallereffecton nitrogenconcentrations
in stemsandfine roots
of woody plants[McGuire et al., 1995a]. On average,overall
nitrogenconcentration
of woodyplantsdecreases
an averageof
15% in responseto an approximatedoublingof atmospheric
CO2 [McGuire et al., 1995a]. Althoughexperimentalresearch
hasconfirmedthatsoil nitrogenavailabilityoftenconstrains
the
responseof plant growth to elevatedCO2, our knowledgeof
how CO2-inducedchangesin plant nitrogenconcentration
influencesthe responseof NPP and carbonstorageis basedon a
small number of studies[McGuire et al., 1995a]. Decreasesin
plantnitrogenconcentration
are hypothesized
to influenceNPP
andcarbonstoragethroughthe dynamicsof carbonassimilation
anddecomposition.
The dynamicsof carbonassimilationare influencedby leaflevel photosynthesis
and ecosystem-level
allocationresponses
to changesin tissuenitrogenconcentrations.Among studies
that manipulateboth CO2 and nitrogenavailability,a linearrelationshipexistsbetweenphotosynthetic
enhancement,
change
in leaf nitrogenconcentration,
and the amountof CO2 change
accordingto the relationship:
phericCO2, the leaf-nitrogenterm indicatesa relative photosyntheticreductionof 6.2%, whereasthe atmospheric-CO2
term
indicatesa relativephotosynthetic
increaseof 60.5%. Therefore
at the tissuelevel, the influenceof reductionsin leaf nitrogen
appearto be secondarycomparedwith the rise in atmospheric
CO2. In contrast,soilnitrogenavailabilityis an importantfactor
that oftenconstrains
the responseof woodyplantgrowthto elevated CO2 [see McGuire et al., 1995a]. This observationis
consistent with numerous studies that have demonstrated that
net primaryproductionin northernandtemperateecosystems
is
knownto be limitedby the availabilityof inorganicnitrogenin
the soil [Mitchelland Chandler,1939; Saffordand Filip, 1974;
Van Cleve and Zasada, 1976; Auchmoodyand Smith, 1977;
Dodd and Lauenroth,1979; Ellis, 1979; Shaverand Chapin,
1980; Risser et al., 1981; Miller, 1981; Aber et al., 1982; Peter-
son, 1982; Pastor et al., 1984; Binkley, 1986; Chapin et al.,
1986; Shaverand Chapin, 1986; Chapin, 1991; Vitousekand
Howarth, 1991]. Furthermore,it has been observedthat productionmay increasesubstantiallyin responseto increasedsoil
nitrogenavailabilityevenwhen thereis little responseof tissue
photosynthesis
[seeVitouseket al., 1993]. The combinationof
evidencefrom CO2-manipulation
studiesandfrom nitrogenfertilization experimentssuggeststhat the ecosystem-levelresponseof carbonassimilationto elevatedCO2 may dependon
how nitrogenuptakeby the vegetationand nitrogenrecycling
within the vegetationinfluencethe ability of plantsto incorporate elevatedCO2 in the constructionof canopy,stem,and root
biomass. If elevatedCO2 leadsto a reductionin the nitrogen
concentration
of planttissues,thenthe responseof NPP to elevatedCO2maybe enhancedbecauselessnitrogenis involvedin
the construction of new biomass.
The effect of elevatedCO2 in reducingnitrogenconcentration of planttissuemay alter soil nitrogenavailabilityby influencingdecomposition.Comparedto leaf litter of woodyplants
grown at baselineCO2, decreasednitrogenconcentrationfor
leaf litterof plantsgrownat elevatedCO2hasbeenobservedfor
numerousspecies[seeMcGuire et al., 1995a]. Ratesof leaf
decomposition
are often correlatedwith severalindicesof nitrogenlitter quality,which includenitrogenconcentration,
carbon/nitrogenratio, and lignin/nitrogenratio [Melillo et al.,
1982]. Nitrogenconcentration
is generallypositivelycorrelated
with decomposition,
whereasthe othertwo indicesgenerallyare
negativelycorrelated.If decomposition
is depressed
becauseof
CO2-inducedchangesin litter quality,soil nitrogenavailability
may be reducedin ecosystems. Becausereducednitrogen
availabilityhas the potentialto limit productivityresponses
to
elevatedCO2,reducedlitter qualityresultingfrom elevatedCO2
has the potentialto causelong-termnegativefeedbackto constrainthe response
of NPP.
Pne/Pnb
= 0.95924 + 0.00298 dLN + 0.00178 dCa
Reductionsin vegetationnitrogenconcentrationmay have
where Pnbis net photosynthesis
per unit leaf area for plants importantimplicationsfor the role of the terrestrialbiospherein
stabilizingthe concentration
of atmospheric
CO2. Thesereducgrown and measuredat both baselineCO2 and the lowestlevel
of fertilizationin the experiment,Pneis the net photosynthesis tions may alter the CO2 responseof NPP, which influences
rate for plants grown and measuredat elevated CO2 and/or vegetationcarbonstorage,andthe CO2 responseof decomposihigherlevelsof nitrogenfertilization,dLNis the percentchange tion, which influencessoil carbon storage. The responseof
in nitrogenconcentration
betweenleavescorresponding
to the vegetationcarbonstoragewill dependon whetherthe potential
measurement
of Pnband thoseof Pne,and dCais the concentra- for lower vegetationnitrogenconcentration
to enhancethe retiondifferencein ppmvbetweenelevatedandbaselineCO2[see sponseof NPP to elevatedCO2 is strongeror weakerthan the
McGuire et al., 1995a]. For a 21% decreasein leaf nitrogen potentialfor lower litter quality to depressthe NPP response.
concentrationassociatedwith a 340 ppmv increasein atmos- The responseof soil carbonstoragewill dependon the degree
175
MCGUIREETAL.:CO2,VEGETATION
[N],ANDGLOBALCARBONSTORAGE
to which decomposition
ratesare depressedby reductionsin
litter qualityassociated
with lower vegetationnitrogenconcen-
Atmospheric Carbon Dioxide
tration.
Althoughreductionsin vegetationnitrogenconcentration
associatedwith elevatedatmospheric
CO2havethe potentialto influencecarbonstorage,it is not clearwhethervegetationor soils
will have a strongerresponse.In addition,thereis uncertainty
about the degreeto which vegetationnitrogen concentration
may respondto elevatedCO2. Most studiesof plantresponses
to elevatedCO2 have involved experimentalmanipulationof
developingseedlings.Becausevegetationnitrogenconcentration generallydecreases
duringdevelopment,
it hasbeenargued
that observedreductionsin nitrogenconcentrationassociated
with elevatedCO2 may, in part, representacceleratedgrowth
[Agren, 1994]. In this study,we evaluateglobal and regional
responses
of NPP, vegetationcarbonstorage,and soil carbon
storageto elevatedCO2 for their sensitivityto changesin vegetationnitrogenconcentration.For evaluatingthe sensitivityof
theseresponses,
we useversion4.0 of the terrestrialecosystem
model(TEM) [McGuireet al., 1995b],whichis a biogeochemical modelthat makesgeographically
referencedestimatesof the
majorcarbonand nitrogenfluxesand pool sizesfor the global
terrestrialbiosphere[seeMelillo et al., 1993].
Model Description
Overview
The TEM usesspatiallyreferencedinformationon climate,
soils, and vegetationto make monthlyestimatesof important
carbonandnitrogenfluxesandpool sizesfor the terrestrialbiosphere(Figure1). The first two versionsof TEM wereusedto
examinepatternsof NPP in SouthAmerica[Raichet al., 1991]
and North America [McGuire et al., 1992]. The third versionof
TEM was used to examinethe responseof NPP to elevated
temperatureandcarbondioxidefor temperateforests[McGuire
et al., 1993] and to generalcirculationmodel (GCM) predicted
climate change for the terrestrialbiosphere[Melillo et al.,
1993]. The carbonstoragepredictionsof the third versionwere
also evaluatedfor global terrestrialecosystems[Melillo et al.,
1995] and for grasslands
and conifer forests[McGuire et al.,
1996]. In this study,we use version4.0 of TEM, which was
modifiedfrom version3 to improvepatternsof soilcarbonstorage along gradientsof temperature,moisture,and soil texture
[seeMcGuire et al., 1995b;vegetation/ecosystem
modelingand
analysisproject (VEMAP) Members, 1995; Pan et al., 1996;
Xiao et al., 1997]. Becausethe NPP and carbonstorageresponsesof TEM to elevatedCO2 and associatedchangesin
vegetationnitrogenconcentration
will dependon the responses
of both plant and soil processes
in the model, we review the
productionanddecomposition
formulations
in TEM.
Production
To understand
how NPP estimatedby TEM 4.0 respondsto
elevatedatmosphericCO2, it is necessaryto understandhow
NPP is determined.For eachmonthlytime stepin a modelrun,
NPP is calculated
asthe differencebetweengrossprimaryproduction(GPP) andplantrespiration(R•). The calculationof RA
considers
both maintenance
respiration[McGuire et al., 1992,
1993] and constructionrespiration[Raich et al., 1991]. The
GPP
•......IR•,
Cv
"
i
!
''
'
Lc
C
VEGETATION
:i
SOIL
'vs
I INs
NUPTAKE
L
NUPTAKE
S
NETNMIN
NINPUT
•
NLOST
V
Figure 1. The Terrestrial EcosystemModel (TEM). The
state variablesare carbon in the vegetation(Cv); structural
nitrogen in the vegetation (Nvs); labile nitrogen in the
vegetation(NvL); organic carbon in soils and detritus (Cs);
organicnitrogenin soils and detritus(Ns); and availablesoil
inorganicnitrogen(N^v). Arrows show carbonand nitrogen
fluxes; GPP, gross primary production; R^, autotrophic
respiration;Ri•, heterotrophicrespiration;Lc, litterfall carbon;
LN, litterfall nitrogen;NUPTAKEs, nitrogenuptake into the
structural nitrogen pool of the vegetation; NUPTAKEt•,
nitrogenuptakeinto the labile nitrogenpool of the vegetation;
NRESORB, nitrogen resorptionfrom dying tissue into the
labile nitrogen pool of the vegetation;NMOBIL, nitrogen
mobilizedbetweenthe structuraland labile nitrogenpools of
the vegetation;NETNMIN, net nitrogenmineralizationof soil
organicnitrogen;NINPUT, nitrogeninputsfrom outsidethe
ecosystem;andNLOST, nitrogenlossesfrom the ecosystem.
flux GPP considers the effects of several factors and is calcu-
latedat eachtime stepasfollows:
GPP = Cma
x f(PAR) f(LEAF) f(T) f(Ca, Gv) f(NA)
where Cma
x is the maximum rate of C assimilation,PAR is
photosynthetically
activeradiation,LEAF is leaf arearelativeto
maximumannualleaf area,T is temperature,
Cais atmospheric
carbondioxide,Gv is relativecanopyconductance,
and NA is
nitrogenavailability.
The response
of GPPto atmospheric
CO2is affectedby three
aspects
of leaf-levelcarbonassimilation
[Farquharet al., 1980;
Wullschleger,1993; Petterssonand McDonald, 1994; Sage,
1994]: carboxylation,
light harvest,and carbohydrate
synthesis.
Under saturatinglight conditionsat low levelsof intercellular
CO2, assimilationis limited by the quantity and activity of
ribulosebisphosphate
carboxylase
(rubisco),the enzymethatis
primarilyresponsible
for capturingatmospheric
carbonin the
production of sugars. Rubisco may accept either CO2
(carboxylation)
or O2 (oxygenation)as a substrate;
oxygenation
leadsto photorespiration.BecauseCO2 competeswith 02 for
rubiscobinding sites,enhancement
of photosynthesis
by elevatedCO2is possiblethroughincreasedcarboxylation
and decreasedoxygenation.Carboxylation
increases
with risingintercellularCO2until levelswherethe regeneration
of rubisco,and
thusthe ability to fix carbon,is limited by the light harvesting
176
MCGUIREET AL.: CO2,VEGETATION[N], AND GLOBALCARBONSTORAGE
main unchangedover the carboxylation-limited
rangeof intercellularCO2. Alternatively,if they representlessallocationof
nitrogento photosynthesis,
thenlower assimilationper unit leaf
areais expectedin plantsgrownat elevatedCO2. This acclimation responseof photosynthesis
is categorizedas down regulaThe assimilation-intercellular
CO2 (A-CO relationshipis the
capacityof plants
empiricalobservation
of carboxylation-limited,
light-limited, tion, which occurswhenthe photosynthetic
in comparisonto plantsgrown
andsynthesis-limited
assimilation
overtherangeof intercellular grownin elevatedCO2decreases
for plantsgrown
CO2[Wullschleger,
1993;Sage,1994]. In TEM, theA-Ci rela- at baselineCO2,but the rateof photosynthesis
andmeasuredat elevatedCO2is not lessthanthe ratefor plants
tionshipis represented
by theproductf(Ca,Gv) f(NA).
Althoughthe individualmechanisms
of assimilationhave grown and measuredat baselineCO2 [see Luo et al., 1994].
been modeled[Farquhar et al., 1980; Farquhar and von Down regulationappearsto be the predominantphotosynthetic
acclimation responseof woody plants to elevated CO2
Caemmerer, 1982], the A-Ci relationshipcan effectively be
1994;McGuire et al., 1995a].
modeledwith a hyperbolicrelationship
that collectivelyrepre- [Gundersonand Wullschleger,
In TEM, it is importantto recognizethatthe response
of GPP
sentsthe mechanisms
of assimilation.In TEM, the overallhyperbolicnatureof the A-Ci relationship
is represented
in the to doubledCO2 is not a constant37% for 1%of 400 ppmvbefunctionf(Ca, Gv) [Raichet al., 1991' McGuire et al., 1993' cause of the effects of f(NA) on the GPP calculation;f(NA)
models the limiting effectsof plant nitrogen statuson GPP
Melillo et al., 1993]:
throughthe feedbackof nitrogenavailabilityon carbonassimif(Ca, Gv)= C/(kc + Ci)
lation [McGuire et al., 1992, 1993; Melillo et al., 1993]. This
machineryof photosynthesis.
At high levelsof intercellular
CO2,the enzymatically
controlledrate of carbohydrate
synthesis, which affectsthe phosphate
regeneration
that is necessary
for harvesting
lightenergy,mayregulatethe fixationof carbon.
where Ci is the concentration
of CO2 within leavesof the can-
opy andkc is the half-saturation
constantfor CO2 uptakeby
plants.The variableC, is theproductof ambientCO2(Ca)and
relativecanopyconductance
to CO2(Gv), a variablewhichincreasesfrom 0 to 1 with increasingwater availabilityand depends on the ratio of actual evapotranspiration
to potential
evapotranspiration:
Gv= -10(EET/PET)
2+ 2.9(EET/PET)
EET/PET
< 0.1
feedbackis dynamicallydeterminedby comparingthe calculation of GPP basedon nitrogensupplyand the calculationof
GPP for no constraints
of nitrogensupply. Nitrogensupplyis
the sum of nitrogenuptake((NUPTAKE), see Figure 1) plus
nitrogen mobilized from the vegetationlabile nitrogen pool
((NMOBIL), seeFigure 1). The C to N ratio of production,
which is represented
by the parameterPcN,is multipliedby the
sum of NUPTAKE
and NMOBIL
to determine
the amount of
NPP that can be supportedfrom the nitrogensupply. Nitrogen
Gv = 0.1 + 0.9 (EET/PET)
EET/PET > 0.1
supplyconstrainsproductionwhen the calculationof unconwhereEET is estimatedevapotranspiration
andPET is potential strainedGPP, that is, potentialGPP (GPPp)for f(NA) equalto
1, exceedsthe sumof autotrophic
respiration(R^) andNPP that
evapotranspiration.The differentform of Gv below EET/PET
is
determined
from
nitrogen
supply.
Thereforethe feedbackof
of 0.1 allows minimumpossibleGv to be 0 insteadof 0.1, a
nitrogen
availability
on
carbon
assimilation,
f(NA), is the ratio
plant responsethat seemspossiblein extremelyarid regions.
of GPP to GPPp. The calculationof GPP, NPP, and f(NA) can
BecauseGv dependson the ratio of EET to PET, the response
be expressed
as
of f(Ca,Gv) to doubledCO2is higherin dry environments.The
NPP = PcN(NUPTAKE+NMOBIL)
value of the parameterko 400 ppmv, has been chosento increasef(Ca, Gv) by 37% for a doublingof atmosphericCO2
GPP = Pcs (NUPTAKE+NMOBIL) + R^
from 340 ppmv to 680 ppmv for maximumrelative canopy
conductance,that is, Gv = 1 [McGuire et al., 1992, 1993].
f(NA) = GPP/GPPp PCN(NUPTAKE+NMOBIL) + R^ _<GPPp
Althoughthe rangeof responseof plant growthis between
and
25 and50% in studiesthatprovideadequatenutrientsandwater
to experimentalplants[Kimball, 1975; Gates, 1985], soil nitroNPP = GPPp- R A
gen is an importantfactorthat often constrains
the responseof
GPP = GPPp
woodyplant growthto elevatedCO2 [McGuire et al., 1995a].
The constraining
effectsof nitrogenavailabilityon the response
f(NA) = 1
PCN(NUFFAKE + NMOBIL) + RA > GPPp
of carbonassimilationto elevatedCO2 represents,
in part, the
acclimationresponseof photosynthesis
to elevatedCO2. BeOn first inspectionthesealgorithmswould appearto comcauseruNsco representsa substantialproportionof leaf nitropletelyconstrainthe responseof NPP to elevatedCO2 in nitrogen [Evans, 1989], photosynthetic
rate is generallycorrelated gen-limitedsystems.However,it is importantto recognizethat
with the nitrogencontentof leaves[Evans, 1989' Field, 1991].
thereis seasonalityin the degreeof nitrogenlimitation. NitroReducednitrogenavailabilityhas often been observedto degen is generallyin greatestsupplyearly in the growingseason
creaseboth leaf nitrogencontentand photosynthesis
[Wong, when vegetationis able to mobilizenitrogenfrom storage. In
1979; Gulmonand Chu, 1981' Evans, 1983; Sageand Pearcy, this case,the vegetationin TEM is able to incorporateelevated
1987a,b; Chapin et al., 1988' Lajtha and WhiOrord,1989].
intercellularCO2 into production.Higher levelsof production
Lower nitrogenconcentrations
of leavesin responseto elevated causegreaterlitterfall to causehigher ratesof decomposition
CO2 may representthe dilutionof nitrogenby higherratesof
and higher rates of nitrogen cycling. One consequenceof
carbonassimilation
or mayrepresentlessallocationof nitrogen greaternitrogencyclingis higherratesof nitrogenuptake. Nito rubiscoand other enzymesthat influencelight harvestand trogencyclingeventuallyequilibratesat a higher level consisphosphate
regeneration.If lower concentrations
primarilyrep- tent with the higherlevelsof productionand nitrogensupply.
resentdilution, then assimilationper unit leaf area shouldreThus elevated CO2 alters the seasonalpattern of carbon-
MCGUIRE ET AL.: CO2, VEGETATION
[N], AND GLOBAL CARBON STORAGE
177
nitrogenstatusin the vegetationof TEM to influenceproduc-
relationships
[seeMelillo et al., 1982]because
thedatausedto
tion.
derivetherelationship
represents
a widerrangeof ligninto nitrogenratiosthandatausedto deriveotherrelationships;
in
TEM,theaggregated
nature
of litterinputs
intothesoilrequires
a relationship
appropriate
to a broadgradient
of litterquality.
Our implementation
of thisrelationship
to determine
the dependence
of Kdonlitterqualityassumes
thatKdisproportional
Decomposition
The response
of decomposition
to elevatedatmospheric
CO2
may directlyinfluencesoil carbonstoragethrougheffectson
litter qualitythat alterheterotrophic
respirationrates. Decompositionmay alsoinfluencesoil carbonstoragethrougheffects to k [see Raich et al., 1991] and that the ratio of k to
-ø'784
isaconstant
[see
Melillo
etal.,1982].Fion inorganicnitrogenavailabilitythatalterNPP andinputsinto (lignin/nitrogen)
tonitrogen
ratioof litterfall
in placeof
the soil;inorganicnitrogenavailabilitydepends,in part, on the nally,weusethecarbon
net nitrogenmineralization((NETNMIN), seeFigure 1) of soil theinitialligninto nitrogenratioof litterfall.
organicnitrogenthat is associated
with decomposition.In
TEM, decomposition
is represented
asheterotrophic
respiration Designof SensitivityAnalyses
(RH):
To examinesensitivity
in the globalresponses
of NPP and
carbonstorageto decreases
in the nitrogenconcentration
of
whereKdis theheterotrophic
respiration
rateat 0øC, Csis car- vegetation,
wecompare
thereference
doubled
CO2responses
of
bonstoragein soils,f(M) is a functiondefiningtheinfluenceof TEM to the doubledCO2responses
of TEM for simulations
in
soilmoisture(M) on decomposition,
andT is meanmonthlyair whichthevegetation
nitrogenconcentration
wasreducedwithtemperature.In TEM, RH is the only losscalculatedfrom the out influencingdecomposition
dynamics("lower N" simuladetritalcompartment
Cs,whichis an aggregated
poolof organic tions)andsimulations
in whichreductions
in vegetation
nitrocarbonin detritusand soils.The functionf(M) is a nonlinear gen concentration
influencedecomposition
dynamics("lower
RH= KdCsf(M) eø'ø693T
relationshipthat modelsthe influenceof soil moistureon microbialactivity at low soil-moisturecontentsand the influence
of oxygenavailabilityon microbialactivityat. high moisture
contents[seeRaich et al., 1991]). This relationshipcausesthe
highestratesof decompositionto occurwhen soilsare 60% to
80% saturatedwith water, which has been observed in numer-
N+D" simulations). We conductedthree lower N simulations
and three lower N+D simulations in which we reduced the ni-
trogenconcentration
of vegetationby 7.5, 15.0, and22.5%. We
choseincremental
reductions
of 7.5% for thesensitivity
analysesbecause
thereductionis half of the average15% reduction
of nitrogenconcentration
in response
to an approximate
dou-
ous laboratoryand field studies[Bartholomewand Norman,
1946; Bhaumikand Clark, 1947; Miller and Johnson,1964; Ino
and Monsi, 1969; Hunt, 1977; Davidson, 1979' Sommerset al.,
1981' Van Veenand Paul, 1981' Stottet al., 1986]. The exponentialfunctionwith T represents
the temperaturesensitivityof
decomposition,
which increaseslogarithmicallywith a Q•0 of
2.0 over all temperatures;
soil respirationin temperateforests
soilshas a Q•0 of 1.988 in relationshipswith mean daily air
temperatureand 1.983 in relationshipswith mean monthlyair
temperature
[Kicklighteret al., 1994].
The parameterKd, which represents
the heterotrophic
respi-
blingof atmospheric
CO2[seeMcGuireet al., 1995a].
In modelingthe acclimationresponseof GPP to elevated
centration,we relateKdto a powerfunctionof the carbonto nitrogenratio of litterfall:
munity[Shaverand Chapin,1991]. The differencebetweenthe
CO2,the assumption
in TEM is that at monthlyand annual
timescales,
thereis a balancebetweencanopycarbonassimilationandnitrogeneconomyof thevegetation.Thusin TEM, nitrogenis allocatedimplicitlyto represent
the tradeoffbetween
canopy
development
andacclimation
of tissue-level
photosynthesissothatcarbonuptakeis maximized
in buildingvegetative
biomass
at a specificcarbonto nitrogen
ratio. Althoughvegetativebiomass
is builtat a specificcarbonto nitrogenratio,the
overallratioof carbonto nitrogenin vegetation
maybe differrationrateat 0øC, is therate-limiting
parameter
in theRHfor- ent. For example,in tundra,thecarbonto nitrogenratioof new
mulation. The value of Kd at a vegetation-specific
calibration production
is around30 whiletheoverallcarbonto nitrogen
rasiteI•2dc
is one of severalrate-limitingparametersthat are deter- tio is around50 [Shaverand Chapin 1991; McGuire et al.,
minedby calibratingTEM to the annualfluxesandpoolsat the 1992]. Theseratiosarerepresentative
for theaggregated
vegecalibrationsitefor an ecosystem.To implementchangesin lit- tationof a plantcommunity,
eventhoughtheratiosmaybedifter qualityassociated
with changesin vegetationnitrogencon- ferentfor individualspecies
andgrowthformswithinthecom-
md-- md
c(Lc/LN)
-0'784
/ (Ccc/CNc)
'0'784
where Lc and LN are the annualfluxes of litterfall carbonand
nitrogen(seeFigure1) andLccandLN•aretheannualfluxesof
litterfallcarbonandnitrogenat thecalibration
sitefor the ecosystem.The implementation
of thispowerfunctionis basedon
therelationship
derivedby Melilloet al. [ 1982]for thedecom-
position
of 13leafandneedle
species
in thelaboratory
studyof
Daubenmire
and Prusso[1963]. The relationship
identifies
thatan inversecurvilinear
relationship
existsbetweenthe rate
constant
for annualmassloss,k [seeJennyet al., 1949],andthe
initialligninto nitrogen
ratioaccording
a powerfunctionwith
theexponent
-0.784.Weusedthisrelationship
instead
of other
carbonto nitrogen
ratiosof newproduction
andoverallvegetativebiomass
arecaused
by theprocesses
of nitrogen
resorption
andmobilization
in plants.In TEM, nitrogen
resorption
is represented
by the flux NRESORB and nitrogenmobilizationis
represented
by theflux NMOBIL (seeFigure1).
Patterns
of nitrogenresorption
fromsenescing
tissueandof
mobilization
of nitrogenfrom storagesuggest
thatplantsallocatenitrogen
to maximizecarbonuptake.Thenitrogenthatappearsin new vegetative
biomass
comesfromnew uptakeand
fromrecycling,
in whichnitrogenis mobilizedfromstorage.
The resorption
of nitrogenfromsenescing
tissues
is primarily
responsible
forthenitrogen
in storage.In general,
thedegree
of
recyclingis sensitive
to the degreeof nitrogenlimitation.For
example,
production
in tundraplantsis morelimitedby nitro-
178
MCGUIREET AL.:CO2,VEGETATION[N],AND GLOBALCARBONSTORAGE
gen availabilitythan is productionin boreal forest, which is
more limited than productionin temperateconifer forest [see
McGuire et al., 1992]. In tundraplants,50 to 80% of nitrogen
in new tissuecomesfrom resorption[Shaverand Chapin,
1991]. This is greaterthan the degreeof recyclingin boreal
forest,in whichapproximately45% of the nitrogenin new biomasscomesfrom resorption[seeMcGuire et al., 1992]. In
temperateconifer forest, the degreeof recyclingis less, with
approximately30% of nitrogenin new biomassderivedfrom
resorption[Sollinset al., 1980], andin tall grassland
only 10%
of the nitrogenin new biomassis derived from resorption
[Risseret al., 1981]. Thesestudiesof nitrogenrecyclingindicate that plants are sensitiveto carbon-nitrogen
balanceand
suggestthatplantsare capableof conservingand allocatingnitrogento maximizecarbonuptake. Becausethe parametersrelated to the nitrogen concentrationof vegetationbiomassin
TEM define carbon-nitrogen
balancein the model,it is important to understand
the sensitivityof CO2 responses
of TEM to
potentialchanges
thattheseparameters
mayexperience
in associationwith elevatedatmospheric
CO2.
Experimentalstudiesthathavemeasuredthe response
of tissue nitrogenconcentration
in plantsexposedto elevatedCO2
usually do not identify whetherthe measurements
represent
changesin new productionor in overallvegetativebiomass.
monthlypercentsunshinedurationin the Cramerand Leemans
database.Thevegetation
dataset,whichidentifies18 dryland
ecosystems,
is similarto that of Melillo et al. [1993], but has
beenmodified
to represent
theterrestrial
boundaries
definedby
the CramerandLeemansglobaldatabase.The elevationdata
represent
an aggregation
to 0.5ø resolution
of the 10' National
Centerfor Atmospheric
Research(NCAR)/Navy[1984] data.
Soiltexture
is basedontheFoodandAgriculture
Organization
(FAO)/ComplexSystemsResearchCenter (CSRC) (undated)
digitization
of the FAO-United
NationsEducational,
Scientific,
andCulturalOrganization
(UNESCO)[1971]soilsmap. Hydrologicaldatafor TEM are determined
by a waterbalance
model[Vorosmarty
et al., 1989]thatusesthesameinputvariables as TEM.
The applicationof TEM to a grid cell requiresthe useof the
monthlyclimaticandhydrologicaldataand the soil- and vegetation-specific
parameters
appropriate
to the grid cell. Although
many of the vegetation-specific
parametersare defined from
publishedinformation,someare determinedby calibratingthe
model to the fluxes and pools sizesof an intensivelystudied
field site.The run for eachgrid cell startswith the Januaryvalues of the variablesthat have monthly temporalresolution.
Thereforethe initial valuesof the pool sizesfor a grid cell are
setto thoseof the Decembervaluesfrom the appropriatevegeTherefore, in both the lower N and the lower N+D simulations, tation-specific
calibration.To determinea solutionfor baseline
cliwe alteredthe parametersin TEM that controlthe vegetation conditions,which in this studyis definedas contemporary
mate at an atmospheric
CO2 concentration
of 340 ppmv, the
carbonto nitrogenratio (VcN, [seeRaich et al., 1991]) and the
modelis run with an opennitrogencycle,andnitrogenis annuproductioncarbonto nitrogenratio (PcN,[seeMcGuire et al.,
ally importedor exportedthroughthe inorganicnitrogenpool
1992]). Becausenitrogenconcentration
is inverselyrelatedto
dependingon whethersoil organicmatteris nitrogenpoor or
carbonto nitrogenratio, we modifiedtheseparametersas follows:
rich in comparison
to the carbonto nitrogenratioof the soil at
the
calibration
site,
SCN. This algorithmsimulatesthe balance
VcN(d[N]) = 100 VCNo/ (100 - d[N])
betweenlong-termnitrogeninputsand outputsso that the grid
and
cell has an equilibriumsoil carbonto nitrogenratio equal to
SCN;it hasthe benefitof reachingequilibriummuchfasterthan
PcN(d[N])= 100 PCNo
/ (100 - d[N])
explicitlysimulatingnitrogenfixation,nitrogendeposition,denitrification,and nitrogenleachinglosses.The grid cell is dewhereVCNoandPcNoarethe originalvaluesof VCNandPCN,and
VcN(d[N]) and PcN(d[N]) are the valuesof VCNand PeNassoci- terminedto havereachedequilibriumwhenthe annualfluxesof
NPP,Lc,andRi•differby lessthan1 g C m'2yr-•,those
of
atedwith the percentdecreasein nitrogenconcentration
(d[N]);
therefore7.5, 15.0, and 22.5% decreases
in nitrogenconcentra- NETNMIN, LN, andtotal N uptakeby vegetationdiffer by less
ofNINPUTandNLOSTdiffer
tion increaseVCNoand PCNoby 8.1, 17.6, and 29.0%, respec- than0.01gN m-2yr-•,andthose
bylessthan0.01g N m'2yr-•. Because
NINPUTandNLOST
tively. In the referenceand lower N simulations,we set K•
are determinedannually,thesefluxes are effectively0 at equiequal to Kdcso that decompositionis not influencedby litter
quality. In the lower N+D simulations,we calculatedKd based librium.
For runninga grid cell underdoubledCO2 (680 ppmv), the
on its relationshipwith litter quality,whichis influencedby the
initial valuesof the poolsfor the grid cell are setto the DecemparameterVcN(d[N]).
ber valuesof the equilibriumbaselinesolutionfor the grid cell,
that is, the solutionfor 340 ppmv CO2. The nitrogencycleis
Application of the Terrestrial EcosystemModel
closedfor the doubledCO2 run so that thereis no nitrogenimThe data setsusedto drive TEM are griddedat a resolution portedor exportedfrom the grid cell. The conditionsfor deterof 0.5ølatitudeby 0.5ølongitude.Thevariables
required
to run miningwhenthe grid cell hasreachedequilibriumarethe same
TEM 4.0 for a grid cell are meanmonthlytemperature,monthly as for the baseline solution. For the reference doubled CO2
precipitation,mean monthlycloudiness,vegetation,elevation, simulation,the mean amountof time for a grid cell to reach
soil textureas percentsilt plus clay, and severalhydrological equilibriumwas76.6 years. For the lower N simulationsassovariables(potentialevapotranspiration,
estimatedevapotrans- ciatedwith 7.5, 15.0, and 22.5 reductionsin vegetationnitrogen
concentration,the mean amountof time for a grid cell to reach
piration,and soil moisture). Mean monthlytemperature,precipitation,and cloudinessin this studyare from the globaldata
equilibriumwas 116.7, 144.3, and 169.2 years,respectively.
setsof CramerandLeeroans(W. Cramer,personalcommunica- For the lower N+D simulations associatedwith 7.5, 15.0, and
tion, 1995), which is a major updateof Leemansand Cramer
22.5% reductionsin vegetationnitrogen concentration,the
[1991]. Monthly percentcloudinessis calculatedas 100 minus mean amountof time for a grid cell to reachequilibriumwas
MCGUIRE ET AL.: CO2,VEGETATION [N], AND GLOBAL CARBON STORAGE
157.8, 192.5, and 219.9 years,respectively.
Becausethe turnovertime of the single-compartment
soilcarbonpool in TEM is
approximately16 years,the simulations
in this studyrepresent
true equilibriumsolutionsby the model. Althoughsoil carbon
is represented
as a singlecompartment,
the equilibriumsolution
is representativeof the solutionin a multiple-compartment
modelthatconsidersthe sametotalpool of soil organiccarbon
with similarformulationsfor eachcompartment.The implementationof the single-compartment
modelhas the advantage
of determiningan equilibriumsolutionin less computational
time thanwould be requiredby a multiple-compartment
model
(seeSchimelet al. [1994] for analysesof single-versusmultipie-compartment
modelsof soil carbondynamics).
Results
BaselineEstimates of NPP and Carbon Storage
At an atmospheric
CO2concentration
of 340 ppmv,TEM 4.0
estimates
globalterrestrial
NPPto be49.010•sg (Pg)C yr-•
(Table 1). Much of the globalNPP occursin equatorialregions
(Figure2a) with 34.5% of globalNPP in tropicalevergreenforest, which occupies13.7% of the terrestrialarea (Table 1). In
contrast,temperateforests(coniferous,deciduous,mixed, and
broadleafevergreen),which occupy 11.3% of the terrestrial
area, accountfor 19.2% of global NPP (Table 1). A troughin
productivity
occurs
between
latitudes
10øand20øN (Figure2a).
These latitudesare largelyoccupiedby desertsand arid shrublands,which accountfor only 3.9% of globalNPP although
they occupy20.2% of global land area (Table 1). Similarly,
high-latitudeecosystems
(polardesert,tundra,borealwoodland,
and borealforest),which occupy22.6% of terrestrialarea,accountfor only 9.6% of globalNPP (Table 1). The leastproductivevegetationtypesincludepolardesert,tundra,and desert,
which collectivelyaccountfor 2.4% of terrestrialNPP (Table
1).
For the baselineatmosphericCO2 concentration,
globalterrestrialcarbonstorageis estimatedby TEM 4.0 to be 1701.8Pg
C, with 771.7 Pg C in soilsand930.1 Pg C in vegetation(Table
1). The estimateof vegetationcarbonstoragedoesnot include
the effectsof landuse. In previousstudies,TEM estimatesthat
landusereducesvegetationcarbonstocksbetween150 and 200
Pg C (D. W. Kicklighter,unpublisheddata,1994). Becausesoil
organicmatteris composedof materialthat represents
a spectrum of turnovertimesalongthe decaycontinuum[seeMelillo
et al., 1989a], we have designedTEM to excludefrom its estimatesof soil carbonstoragethat portionof soil organicmatter
whichis biologicallyunreactivein the contextof globalchange
thatmightoccuroverthenextcenturyor so. In previousstudies
with TEM, we have estimatedthat the biologicallyunreactive
soilorganiccarbonin thiscontextrepresents
400 to 500 Pg C of
the globalsoil carboninventory[Melillo et al., 1995]. Several
studiesshedlight on the spectrumof tumovertimesrepresented
in soil carbon estimates of TEM.
The terrestrial model of Sar-
mientoet al. [1995] estimatesthat approximately350 Pg C of
soil organiccarbonhasa turnovertime rangingbetween1 and
20 years. Harrison et al. [1993] indicatethat 50% of the soil
column containsfast-cyclingcarbon and estimatethe global
fast-cyclingpool to be 500 Pg C with components
that have
residenttimes rangingfrom 10 to 100 years. Trumboreet al.
[1996] demonstrate
fast rumover(7 to 65 years)for 50 to 90%
179
of carbon
in theupper20 cmof soil.In theupper20 cmof soil,
theCenturymodelestimates
thatapproximately
10%of thesoil
organicpool is represented
in the soil microbialand detrital
poolswith0- to 10-yearturnover,
approximately
50%is represented
in a slowpoolwith 20- to 100-yearturnover,
andapproximately
40% is represented
in a passive
poolwith 1000-to
5000-yearrumover[Schimelet al., 1994]. On the basisof the
837.2PgC estimated
by thethirdversion
of TEM for organic
carbonstocksin theupper20 cm of soil [Melilloet al., 1995]
and the 60% estimateof soil carbonwith 0- to 100-year turnover, 502 Pg C of global soil carbonstockshas 0- to 100-year
turnover,which agreescloselywith the 500 Pg C estimateof
Harrison et al. [1993]. On the basisof theseestimates,we in-
terpretthat the 771.7 Pg C pool estimatedby TEM represents
soilcarbonwith turnovertimesminimallyin the rangefrom 0 to
100 yearsand perhapsto severalhundredyears,but that the
poolexcludessoilcarbonwith millenialrumover. Thereforewe
feel that TEM minimallycharacterizes
the decadalresponseof
soil carbonup to about a century,which is an appropriate
timescalefor consideringpotentialterrestrialresponses
to doubled atmospheric
CO2. The modelmay characterize
soil carbon
responses
beyonda century,but the timescaledependson what
fractionof soil carbonwith 100- to 1000-yearturnoveris representedby the model;soil reservoirswith 400-yearturnovermay
startto dominatesoil carbonresponsesto rising atmospheric
CO2afterabout100 years[Sarmientoet al., 1995].
In contrastto NPP, globalcarbonstoragehas a bimodallatitudinaldistribution,with substantialcarbonstoragein both the
equatorialand boreal regions (Figure 2b). Vegetation carbon
accountsfor most of the storagein the tropics (Figure 2b);
74.5% of the carbonin tropicalevergreenforest is storedin
vegetation(Table 1). In contrast,soil carbonaccountsfor more
thanhalf of the storagein the borealregion(Figure2b); 53.9%
of the carbonin boreal forest is storedin soils (Table 1). Becausehighertemperaturecausessoil organicmatterto decom-
poseat a higherrate,temperature
is the mostimportantvariable
determiningthe latitudinalpattem of the contributionof soil
carbonto total carbonstorage,althoughsoil moisturealsoplays
a role [seeMcGuire et al., 1995b].
ReferenceResponsesof NPP and Carbon Storage
In response
to doubledatmospheric
CO2,TEM estimates
that
global
NPPincreases
4.1PgC yr-•(Table
2),which
isa relative
increaseof 8.3%. Much of the NPP responseis concentrated
in
the tropicswheretropicalevergreenforestaccountsfor 32.8%
of the globalresponse(Figure3a). The latitudinaldistribution
of responsebroadlyextendsthroughoutthe northernand southem temperateregions(Figure 3a), where temperateforestsaccountfor 14.3% of the response
(Table2). Althoughthe relative
NPP responses
of deserts(36.0% increase)and add shrublands
(24.3% increase)are high, theseecosystemsaccountfor only
12.5% of the global response(Table 2) becauseof their low
productivity.The latitudinaldistributionof NPP response
drops
off substantiallybetween the temperate and polar regions
(Figure 3a) where high-latitudeecosystemsaccountfor only
3.5% of the globalresponse(Table2).
For doubledatmosphericCO2, the estimateof global total
carbonstorageby TEM increases114.2 Pg C (Table 2). Soil
carbonaccountsfor 41.7% of the increasedcarbonstorage,and
vegetationcarbonaccountsfor 58.3% (Table2). The latitudinal
180
MCGUIREET AL.: CO2,VEGETATION[N], AND GLOBALCARBONSTORAGE
-o
o
o
o
o
.,•
o
o
o
o•
-o o
.,•
o
.,•
.,.•
o
ß•,
o
MCGUIRE ET AL.' CO2,VEGETATION [N], AND GLOBAL CARBON STORAGE
(Figures4b and 4c). For each7.5% reductionin nitrogenconcentration,the responseof vegetationcarbon increasesapproximatelyan additional40 Pg C and soil carbonan additional
30 Pg C. The responses
of vegetationand soil carboncombine
so that total carbonstorageincreasesapproximatelyan additional 70 Pg C for each 7.5% reductionin nitrogenconcentration (Figure4d).
For simulationsin which lower vegetationnitrogenconcentration is coupledwith decompositiondynamics(lower N+D),
the responseof NPP is relativelyinsensitiveto differencesin
the reductionof nitrogenconcentration
(Figure4a). The NPP
600
(a)
50O
4O0
30O
20O
100
0
-60-50-40-30-20-10
181
0
10 20
30
40
50
60
70
80
90
increase
is5.0PgC yr'• fora 22.5%reduction
compared
witha
4.1PgC yr'• increase
fornoreduction.
Theinsensitivity
in the
coupledsimulationsoccursbecausepotentialenhancements
in
NPP associatedwith reducedvegetationnitrogenconcentration
14000
are approximatelyoffset by lower nitrogenavailabilityassoci12000 atedwith thedecomposition
dynamicsof reducedlitter nitrogen
concentration. Becauseof this insensitivity,the responseof
10000 vegetationcarbon storageis also insensitiveto reductionsin
8000 vegetationnitrogenconcentration(Figure4b). For a 22.5% re6oooductionin nitrogenconcentration,vegetationcarbonincreases
4000 83.3 Pg C comparedwith 66.6 Pg C for no changein nitrogen
2000 concentration.
In contrast,theresponse
of soilcarbonstorageis
0
larger for greaterreductionsin nitrogenconcentration(Figure
4c). Lower nitrogenconcentrationof litter input to the soil
-60-50-40-30-20-10
0 10 20 30 40 50 60 70 80 90
slowsdecomposition
so that for each7.5% reductionin vegetation nitrogenconcentration,
soil carbonincreasesapproximately
Latitude
an additional60 Pg C (Figure4c). The responses
of vegetation
Figure 2. Latitudinal distributionsfor the baselineestimates and soil carbonstoragecombineso that total carbonstorageinof (a) net primary productionand (b) carbon storageby the creasesapproximately65 Pg C for each 7.5% reductionin niTerrestrial Ecosystem Model at 340 ppmv CO2 and trogen concentration(Figure 4d). Thus, for simulationsin
contemporaryclimate. Estimates of soil carbon do not whichreductionsin nitrogenconcentration
are coupledwith deinclude biologically unreactive soil organic matter. The
compositiondynamics,the global sensitivityof total carbon
resolution
of NPP andcarbonstorage
is 0.5olatitude.
storageto reductionsin nitrogenconcentration
is similarto that
for simulationsin whichreductionsareuncoupledfrom decompositiondynamics. Althoughthe global sensitivityis similar,
distributionof the total carbonresponseis more bimodalthan the responseof soil carbonis more importantin the coupled
theNPP responsewith peaksin the tropicsnearthe equatorand simulations.This has implicationsfor the latitudinaldistribuin the northtemperatezone (Figure3b). The peak in the equa- tionsof carbonstorageresponses.
torial tropicsis causedprimarilyby a peak in the responseof
For the simulationsin which changesin vegetationnitrogen
vegetationcarbon,which accountsfor 75.0% of the increased
concentration
areuncoupledfrom decomposition
dynamics,the
carbonstoragein tropicalevergreenforest. The peak in the latitudinaldistributionof NPP responseis sensitiveto reducnorth temperatezone is causedby peaksin the responses
of
tions in nitrogenconcentrationthroughoutthe terrestrialbiobothvegetationandsoilcarbon.In contrast,in high-latitude
re- sphere(Figure5a) but is insensitive
for the coupledsimulations
gions,the responseof soil carbonincreasinglyaccountsfor (Figure 5b). Thus, in the coupled simulations,potential enmore of the total carbonresponseas latitudeapproaches
the hancements
in NPP associated
with reducedvegetationnitrogen
poles(Figure3b). For example,soil carbonaccountsfor 49.3%
concentrationare approximatelyoffsetby lower nitrogenavailof theresponse
in borealforest,70.8% of theresponse
in boreal ability associatedwith the decomposition
dynamicsof reduced
woodland,and 92.6% of the response
in tundraandpolardes- litter quality. The latitudinalsensitivityof NPP responsecauses
ert.
a similarpatternin the responseof vegetationcarbonwherethe
latitudinaldistributionis sensitiveto the reductionof nitrogen
Sensitivityof NPP and Carbon StorageResponses
concentrationin the uncoupledsimulations(Figure 6a) but is
to Changesin Plant Nitrogen Concentration
insensitivein the coupledsimulations(Figure6b). In contrast,
For simulationsin which lower vegetationnitrogenconcen- the latitudinaldistributionof soil carbonresponseis sensitiveto
trationis uncoupledfrom decomposition
dynamics(lower N),
reductionsin nitrogen concentrationfor both the uncoupled
simulations(Figure7a) andthe coupledsimulations(Figure7b).
the responseof globalNPP to doubledatmospheric
CO2 increases
approximately
2 PgC yr-1foreachincremental
7.5%re- Throughoutthe terrestrialbiosphere,the responseof soil carductionin vegetationnitrogenconcentration
(Figure4a). The bon is moresensitivefor the coupledsimulationsbecauseof the
dynamicsassociated
with lower litter quality. In
responseof NPP causeslargerincreasesin both vegetationand decomposition
soil carbon for greaterreductionsin nitrogen concentration both setsof simulations,the responseof soil carbonis substan16000
-
--Total
(b)
.
182
MCGUIREET AL.: CO2,VEGETATION[N], AND GLOBALCARBONSTORAGE
¸
¸
¸
¸
¸
¸
¸
¸
.,_•
¸
¸
¸
..,•
MCGUIRE ET AL.: CO2,VEGETATION [N], AND GLOBALCARBONSTORAGE
50-
(a)
•,
40-
•
30-
•
2o-
o'-
10-
•
0
-60-50-40-30-20-10
1400
-
•'
1200
-
m
1000
-
""
800
-
0
600 -
•
400
-
•:
200
-
o1
.c:
0
10 20 30 40 50 60 70 80 90
(b)
......
Total
0
-200
-60-50-40-30-20-10
0
10 20 30 40 50 60 70 80 90
Latitude
Figure 3. Latitudinaldistributions
for the referenceresponse
of (a) net primaryproduction(NPP) and (b) carbonstorage
estimatedby the TerrestrialEcosystemModel for an increase
in atmospheric
CO2from 340 to 680 ppmvwith no changein
climate from contemporary
and no changein vegetation
nitrogen concentration. The resolutionof NPP and carbon
storage
responses
is 0.5ølatitude.
183
The global NPP estimatealso differs from the 53.2 Pg C reportedby Melillo et al. [1993], which useda differentversion
of TEM (version3), differentinputdatasets,differentterrestrial
boundariesassociated
with the input data sets,and a different
levelof baselineCO2. Althoughthe globaltotalsof the two applicationsdiffer, botharewithinthe rangeof estimatesreported
in the literature,and both applicationsreport similar distributions of NPP amongterrestrialecosystems.For example,the
34.5% of globalNPP in tropicalevergreenforestin the baseline
simulationof this studyis similar to the 35.9% reportedby
Melillo et al. [1993]. Similarly,2.4% of globalNPP in the three
leastproductiveecosystems
is similarto the 3.0% reportedby
Melillo et al. [1993]. The sensitivityof annualNPP estimates
by TEM 4.0 to differentalternativeinput data setsof temperature, precipitation,solar radiation, and soil texture has been
evaluatedby Pan et al. [1996].
Similarto NPP, the baselineestimates
of carbonstorageare
slightlyhigher than thosereportedby Xiao et al. [1997], in
whichTEM 4.0 estimated
globalterrestrialcarbonstorageto be
1659Pg C, with 909 Pg C in vegetationand750 Pg C in soils.
The estimateof soil carbonstorageis alsodifferentfrom that
reportedby McGuire et al. [1995b], in which TEM 4.0 estimatedglobalsoilcarbonstorageof 706.5 Pg C. The application of TEM by McGuire et al. [1995b] useddifferentinput
data sets, different terrestrial boundaries associatedwith the in-
put data sets, and a different level of baselineCO2 (312.5
ppmv). Differencesbetweenthe globalestimates
of vegetation
carbonstoragein thisstudy(930.1 Pg C) andthe 977 Pg C reportedby the applicationof version3 of TEM in Melillo et al.
[1995] are largelyrelatedto differencesin NPP estimates.Differencesin soil carbonstoragebetweenversions3 and 4 of
TEM largelydependon whetherunreactivesoil organicmatter
has been included in the estimates[see McGuire et al., 1995b;
Melillo et al., 1995].
The ReferenceResponsesof NPP and Carbon Storage
tiallymoresensitive
in temperate
andborealregionsthanin the
tropics.The combinedresponses
of vegetationand soil carbon
indicatethattheresponse
of totalcarbonstoragein bothsetsof
simulations
is bimodallydistributed
betweenthetropicsandthe
northtemperate-boreal
region(Figures8a and 8b). As the reductionin vegetation
nitrogenconcentration
getsgreaterin the
coupledsimulations,
more of the additionalcarbonstorage
tendsto becomeconcentrated
in thenorthtemperate-boreal
region in comparison
to the tropics(Figure8b). Also, the additionalcarbonstorage
in thenorthtemperate-boreal
regiontends
to shiftmorenorthward
for greaterreductions
in nitrogenconcentration.
Discussion
BaselineEstimatesof NPP and CarbonStorage
The estimate
of globalannualNPPby TEM 4.0 in thisstudy
for baselineatmospheric
CO2(49.0Pg C) is similarto manyof
theestimates
thathaveappeared
in theliterature(mean53.1Pg
C; N = 13;range40.5 Pg C to 78.0 Pg C; standard
deviation9.3
Pg C [seeMelillo et al., 1993]). Thisestimateis slightlyhigher
thanthe47.9PgC yr-• reported
byXiaoetal. [1997]in which
TEM 4.0 wasappliedwiththe sameclimate,soils,andvegetationinputsat an atmospheric
CO2concentration
of 315 ppmv.
The 8.3% increaseof globalNPP in the referencesimulation
of doubledCO2response
is similarin magnitudeto the 9% responseof TEM 4.0, the 11% responseof Biome-BGC (for BiomeBioGeoChemical
Cycles),andthe 5% response
of Century
in equilibriumdoubledCO2 simulationsfor the conterminous
United States in the Vegetation/Ecosystem
Modeling and
AnalysisProject(VEMAP) [seeVEMAP Members,1995]. All
of theseresponses
are substantiallylower than the 25 to 50%
responseobservedamongstudiesthat provideadequatenutrients and water to experimentalplants [Kimball, 1975; Gates,
1985]. Althoughthe modelstendto agreeon the magnitudeof
continental-scale
NPP responses
to doubledatmosphericCO2,
differentfactorsconstrainthe responseof NPP [Pan et al.,
1995]. All the modelsagreethat relativeNPP tendsto increase
alonggradientsof decreasing
precipitation.However,along
gradientsof decreasing
temperature,estimatesof relativeNPP
responses
tend to decreasefor TEM, increasefor Biome-BGC,
and haveno patternfor Century[Pan et al., 1995]. For TEM
simulations,the spatialpatternsand magnitudeof NPP responses
in the referencesimulationsof doubledCO2 response
representinteractions
of carbon,water,and nitrogendynamics
in theformulations
of ecosystem
processes
in themodel.
The magnitudeof the NPP responseis limited in the TEM
simulationsbecausethe availabilityof nitrogenconstrains
the
184
MCGUIRE ET AL.' CO2,VEGETATION [N], AND GLOBAL CARBON STORAGE
12"/'22]Reference
20-
•
(a)
•
-7.5%
[N]
-15%[N]
I==:::] -22.5% [N]
15-
0
500
o•
(b)
e 400
o
•
300
10-
o• 200
5-
•
o
LowerN LowerN+D
• 100
.c:
0
(.3
LowerN LowerN+D
500
500
(c)
t(d)
400
400
300
300
200
200
100
100
0
0
Lower N Lower N+D
0
Lower N Lower N+D
Figure 4. Sensitivityto 7.5, 15.0, and 22.5% decreases
in vegetationnitrogenconcentrationfor the global
responses
of (a) net primaryproduction(NPP), (b) vegetationcarbonstorage,(c) soil carbonstorage,and (d)
total carbonstorageestimatedby the TerrestrialEcosystemModel for a changein atmospheric
CO2 from 340
to 680 ppmv with no changein climate from contemporary.Sensitivityis shownfor simulationsin which
vegetationnitrogenconcentrationis uncoupledfrom decompositiondynamics(Lower N) and coupled to
decompositiondynamics(Lower N+D). The referenceresponsesfor no changein vegetationnitrogen
concentration
are shownfor comparison.
responseof plant growthto elevatedCO2 for much of the terrestrialbiosphere.In environments
with adequatesoilmoisture,
TEM estimatesthatproductionis lesslimitedby nitrogenavailability in warmer environmentsbecauseof a closermatchbetweenpotentialcarbonassimilationand carbonassimilationthat
can be supportedby nitrogencycling [McGuire et al., 1992,
1993]. Becausetemperatureplays an importantrole in the
availability of nitrogen estimatedby TEM, relative NPP responses
in moistenvironments
tendto be highestin the tropics,
intermediatein the temperatezone,and lowestin high-latitude
regions. The modelestimatesthat low nitrogenavailabilityassociatedwith cold soilssubstantiallylimits the ability of vegetationin high-latitudeecosystems
to incorporate
elevatedCO2
into production[seealsoMcGuire et al., 1993; Melillo et al.,
1993]. In comparisonto moistenvironments,
TEM estimates
Similar to NPP, the 6.7% increaseof global carbonstorage
in the referencesimulationof doubledCO2 responseis similar
in magnitudeto the 9% responseof TEM 4.0, the 7% response
of Biome-BGC,andthe 2% response
of Centuryin equilibrium
theseecosystems.
Althoughchanges
in vegetation
carbonin the
referencesimulationof doubledCO2 response
havea similar
spatialpatternto changes
in NPP,thespatialpatternof soilcar-
simulations.
Althoughnumerousexperimentalstudieshave investigated
tissueand plant responsesto elevatedCO2 [Kimball, 1975;
bon response
estimatedby TEM is different.
Kimball and Idso, 1983; Eamus and Jarvis, 1989; Poorter,
doubled CO2 simulations for the conterminousUnited States
[VEMAP Members, 1995]. The partitioningbetweenincreases
in global vegetationand carbonstorageby TEM in this study
(58.3% in vegetation)is similarto the rangeof partitioningestimatedamongthe modelsin the VEMAP study(57 to 67% in
vegetation).For the TEM simulations,the relativeresponses
of
vegetationcarbon storageare similar both in magnitudeand
spatialpatternto the NPP responses.Becauseof the temperature sensitivityof decomposition
in TEM, the portionof the total carbonresponseaccountedfor by increasesin soil carbonare
highestin high-latituderegions,intermediatein the temperate
thattherelativeNPP response
is substantially
higherin deserts zone, and lowestin the tropics. BecauseTEM estimatesmuch
and arid shrublands
becausethe response
of fica, Gv) to dou- of the globalNPP responseto doubledCO2 is concentrated
in
bledCO2is higherin dryenvironments
andnitrogen
availability the tropics,increasesin globalvegetationcarbonstoragetendto
is notasimportant
aswateravailability
in limitingproduction
in be higher than increasesin soil carbon storagein the model
185
MCGUIREETAL.:CO2,VEGETATION
[N],ANDGLOBALCARBONSTORAGE
Reference
100
80-
2ooo
]! (a)
2500
-7.5% [N]
-
(a)
-A,
......1s'ø%
[N]
....
-I
/
'\,., •r',-,,,.,.-•\-..\
5oo
20-
o
..Q
0
I/.
I/.
.... 22.5%
[N]
1000
'.
.:..'•V* f
....... 15.0%[N]
20001
•i'"i""
"""• _ •,••
60-
Reference
-7.5%
[N]
j..df:•'.k,.'%
t..•.'4;-...:...
-'....,
.. -:.,..',..•
..,•.......
.....
-500
-60-50-40-30-20-10
-60-50-40-30-20-10
z
0
0
10 20 30 40 50 60 70 80 90
0
10 20 30 40 50 60 70 80 90
10 20 30 40 50 60 70 80 90
o
lOO (b)
3000 (b)
2500
8O
2000
60
1500
1000
5O0
20
0
0
-500
-60-50-40-30-20-10
-60-50-40-30-20-10
0
10 20 30 40 50 60 70 80 90
Latitude
Figure5. Sensitivity
to 7.5, 15.0,and22.5% decreases
in
vegetationnitrogen concentration
for the latitudinal
distributions
of NPP responses
to doubledCO2as estimated
by theTerrestrial
Ecosystem
Modelfor simulations
in which
changesin vegetationnitrogen concentrationare (a)
uncoupled
fromdecomposition
dynamics
and(b) coupledto
decomposition
dynamics.
Thedoubled
CO2responses
arefor
a change
in atmospheric
CO2from340 to 680 ppmvwithno
changein climatefrom contemporary.The latitudinal
Latitude
Figure 6. Sensitivity
to 7.5, 15.0, and 22.5% decreases
in
vegetationnitrogen concentrationfor the latitudinal
distributions
of vegetationcarbonresponses
to doubledCO2
as estimatedby the Terrestrial EcosystemModel for
simulations in which changes in vegetation nitrogen
concentrationare (a) uncoupled from decomposition
dynamics
and(b) coupledto decomposition
dynamics.The
doubledCO2responses
are for a changein atmospheric
CO2
from 340 to 680 ppmv with no changein climate from
contemporary.The latitudinaldistributionof the reference
distributionof the referenceNPP responsefor no changein response
of vegetationcarbonfor no changein vegetation
vegetation
nitrogenconcentration
is shownfor comparison. nitrogenconcentration
is shownfor comparison. The
Theresolution
of NPPresponses
is 0.50latitude.
resolution
of carbon
storage
responses
is 0.5olatitude.
1993; Ceulemansand Mousseau,1994; Idso and Idso, 1994;
McGuireet al., 1995a;Wullschleger
et al., 1995],thereareonly
taintyin biospheric
responses
to globalchange.Oneof these
a handfulof studiesthatdocumentecosystem-level
responses
to
elevatedCO2[Mooneyet al., 1991;McGuireet al., 1995a].Be-
nitrogenconcentration
on the responses
of NPP andcarbon
causeecosystem-level
responses
of NPP andcarbonstorage
to
elevatedCO2havebeenpoorlydocumented,
the controlsover
uncertainties
involvesthe influenceof changesin vegetation
storage
to increases
in atmospheric
CO2.
Sensitivityof NPP and Carbon StorageResponses
theresponse
arepoorlyunderstood.
Amongspatially
explicit
to Changesin Plant NitrogenConcentration
biogeochemical
models,
thislackof understanding
hasled to
The globalresponses
of NPP andvegetation
carbonstorage
different conceptualizations
about how ecosystemprocesses
controlthe response
of NPP and carbonstorageto elevated for the simulationsin which lower vegetationnitrogenconcenwithdecomposition
dynamics
arerelatively
CO2. The conceptualization
and formulations
in TEM and trationis coupled
to changesin vegetationnitrogenconcentration
othermodelsrepresent
a rangeof empiricalfindingsobserved insensitive
of 4 versus
5 PgC). Theinsensitivity
occurs
because
in variousfield andlaboratory
studies
butdonotencompass
the (increases
attributable
to reducednitrogenconcentration
of
full rangeof environmental
andecosystem
conditions
thatare enhancements
areapproximately
offsetbydecreases
attributable
simulated
in continentalandglobal-scale
studiesof terrestrial thevegetation
rates. In
responses
to elevatedCO2(Y. Panet al., unpublished
manu- to reducedlitter qualityand slowerdecomposition
theresponses
of totalcarbonstoragearesensitive
to
script,1997). Althoughour experimental
understanding
is contrast,
with changesin vegetationnitrogen
poor,spatiallyexplicitbiogeochemical
modelslike TEM are the feedbackassociated
(increases
rangefrom 100Pg C for no changeto
usefultoolsfor exploringthepotentialconsequences
of uncer- concentration
186
MCGUIRE ET AL.: CO2,VEGETATION [N], AND GLOBAL CARBON STORAGE
3000
Reference
(a)
2500
2000
.......
....
1500
-7.5% [N]
Reference
3000
-7.5% [N]
15.0% [N]
22.5ø/ø[N]
A
]
o 2500
5•...2:L• ......
ø.'
.... ':"
100
'::
133 1000
-500
•
-60-50-40-30-20-lO
o
.... 22.5%
[N]
"":""
.
•,",ilf
"•':"v'.
;.?.,...,....,,.,
'
•,
'.
0
lO 20 30 40 50 60 70 80 90
O
-500
•
'c
I
e0
o
o
•.
131
20001
1000
o
....... 15.0ø/ø
[N]
(a)
3ooo
t
(b)
1500
-60-50-40-30-20-10 0 10 20 30 40 50 60 70 80 90
2500
3000 2500 -
.,'/'"/'X,./
, '.
:,,,'¾ ,
(b)
2000 -
1000
'"."
"'.!.
.....,,,,
.',,r...',J:'-x..'.•,
.'"......'.._, ./:....'.,.-'""
1500lOOO -
o-•
-500 ,
,
•
-60-50-40-30-20-10
,
.
0
•
500
•
.
•
.
10 20 30 40 50 60 70 80 90
Latitude
-
o
-500
-60-50-40-30-20-10
Figure 7. Sensitivityto 7.5, 15.0, and 22.5% decreases
in
vegetation nitrogen concentration for the latitudinal
0 10 20 30 40 50 60 70 80 90
Latitude
distributions
of soil carbonresponses
to doubledCO2 as Figure 8. Sensitivityto 7.5, 15.0, and 22.5% decreases
in
estimated
by theTerrestrial
Ecosystem
Modelfor simulations vegetation nitrogen concentration for the latitudinal
in whichchanges
in vegetation
nitrogen
concentration
are(a) distributions
of totalcarbonresponses
to doubledCO2as
uncoupled
fromdecomposition
dynamics
and(b) coupled
to estimated
by theTerrestrial
Ecosystem
Modelfor simulations
decomposition
dynamics.
Thedoubled
CO2responses
arefor in whichchanges
in vegetation
nitrogen
concentration
are(a)
a change
in atmospheric
CO:from340to 680ppmvwithno uncoupled
fromdecomposition
dynamics
and(b) coupled
to
changein climatefrom contemporary.The latitudinal decomposition
dynamics.
Thedoubled
CO2responses
arefor
distribution
of thereference
response
of soilfornochange
in a change
in atmospheric
CO2from340to 680ppmvwithno
vegetation
nitrogen
concentration
is shownfor comparison.changein climatefrom contemporary.The latitudinal
Theresolution
of carbon
storage
responses
is 0.50latitude.
distribution
of the reference
response
of totalcarbonfor no
changein vegetationnitrogenconcentration
is shownfor
comparison.
The resolution
of carbonstorage
responses
is
330 Pg C for a 22.5% decrease).For each7.5% reduction
in
0.50 latitude.
vegetation
nitrogen
concentration,
soil carbonincreases
approximately
an additional
60 Pg C, whilevegetation
carbon
storage
increases
by onlyapproximately
5 PgC. Thusgreater changes
in plantnitrogen
concentration
influence
carbon
cycarbonstorage
is caused
primarilyby increased
soilcarbonstor- cling at the ecosystem
level is basedon a smallnumberof
ageassociated
withlowerdecomposition
ratesandsecondarily studies
[seeMcGuireet al., 1995a]. At present,
theavailable
by increased
vegetation
carbonstorage
associated
with lower datasuggest
thatCO2-induced
reductions
in litterquality
may
vegetation
nitrogen
concentration.
In temperate
andborealre- depress
decomposition
rates,butthedataareambiguous.
Efgions,
theresponses
of soilcarbon
storage
to changes
in vege- fectson nitrogenmineralization
arelesswell documented.
Adtation nitrogenconcentration
are more sensitivethan in the ditional
experimental
research
isrequired
attheecosystem
level
tropicsbecause
reduced
litterqualitycauses
carbonto accumu- to understand
howinteractions
of thenitrogen
cycleandelelatein soilsmoreatlowtemperature
thanathightemperature.vatedCO2affectNPPandcarbonstorage.
Thustheresponse
of vegetation
nitrogen
dynamics
to elevated
Toourknowledge,
thisisthefirststudy
toreport
howpotenatmospheric
CO2andthecouplingof thosedynamics
withthe tialchanges
in vegetation
nitrogen
concentration
mayinfluence
dynamics
of decomposition
haspotential
consequences
forthe theCO2response
of globalterrestrial
carbon
storage
in a geoboththenature
andspatial
pattern
of carbon
storage
responsesgraphically
specific
manner.
Thesensitivity
analysis
reported
in
in theterrestrial
biosphere.
thisstudyis in thecontext
of TEM, whichis a generalized
abAlthough
numerous
experimental
studies
haveinvestigatedstraction
of processes
thatinfluence
carbon
cycling
in terrestrial
theroleof nitrogen
in leaf-andplant-level
responses
of carbon ecosystems
at large spatialscales. The structureof TEM is
uptaketo elevated
CO2,ourknowledge
of howCO2-induced highlyaggregated
bothfor computational
efficiency
andfor
MCGUIREET AL.: CO2,VEGETATION[N], AND GLOBALCARBONSTORAGE
parsimonious
representation
of processes
that are incompletely
understood
at largespatialscales.BecauseTEM is a highlyaggregatedmodel, it doesnot explicitlyrepresentcanopydevelopmentandtissuenitrogenconcentration.
Althoughvegetation
nitrogenconcentration,
canopydevelopment,and photosynthesis are implicitlycoupledin the model,we do not know if the
globaland regionalsensitivityof carboncyclingto changesin
nitrogenconcentration
would be alteredby moreexplicitrepresentations
of vegetationstructureanddevelopment.A comparisonof the resultsin thisstudyto similarsimulationsby models
that explicitly representthe tradeoffbetweencanopydevelopmentandthe acclimationresponse
of tissue-levelphotosynthesiswouldhelp determineif the structure
of TEM shouldbe less
aggregated.
The sensitivityanalysisreportedin this studyis also in the
contextof a closednitrogencycle. Globally,the naturalterrestrial nitrogencyclemay be approximately
balancedwith an es-
187
by doubledatmospheric
CO2,TEM estimatesthatthe effectsof
elevatedCO2 morethanoffsetthe effectsof climatechangeto
cause increasedcarbon storage [Melillo et al., 1993, 1995;
McGuire et al., 1996]. The resultsof this studyindicatethat
carbonstoragewould be furtherenhancedby consideration
of
the influenceof changesin plantnitrogenconcentration
on carbon assimilationand decomposition
rates. Thus, the influence
of changesin vegetationnitrogenconcentration
may have importantimplicationsfor the abilityof the terrestrialbiosphereto
mitigateincreasesin the atmospheric
concentration
of CO2 and
climatechangesassociated
with the increases.
Acknowledgments. We thank JorgeSarmiento,David Schimel,
and four anonymousreviewersfor thoughtfulcommentson earlier
draftsof this paper. This studywas fundedby the Earth Observing
SystemProgramof theNationalAeronauticsandSpaceAdministration
(NAGW-2669) andthe ElectricPower ResearchInstitute.
timated
1601012
g (Tg)N yr-1entering
terrestrial
ecosystems
in
nitrogenfixationandleavingterrestrialecosystems
in river flow
and denitrification[Schlesinger,1991]. Although the natural
nitrogen cycle may be approximatelybalanced,nitrogen releasedfrom fossilfuel burningdepositsbetween50 and 80 Tg
N yr-• globally,
withapproximately
20 Tg N yr-1deposited
on
temperateand boreal forestsof North America and Europe
[Melillo et al., 1996a]. Experimental
evidencesuggests
thatthe
interactionbetweenelevatedCO2 and nitrogendepositionmay
haveimportantimplications
for forestgrowthin regionsof the
world that receivesubstantial
inputsof anthropogenic
nitrogen
from the atmosphere[seeMcGuire et al., 1995a]. Severaleffortsin the internationalbiogeochemical
sciencecommunityare
attemptingto organizespatiallyexplicit data setsof nitrogen
depositionfor both the terrestrialand oceaniccomponentsof
the biosphere.Oncethesedataare available,terrestrialbiogeochemicalmodelslike TEM will be ableto conductanalysesthat
investigatehow the interactionbetweenfossilfuel burning,rising levels of atmosphericCO2, and anthropogenicnitrogen
depositionmay influenceterrestrialcarbonstorage.
Conclusion
During the next century,substantialsimultaneouschanges
areexpectedto occurin a varietyof environmental
variablesincluding atmosphericCO2, temperature,precipitation,cloudiness,and atmosphericdepositionof nutrients[Melillo et al.,
1989b; Mitchell et al., 1990; Watsonet al., 1992]. Also, human
References
Aber, J. D., J. M. Melillo, and C. A. Federer,Predictingthe effectsof
rotationlength,harvestintensity,andfertilizationon fiberyieldfrom
northernhardwoodforestsin New England, For. Sci., 28, 31-48,
1982.
Agren,G.I., The interactionbetweenCO2 and plant nutrition:Comments on a paper by Coleman, McConnaughayand Bazzaz,
Oecologia,98, 239-240, 1994.
Auchmoody,L. R., and H. C. Smith,Responses
of yellow-poplarand
red oak to fertilizationin West Virginia, Soil Sci. Am. J., 41, 803807, 1977.
Bartholomew, M. V., and A. G. Norman, The thresholdmoisture con-
tent for activedecomposition
of somematureplant materials,Soil
Sci. Am. J., 11,270-279, 1946.
Bhaumik, H. D., and F. E. Clark, Soil moisture tension and microbi-
ologicalactivity,SoilSci.Am.J., 12, 234-238, 1947.
Binkley, D., Forest NutritionManagement,JohnWiley, New York,
1986.
Ceulemans,R., andM. Mousseau,Effectsof elevatedatmospheric
CO2
onwoodyplants,NewPhytol.,127, 425-446, 1994.
Chapin,F. S., III, Effectsof multipleenvironmental
stresses
on nutrient
availabilityanduse,in Response
of Plantsto Multiple Stresses,editedby H. A. Mooneyet al.,pp. 67-88, Academic,SanDiego,Calif.,
1991.
Chapin,F. S., III, P.M. Vitousek,andK. Van Cleve,The natureof nutrientlimitationin plantcommunities,
Am.Nat., 127, 48-58, 1986.
Chapin,F. S., III, C. S. H. Walter,andD. T. Clarkson,Growthresponse
of barley and tomatoto nitrogenstressand its controlby abscisic
acid, water relations,and photosynthesis,
Planta, 173, 352-366,
1988.
Daubenmire,R., and D. Prusso,Studiesof decomposition
ratesof tree
litter,_Ecology,
44, 589-592, 1963.
Davidson,R. L., Micro-organisms,
in GrasslandEcosystems
of the
World: Analysisof Grasslandsand Their Uses,editedby R. T.
Coupland,
pp. 267-276,Cambridge
Univ.Press,New York, 1979.
Dodd, J. L., and W. K. Lauenroth,Analysisof the responseof a grasslandecosystem
to stress,in Perspectives
in GrasslandEcology,editedby N. French,pp. 43-58, Springer-Verlag,
New York, 1979.
Eamus,D., and P. G. Jarvis,The directeffectsof increasein the global
atmospheric
CO2 concentration
on naturalandcommercialtemperate treesand forests,Adv. Ecol. Res., 19, 1-55, 1989.
Ellis, R. C., Responseof crop treesof sugarmaple,white ash, and
blackcherryto releaseandfertilization,Can. J. For. Res.,9, 179-
activityis substantiallyalteringthe vegetationof the terrestrial
biosphere[Houghtonet al., 1983, 1991; Melillo et al., 1988;
Houghtonand Skole,1990;Myers, 1991;Melillo et al., 1996b].
Our purposein thisstudyis not to defineterrestrialresponses
to
global change but to assesswhether the coupling between
vegetationnitrogendynamicsand decomposition
dynamicshas
potentialconsequences
for the globalcarboncycle. In the context of TEM, our analysesindicatethat inclusionof theseprocesseshas consequences
not only for carboncycle responses
to
elevatedatmosphericCO2 but alsofor climatechangethat may
188, 1979.
accompany
risinglevelsof CO2. For climatechangeswithout
Evans,J. R., Nitrogenand photosynthesis
in the flag leaf of wheat
changesin atmosphericCO2, previoussimulationswith TEM
(TriticumaestivumL.), Plant Physiol.,72, 297-302, 1983.
indicatethat globalNPP is relativelyinsensitiveand that total Evans,J. R., Photosynthesis
andnitrogenrelationships
in leavesof C3
carbonstoragedecreases
[Melillo et al., 1993, 1995; McGuire
plants,Oecologia,78, 9-19, 1989.
Systems
ResearchCenter
et al., 1996]. However,when climatechangeis accompanied FoodandAgricultureOrganization/Complex
188
MCGUIREET AL.: CO2,VEGETATION[N], AND GLOBALCARBONSTORAGE
(FAO/CSRC),0.5ø digitization
of FAO-UNESCO[1971] by ComplexSystems
Research
Center,Univ.of N.H., Durham,undated.
FoodandAgricultureOrganization-United
NationsEducational,
Scientific Organization(FAO-UNESCO), Soil map of the world,
1:5,000,000, Paris, 1971.
restrialgrid, Res. Rep. RR-91-18, Int. Inst. of Appl. Syst.Anal.,
(IIASA), Laxenburg,Austria,1991.
Luo, Y., C. B. Field,andH. A. Mooney,Predictingresponses
of photosynthesis
and root fractionto elevated[CO2]a2Interactions
among
carbon,nitrogen,and growth,Plant Cell Environ.,17, 1195-1204,
1994.
Farquhar,G. D., andS. von Caemmerer,
Modellingof photosynthetic
response
to environmental
aonditions,
in Encyclopedia
of Plant McGuire, A.D., J. M. Melillo, L. A. Joyce,D. W. Kicklighter,A. L.
Grace, B. Moore III, and C. J. Vorosmarty, Interactionsbetween
Physiology,New Ser., vol 12 B, editedby O. L. Langeet al., pp.
carbonandnitrogendynamicsin estimatingnet primaryproductivity
549-587,Springer-Verlag,
New York, 1982.
for potentialvegetation
in NorthAmerica,GlobalBiogeochem.
CyFarquhar,G. D., S. von Caemmerer,and J. A. Berry, A biochemical
cles, 6, 101-124, 1992.
modelof photosynthetic
CO2 assimilationin leavesof C3 species,
Planta, 149, 79-90, 1980.
Field, C. B., Ecologicalscalingof carbongain to stressand resource
availability,in Response
of Plantsto Multiple Stresses,
editedby H.
A. Mooneyand W. E. Winner, pp. 35-65, Academic,San Diego,
Calif., 1991.
McGuire, A.D., L. A. Joyce,D. W. Kicklighter, J. M. Melillo, G.
Esser,and C. J. Vorosmarty,Productivityresponseof climax tem-
perateforeststo elevatedtemperature
andcarbondioxide:A North
Americancomparison
betweentwo globalmodels,Clim. Change,
24, 287-310, 1993.
Gates, D. M., Global biosphericresponseto increasingatmospheric McGuire, A.D., J. M. Melillo, andL. A. Joyce,The role of nitrogenin
carbondioxide concentrations,
in Direct Effectsof IncreasingAtthe responseof forestnet primaryproductionto elevatedatmosmosphericCarbon Dioxide on Vegetation,editedby B. R. Strain
phericCO2,Annu.Rev.Ecol.Syst.,26, 473-503,1995a.
andJ. D. Cure,Rep.DOE/ER-0238, pp. 171-184,U.S. Department McGuire, A.D., J. M. Melillo, D. W. Kicklighter, and L. A. Joyce,
of Energy,Washington,D.C., 1985.
Equilibriumresponses
of soil carbonto climatechange:Empirical
Gulmon, S. L., and C. C. Chu, The effectsof light and nitrogenon
andprocess-based
estimates,
J. Biogeogr.,22,785-796, 1995b.
McGuire, A.D., D. W. Kicklighter,and J. M. Melillo, Globalclimate
photosynthesis,
leaf characteristics,
and dry matterallocationin the
changeand carboncyclingin grasslandsand coniferforests,in
chaparralshrub,Diplacus aurantiacus,Oecologia,49, 207-212,
1981.
Global Change: Effectson ConiferousForests and Grasslands
(SCOPE56), editedby A. I. Breymeyer,D. O. Hall, J. M. Melillo,
Gunderson,C. A., and S. D. Wullschleger,Photosynthetic
acclimation
andG.I. Agren,pp. 388-411,JohnWiley, New York, 1996.
in treesto risingatmospheric
CO2: A broaderperspective,
PhotoMelillo, J. M., J. D. Aber, andJ. F. Muratore,Nitrogenandlignin consynth.Res.,39, 369-388, 1994.
trol of hardwoodleaf litter decomposition
dynamics,Ecology,63,
Harrison,K., W. Broecker,andG. Bonani,A strategyfor estimatingthe
621-626, 1982.
impactof CO2 fertilizationon soil carbonstorage,Global BiogeoMelillo, J. M., J. R. Fruci,R. A. Houghton,B. Moore, andD. L. Skole,
chem.Cycles,7, 69-80, 1993.
Land-usechangein the Soviet Union between 1850 and 1980:
Houghton,R. A., and D. L. Skole, Carbon,in The Earth as TransCausesof a netreleaseof CO2to the atmosphere,
Tellus,Ser.B, 40,
formed by Human Action, editedby B. L. Turner,W. C. Clark, R.
116-128, 1988.
W. Kates,J. F. Richards,J. T. Mathews,andW. B. Meyer, pp. 393Melillo, J. M., J. D. Aber, A. E. Linkins, A. Ricca, B. Fry, and K. J.
408, CambridgeUniv. Press,New York, 1990.
Nadelhoffer,Carbondynamicsalong the decay continuum:Plant
Houghton,R. A., J. E. Hobbie,J. M. Melillo, B. Moore, B. J. Peterson,
litterto soilorganicmatter,in Ecologyof Arable Land,editedby M.
G. R. Shaver,and G. M. Woodwell, Changesin the carboncontent
of terrestrial biota and soils between 1860 and 1980: A net release
Clarholmand L. Bergstrom,pp. 53-62, Kluwer Acad., Norwell,
Mass., 1989a.
of CO2to theatmosphere,
Ecol.Monogr.,53, 235-262, 1983.
Melillo, J. M., P. A. Steudler, J. D. Aber, and R. D. Bowden, AtmosHoughton,R. A., D. L. Skole, and D. S. Lefkowitz, Changesin the
phericdeposition
andnutrientcycling,in Exchangeof TraceGases
landscape
of Latin Americabetween1850 and 1985, I, Progressive
BetweenTerrestrialEcosystems
and the Atmosphere,editedby M.
lossof forests,For. Ecol. Manage., 38, 143-172, 1991.
O. AndreaeandD. S. Schimel,pp. 263-280,JohnWiley, New York,
Hunt, H. W., A simulationmodel for decompositionin grasslands,
1989b.
Ecology,58, 469-484, 1977.
Idso,K. E., andS. B. Idso,Plantresponses
to atmospheric
CO2enrich- Melillo, J. M., A.D. McGuire, D. W. Kicklighter,B. Moore III, C. J.
Vorosmarty,and A. L. Schloss,Global climate changeand terresmentin the face of environmental
constraints:
A reviewof the past
trial netprimaryproduction,
Nature,363, 234-240, 1993.
10 yearsresearch,
Agric. For. Meteorol.,69, 153-203, 1994.
Melillo, J. M., D. W. Kicklighter,A.D. McGuire,W. T. Peterjohn,and
Ino, Y., andM. Monsi, An experimentalapproachto the calculationof
K. M. Newkirk, Globalchangeandits effectson soil organiccarbon
CO2 amountevolvedfrom severalsoils,Jpn. J. Bot., 20, 153-188,
1969.
stocks,in Role of NonlivingOrganicMatter in the Earth's Carbon
Cycle, editedby R. G. Zepp and C. Sonntag,pp. 175-189, John
Intergovernmental
Panelon Climate ChangeWorking GroupI (IPCC
Wiley, New York, 1995.
WGI), Climate Change 1995- The Scienceof Climate Change.'
Melillo, J. M., I. C. Prentice,G. D. Farquhar,E. -D. Schulze,andO. E.
Contributionof WorkingGroupI to the SecondAssessment
Report
Sala,Terrestrialbioticresponses
to environmental
changeandfeedof the Intergovernmental
Panel on Climate Change,editedby J. T.
backsto climate,in ClimateChange1995 - The Scienceof Climate
Houghton,L. G. Meira Filho, B. A. Callander,N. Harris, A. KattenChange:Contributionof WorkingGroupI to the SecondAssessberg,andK. Maskell, CambridgeUniv. Press,New York, 1996.
mentReportof the Intergovernmental
Panel on ClimateChange,
Jenny,H., S. P. Gessell,and T. Bingham,Comparativestudyon deeditedby J. T. Houghton,L. G. Meira Filho, B. A. Callander,N.
compositionratesof organicmatter in temperateand tropicalregions,Soil Sci.,68, 419-432, 1949.
Harris, A. Kattenberg,and K. Maskell, pp. 447-481, Cambridge
Univ. Press,New York, 1996a.
Kicklighter,D. W., J. M. Melillo, W. T. Peterjohn,E. B. Rastetter,A.
Melillo, J. M., R. A. Houghton,D. W. Kicklighter,andA.D. McGuire,
D. McGuire,andP. A. Steudler,Aspectsof spatialandtemporalaggregationin estimatingregionalcarbondioxidefluxesfrom temperTropicaldeforestation
andtheglobalcarbonbudget,Annu.Rev.Energy Environ.,21,293-310, 1996b.
ateforestsoils,J. Geophys.Res.,99, 1303-1315, 1994.
Forestry,54,
Kimball, B. A., Carbondioxideand agriculturalyield: An assemblage Miller, H. G., Forestfertilization:Someguidingconcepts,
157-167, 1981.
and analysisof 430 prior observations,
Agron. J., 75, 779-788,
1975.
Kimball,B. A., andS. B. Idso,Increasingatmospheric
CO2:Effectson
cropyield, wateruseandclimate,Agric. WaterManage., 7, 55-72,
1983.
Miller, R. D., and D. D. Johnson,The effect of soil moisturetensionon
carbondioxideevolution,nitrification,and nitrogenmineralization,
Soil Sci. Soc.Am. J., 28, 644-647, 1964.
Mitchell, H. L., and R. F. Chandler Jr., The nitrogen nutrition and
Lajtha, K., and W. G. Whitford, The effect of water and nitrogen
growth of certaindeciduoustrees of northeastern
United States,
Black Rock For. Bull., 11, 1-94, 1939.
amendments
on photosynthesis,
leaf demography,
and resource-use
efficiencyin Larrea tridentata,a desertevergreen
shrub,Oecologia, Mitchell, J. F. B., S. Manabe, V. Meleshko, and T. Tokioka, Equilib80, 341-348, 1989.
riumclimatechange--Andits implications
for thefuture,in Climate
LeemansR., andW. P. Cramer,The IIASA databasefor meanmonthly
Change:The IPCC ScientificAssessment,
editedby J. T. Houghton
valuesof temperature,precipitationand cloudinessof a global teret al., pp. 131-172,CambridgeUniv. Press,New York, 1990.
MCGUIRE ET AL.: CO2,VEGETATION [N], AND GLOBAL CARBON STORAGE
Mooney, H. A., B. G. Drake, R. J. Luxmoore,W. C. Oechel, and L. F.
Pitelka,Predictingecosystem
responses
to elevatedCO2 concentrations, BioScience, 2, 96-104, 1991.
Myers, N., Tropical forests:Presentstatusand future outlook, Clim.
Change,19, 3-32, 1991.
NationalCenterfor AtmosphericResearch(NCAR)/Navy, Global 10minuteelevationdata, digital tape availablethroughNat. Oceanic
and Atmos. Admin., Nat. Geophys.Data Cent., Boulder, Colo.,
1984.
Pan, Y., Sensitivityof terrestrialecosystems
to elevatedatmospheric
CO2:Comparisonof modelsimulationstudiesto CO2 effects.Bull.
Ecol. Soc.Am., 76, 205, 1995.
Pan,Y., A.D. McGuire,D. W. Kicklighter,andJ. M. Melillo, The importanceof climate and soilsfor estimatesof net primary production: A sensitivityanalysiswith the TerrestrialEcosystemModel,
Global ChangeBiol., 2, 5-23, 1996.
Pastor,J., J. D. Aber, C. A. McClaugherty,and J. M. Melillo, AbovegroundproductionandN andP cyclingalonga nitrogenmineralization gradienton BlackhawkIsland,Wisconsin,Ecology,65, 256268, 1984.
Peterson,C., Regional growth and responseanalysisfor unthinned
Douglas-fir,in RegionalForestNutritionResearchProjectBiennial
Report, 1980-82, pp. 3-25, Univ. of Wash. Coll. of For. Resour.,
Seattle, 1982.
Pettersson,
R., andA. J. S. McDonald,Effectsof nitrogensupplyon the
acclimationof photosynthesis
to elevatedCO:, Photosynth.
Res.,39,
389-400, 1994.
Poorter,H., Interspecificvariationin the growthresponseof plantsto
an elevatedambientCO2 concentration,Vegetatio,104/105, 77-97,
1993.
189
effect of water potentialon decomposition
processes
in soils,in
WaterPotentialRelationsin SoilMicrobiology,editedby J. F. Parr,
W. R. GardnerandL. F. Elliott, Spec.Pub.9, Soil Sci. Soc.Am., 9,
97-117, 1981.
Stott,D. E., L. F. Elliott, R. I. Papendick,and G. S. Campbell,Low
temperatureor low water potentialeffectson the microbialdecompositionof wheatresidue,Soil Biol. Blochem.,18, 577-582, 1986.
Trumbore,S. E., O. A. Chadwick,andR. Amundson,
Rapid exchange
betweensoilcarbonandatmospheric
carbondioxidedrivenby temperaturechange,Science,272, 393-396, 1996.
Van Cleve,K., andJ. Zasada,Response
of 70-year-oldwhite spruceto
thinningand fertilizationin interiorAlaska, Can. J. For. Res., 6,
145-152, 1976.
Van Veen,J. A., andE. A. Paul,Organiccarbondynamicsin grassland
soils,1, Backgroundinformationand computersimulation,Can. J.
Soil Sci., 61, 185-201, 1981.
VEMAP Members,Vegetation/ecosystem
modelingand analysisproject: Comparingbiogeographyand biogeochemistry
modelsin a
continental-scale
studyof terrestrialecosystem
responses
to climate
changeandCO2doubling,GlobalBiogeochem.Cycles,9, 407-437,
1995.
Vitousek,P.M., andR. W. Howarth,Nitrogenlimitationon landandin
the sea:How canit occur?,Biogeochemistry,
13, 87-115, 1991.
Vitousek, P.M., P. R. Ehrlich, A. H. Ehrlich, and P. A. Matson, Human
appropriationof the productsof photosynthesis,
BioScience,36,
368-373, 1986.
Vitousek, P.M., L. R. Walker, L. D. Whiteaker, and P. A. Matson,
Nutrientlimitationto plantgrowthduringprimarysuccession,
Biogeochemistry,
23, 197-215, 1993.
Vorosmarty,C.J., B. Moore III, A. L. Grace, M.P. Gildea, J. M.
Raich, J. W., E. B. Rastetter,J. M. Melillo, D. W. Kicklighter,P. A.
Melillo, B. J. Peterson,E. B. Rastetter, and P. A. Steudler, Conti-
Steudler, B. J. Peterson, A. L. Grace, B. Moore III, and C. J.
nentalscalemodelsof waterbalanceandfluvialtransport:
An applicationto SouthAmerica,Global Biogeochem.Cycles,3, 241-265,
Vorosmarty,Potentialnet primary productivityin SouthAmerica:
Applicationof a globalmodel,Ecol.Appl., 1,399-429, 1991.
Risser,P. G., E. C. Birney, H. D. Blocker, S. W. May, W. J. Parton,
and J. A. Wiens, The True Prairie Ecosystem,Van NostrandReinhold, New York, 1981.
Safford,L. O., andS. M. Filip, Biomassandnutrientcontentof 4-yearold fertilized and unfertilized hardwood stands,Can. J. For. Res., 4,
549-554, 1974.
Sage,R. F., Acclimationof photosynthesis
to increasingatmospheric
CO2:The gasexchangeperspective,
Photosynth.Res.,39, 351-368,
1994.
1989.
Watson,R. H., H. Rodhe,H. Oeschger,and U. Siegenthaler,
Greenhousegasesand aerosols,in Climate Change:The IPCC Scientific
Assessment,
editedby J. T. Houghtonet al., pp. 5-40, Cambridge
Univ. Press,New York, 1990.
Watson, R. T., L. G. M. Filho, E. Sanhueza, and A. Janetos, Green-
housegases:SourcesandSinks,in ClimateChange1992: The SupplementaryReport to the IPCC ScientificAssessment,
editedby J.
T. Houghtonet al., pp. 25-46, CambridgeUniv. Press,New York,
1992.
Wong, S.C., Elevatedatmospheric
partial pressureof CO2 and plant
Sage,R. F., andR. W. Pearcy,The nitrogenuseefficiencyof C3 andC4
growth,I, Interactionsof nitrogennutritionand photosynthetic
caplants,I, Leaf nitrogen,growth, and biomasspartitioningin Chepacityin C3andC4plants,Oecologia,44, 68-74, 1979.
nopodiumalbion (L.) and Amaranthusretrofiexus(L.), Plant
Wullschleger,S. D., Biochemicallimitationsto carbonassimilationin
Physiol.,84, 954-958, 1987a.
C3 plants --A retrospectiveanalysisof the A/Ci curvesfrom 109
Sage,R. F., andR. W. Pearcy,The nitrogenuseefficiencyof C3 andC4
species,J. Exp.Bot., 44, 907-920, 1993.
plants,II, Leaf nitrogeneffectson the gas exchangecharacteristics
Wullschleger,
S. D., W. M. Post,andA. W. King,On thepotentialfor a
of Chenopodium
albion(L.) andAmaranthusretrofiexus(L.), Plant
CO2 fertilizationeffect in forests:Estimatesof the biotic growth
Physiol.,84, 959-963, 1987b.
factorbasedon 58 controlled-exposure
studies,in Biotic Feedbacks
Sarmiento,J. L., C. Le Quere,andS. W. Pacala,Limitingfutureatmosin the Global ClimaticSystem,editedby G. M. WoodwellandF. T.
pheric carbon dioxide, Global Biogeochem.Cycles, 9, 121-137,
1995.
Mackenzie,pp. 85-107, OxfordUniv. Press,New York, 1995.
Xiao, X., D. W. Kicklighter,J. M. Melillo, A.D. McGuire,P. H. Stone,
Schimel, D. S., B. H. Braswell, E. A. Holland, R. McKeown, D. S.
andA. P. Sokolov,Responses
of primaryproductionand total carOjima, T. H. Painter,W. J. Parton,and A. R. Townsend,Climatic,
bon storageto changesin climateand atmospheric
CO2 concentraedaphic,andbioticcontrolsoverstorageandturnoverof carbonin
tion,Tellus,in press,1997.
soils,GlobalBiogeochem.
Cycles,8, 279-293, 1994.
Schlesinger,
W. H., Biogeochemistry:
An Analysisof Global Change,
J. Helfrich,D. W. Kicklighter,J. M. Melillo, Y. Pan,andX. Xiao,
Academic,SanDiego,Calif., 1991.
The Ecosystems
Center,MarineBiologicalLaboratory,
WoodsHole,
Shaver,G. R., andF. S. Chapin,III, Responseof fertilizationby various
MA 02543.(e-mail:[email protected];
[email protected];
plantgrowthformsin an Alaskantundra:Nutrientaccumulation
and
jmelillø@lupine'mbl'edu;
[email protected];
[email protected].
growth,Ecology,61,662-675, 1980.
Shaver,G. R., andF. S. ChapinIII, Effect of fertilizationon production edu)
A.D. McGuire, U.S. GeologicalSurvey,Alaska CooperativeFish
andbiomassof tussocktundra,Alaska,U.S.A., Arct. Alp. Res.,18,
andWildlife ResearchUnit, Universityof AlaskaFairbanks,Fairbanks,
261-268, 1986.
Shaver,G. R., and F. S. ChapinIII, Production:Biomassrelationships AK 99775. (e-mail: ffadm@aurøra'alaska'edu)
B. Moore III, A. L. Schloss,and C. J. Vorosmarty,Complex Sysand elementcycling in contrastingarctic vegetationtypes, Ecol.
tems ResearchCenter, Institute for the Study of Earth, Oceans, and
Monogr.,61, 1-31, 1991.
Space,Universityof New Hampshire,Durham, NH 03824. (e-mail:
Sollins,P., C. C. Grief, F. M. McCorison,K. Cromack,Jr., R. Fogel,
[email protected];
cv@cycling'unh'edu)
andR. L. Fredriksen,The internalelementcyclesof an old-growth b'møøre@unh'edu;
Douglas-firecosystem
in westernOregon,Ecol. Monogr.,50, 261285, 1980.
(ReceivedFebruary1, 1996;revisedDecember1, 1996;
Sommers,L. E., C. M. Gilmour,R. E. Wildung,and S. M. Beck, The
acceptedJanuary3, 1997.)