O in atmospheric CO - Semantic Scholar
JOURNAL
OF GEOPHYSICAL
RESEARCH,
VOL. 102, NO. D5, PAGES 5857-5872, MARCH
A three-dimensional synthesis studyof 1.
Surface
20, 1997
in atmospheric COz
fluxes
PhilippeCiais,• A. Scott Denning,2 Pieter P. Tans,3 JosephA. Berry,4 David A. Randall, 2 G. James Collatz, s Piers J. Sellers,s James W. C. White, 6
Michael Trolier,3,6Harro A. J. Meijer,7 Roger J. Francey,s Patrick Monfray,9 and Martin Heimann•ø
Abstract.Theisotope•SOin CO2is of particular interestin studying theglobalcarbon cyclebecauseit is sensitiveto the processes by which the global land biosphereabsorbs and respiresCO2. Carbon dioxideand water exchangeisotopicallyboth in leavesand in
soils,andthe•SOcharacter of atmospheric CO2is strongly influenced bythelandbiota, which shouldconstrainthe grossprimary productivityand total respirationof land
ecosystems. In thisstudywecalculate theglobalsurface fluxesof •SOforvegetation and
soilsusingthe SiB2 biospheremodel coupledwith the ColoradoState Universitygeneral circulationmodel. This approachmakesit possibleto use physiologicalvariablesthat are consistentlyweightedby the carbonassimilationrate and integratedthrough the exchangeof rsO and the isotopic vegetation canopy. We alsocalculate theair-sea characterof fossilemissionsand biomassburning.Global mean valuesof the isotopic
exchange witheachreservoir areusedto closetheglobalbudgetof •SOin CO2.Our
resultsconfirm the fact that the land biota exert a dominant control on the (5•SOof the
atmosphericreservoir.At the global scale,exchangewith the canopyproducesan isotopic enrichmentof CO2, whereasexchangewith soilshas the oppositeeffect. vated atmosphericCO2 levels,changingnutrient availability, temperature,and precipitationpatterns. Increasingattentionhasbeen givenrecentlyto the terrestrial Modelsof ecosystem functioninghavebecomeprogressively biospherein controllingatmosphericCO2 levelsbecausethe more processoriented,especiallyregardingthe photosynthetic carbonstoredin the abovegroundbiomassand in soil organic uptake of CO2. A few global mechanisticmodels based on mattercanbe exchangedrapidlywith the atmosphere.It iswell external"climatic"forcing,suchas the incidentsolar flux, the recognizedthat land ecosystems take up and release large water availabilityfor plants,and the temperature,are able to quantitiesof CO2 not only on a daily and seasonaltimescale calculate the gross fluxes of CO2 exchangedbetween land but also in the long term. Severalmodelsof the global bio- ecosystemsand the atmosphere.Figure la givesa schematic sphereon land have been developedwhich simulatethe be- pictureof the cyclingof carbonbetweenplants,soils,and the havior of various ecosystemsand have been used in order to atmosphere.Of particularimportanceto the atmosphericCO2 (A is gross predictthe carbonfluxesexchangedwith the atmosphereun- budget is the uptake of CO2 by photosynthesis (GPP) minusleaf respiration•d) andthe der specificscenariosof future climate changeincludingele- primaryproductivity accompanying ecosystemtotal respiration(•). Respiratory 1.
Introduction
•Laboratoirede Mod61isationdu Climat et de l'Environnement, Commissariath l'Energie Atomique l'Orme des Merisiers, Gif sur Yvette, France.
CO2emissions includeaboveground plantrespiration ^ (•plants) andbelowground rootrespiration B (?•Plants) aswell as heterotrophicsoilrespiration(•soi•s),the total CO2 effiuxfrom soils
2Department of Atmospheric Sciences, ColoradoStateUniversity, being called Fsoils. Over the course of the year, the annual Fort Collins. mean assimilationA is almostentirely compensatedby respi3ClimateMonitoringandDiagnostic Laboratory, NOAA, Boulder, ration emissions. Colorado.
The CO2 biosphericfluxescalculatedby ecosystemmodels can be partially validated againstatmosphericobservations. For instance,a very usefulvalidationis to comparethe seaSNASAGoddardSpaceFlightCenter,Greenbelt,Maryland. 6Institute of ArcticandAlpineResearch andDepartmentof Geo- sonalvariation in atmosphericCO2 simulatedwith givenbiologicalSciences,Universityof Colorado,Boulder. sphericfluxesto the well-documentedobservationalrecordat 7Centrum voorIsotopenOnderzoek, University of Groningen, Gronumeroussitesaroundthe world [Conwayet al., 1994].This is ningen,Netherlands. 8Division of Atmospheric Research, Commonwealth Scientific and commonly done by coupling the calculated field of the net IndustrialResearchOrganisation,Melbourne,Victoria, Australia. ecosystem flux of CO2, the differencebetweenA and •, to an 9CentredesFaiblesRadioactivit6s, Laboratoirede Mod61isation du atmospherictransport model, and comparingthe results to Climat et de l'Environnement, Gif sur Yvette, France. observations. This approachhasprovenvery valuablefor test•øMax-Planck-Institut ffir Meteorologie, Hamburg,Germany. ing the seasonality of net CO2 fluxes[e.g.,Funget al., 1987]but Copyright1997 by the American GeophysicalUnion. doesnot constrainthe grossfluxesof CO2 separately. The globalbudgetof atmosphericCO2 hasalsobeenstudied Paper number 96JD02360. 4Department of PlantBiology,Carnegie Institution of Washington,
Stanford, California.
0148-0227/97/96JD-02360509.00
usingmeasurements of the /5•3Cof atmospheric CO2.The
5857
5858
CIAIS ET AL.: STUDY OF &•80IN ATMOSPHERICCO2,1
FALingoing CO2
FLA retrodiffusion ofCO2
A
•Plants
stem, twigs '•' respiration
Fso,s
•
totalsoil • total respiration
respiration
•lant
!plant belowsoilheterotrophic :ground respiration respiration Figure la. (a) The cyclingof carbonbetweenthe land biosphereand the atmosphere.The ecosystem representedis in equilibriumsincethe annualmeanuptakeof CO2by photosynthesis (A) compensates exactly the total respiratoryloss(•).
methodrelieson theinterpretation of atmospheric 8•3Cvariationsas indicatingnet biosphericfluxes[Tans et al., 1993; Franceyet al., 1995;Keelingetal., 1995;Entinget al., 1993,1995; Ciais et al., 1995]. The methodis limited by uncertaintyconcerningthe influenceof isotopicdisequilibriabetweenatmosphereand surfacereservoirs.Suchdisequilibriacan be transferred to the atmosphereby grossexchangefluxes,evenin the absenceof net exchange. The Earth'svegetationlikely exertsa major influenceon the
For CO> all isotopicvalues are given relative to the standardisotopicratio Vienna Pee Dee belemnite(VPDB)-CO: = 0.002088349077 as recommendedbyAllisonet al. [1995].For H:O we expressisotopicabundancerelative to the standard Vienna SMOW (VSMOW) = 0.00200520[Baertchi andMackfin, 1965]. We must subtract41.47%o to expressVSMOW valuesin the VPDB-CO: scale.This includesa differenceof -30.9%0 between VSMOW and VPDB-calcite [Hut, 1987]
andaccounts forthe•gOfractionation duringCO: evolution at •80/•60 ratioof atmospheric CO2[Keeling, 1995].Francey and 25øCwith 100%phosphoricacid [Friedmanand O'Neill,1977]
Tans [1987] first pointedout that the isotopicexchangewith water in leaves (and possiblysoils) may determine the ob-
betweenVPDB-calcite and VPDB-CO2.
served persistent north-south differences in 1gOof atmospheric 1.2. Climate Variables Used in This Study:
CO2.Farquharet al. [1993]furtherquantifiedthe globalrole of leaf exchangeand calculateda globalatmosphericbudgetof
CSU GCM
and SiB2 Model
The Colorado State University (CSU) generalcirculation model(GCM) is derivedfrom the Universityof California,Los is controlledby the fluxesA and •. We presenthere a syn- Angeles,(UCLA) GCM, whichwasdevelopedat UCLA over thesissimulation of •gOin CO2thatwe compare with atmo- a period of 20 yearsby A. Arakawa and collaborators.A copy sphericmeasurements. In the presentpaper we focuson the of the model was brought to the Goddard Laboratory for mechanisms thatgovernthe•gO/•60ratioin CO:. Specifically,Atmospheresin 1982 and from there to CSU in 1988. Many we havecalculatedon a 4øby 5ø grid the isotopicfluxesasso- changeshavebeen madesincethe modelleft UCLA, including ciated with the terrestrial and oceanic reservoirs, as well as revised parameterizationsof solar and terrestrial radiation with anthropogenicCO: emissions. In the companionpaperby [Harshvardhanet al., 1987], the planetary boundary layer Ciaiset al. [this issue]we haveprescribedthesefluxesin the (PBL) [Randallet al., 1992],cumulusconvection[Randalland processes [Fowleret al., 1995], three-dimensionalatmospherictracer model TM2 and com- Pan, 1993],cloudmicrophysical [Sellerset al., 1986,1992a,b, 1996a, paredthesimulated 8•80 valuesto atmospheric observations.and land-surfaceprocesses b]. Somerecent resultsare presentedby Randall et al. [1989, 1.1. Conventions and Units 1991, 1996], Fowler et al. [1995], and Fowler and Randall [1995a, b]. In this paper, sinkscorrespondto a negativenet flux of The prognosticvariablesof the CSU GCM are potential carbon(CO2 is removedfrom the atmosphere)and sources correspondto a positivenet flux (CO2 is releasedto the atmo- temperature;the horizontal wind components;the surface sphere).Isotopicratiosare expressed in per mil (%0), definedas pressure;the PBL's depth and turbulencekinetic energy;the mixingratio of three phasesof water plus rain and snow;the ( O/ O)sampl e -( )standard temperaturesof the plant canopy,the groundsurface,and the 8180 = 1000 •J/ •JJstandard 18t•/16t•\ deep soil; the water contentsof four abovegroundand three
180in CO2.Specifically, thelgo/•60ratioof atmospheric CO:
1816
180/160
CIAIS ET AL.: STUDY OF 8•80 IN ATMOSPHERICCO2,1 belowgroundmoisturestores;the stomatalconductanceof the plant canopy;and the ice temperatureat land ice and sea ice points.The governingequationsare finite-differenced,using highlyconservative schemes[Arakawaand Lamb, 1977,1981]. The modelis formulatedin termsof a modifiedsigmacoordinate, in which the PBL top is a coordinatesurface,and the PBL itself is identifiedwith the lowestmodel layer [Suarezet al., 1983].The masssourcesand sinksfor the PBL consistof large-scaleconvergence or divergence,turbulententrainment, and the cumulus
mass flux. Turbulent
entrainment
can be
drivenbypositivebuoyancyfluxesor by shearof the meanwind in the surfacelayer or at the PBL top. For vegetatedland pointsthe surfacefluxesof sensibleand
5859
Direct isotopicexchangebetweenCO2 and H20 vapor is excludedbecausethe rate of hydrationis slow (severalminutes)andonlya verysmallfractionof CO2 is dissolved in liquid water at any time [Franceyand Tans, 1987]. However, the enzymecarbonicanhydrase(CA), ubiquitousin plant tissues, catalyzesthe hydration and stronglyacceleratesthe rate of reaction(1) [Silverman,1982].In livingplant tissuesthe isotopic equilibrium between CO2 and H20 is reached quasiinstantaneously. Little is knownof CA activityin soils,but CO2 producedfrom decayingplant tissuesremainsin contactwith soil water for sufficienttime (see below) for reaction(1) to occur and most likely yield full isotopicequilibrationof CO2
withwater,evenin the absence of CA. The •80 of CO2in
latent heat, radiation, moisture, and momentum are deter-
leavesand in soilscan thereforebe predictedby reaction(1),
minedusingthe simplebiosphere(SiB) parameterizationdeveloped by Sellerset al. [1986]. SiB has recentlyundergone substantialmodification[Sellerset al., 1996a,b; Randall et al., 1996] and is now referred to as SiB2. The numberof biomespecificparametershas been reduced,and most are now derived directly from processedsatellite data rather than prescribedfrom the literature. The vegetationcanopyhas been reduced to a single layer. Another major change is in the parameterizationof stomataland canopyconductance [Collatz et al., 1991, 1992;Sellerset al., 1992a,b, 1996a] used in the calculationof the surfaceenergybudget over land. This pa-
provided weknowthe•80 ofwaterreacting withCO2andthe
rameterization
involves
the direct
calculation
of the rate of
carbon assimilationby photosynthesis, making possiblethe calculationof CO2 exchangebetweenthe atmosphereand the terrestrialbiota at the dynamictime step(6 min) of the CSU GCM [Denning,1994;Denninget al., 1996;Dennyand Randall, 1996]. Details of the carbonflux calculationsand their use in isotopicexchangecalculationsare presentedin Appendix B.
2. OxygenIsotope Fractionation BetweenCOz and HzO Of major importancefor the isotopiccompositionof CO2 in the atmosphereis the fact that dissolvedCO2 mayexchangean
•80 atomwith water according to the isotopicequilibrium reaction(1):
COO + H2180 transition, equalto 1 + (•_vap/1000)----R•/RvWap; usedto distinguish a•' kineticfractionation of H2180versus H2160in the {sv•p •tm) inthelimith -->0 andplanttranspired vapor ({5180 --> w
diffusionof water vapor acrossthe stomatalcavity and leaf boundarylayer,equalto 1 + (•'/1000);
Rv•p 180/160 ratioofwatervapor in theairoutside the
{5)")in the limit h --> 1. We did not extrapolatethis empirical
regression to the globallevelto correct{5180of vaporin canopy, however.By usingthe GISS model fieldswe insteadas-
leaf;
sumea lowerboundary for {5180 in watervaporandhencefor {5•80of CO2in leaves. roots. Plate2c shows{5•80of leaf CO2,whichdecreases toward Overdryareas,{5180 of leafCO2islargerthan The firstimportantparameter in (12') is {5•',the {5180of highlatitudes. R•' •80/160ratioof groundwater whichistakenup by
groundwaterdeliveredto the leaf. At steadystate an equivalent amountof water deliveredto the leaf and lost by transpiration mustbe pumpedfrom the soil by the root system.Following the hypothesisof the SiB2 model soil hydrology,we considerthat the rootspumpgroundwaterfrom an intermediate soillayerbeneaththe surface(FigureA1 in AppendixA). Assumingthat no isotopicfractionationoccursduringthe root uptakeof water [Bariacet al., 1994b],we calculate{5•"from the
6%0, with a maximum over the Sahara Desert of 15%o. This is
mostlydue to low relative humiditywhich increasesthe value of {SLin (12). Note, however,that the maximumvaluesob-
tainedin deserts isnotlikelyto influence theatmospheric {5180 becauseit is associated with a negligibleexchangeof CO2. The simulatedisotopiccompositionof leaf CO2 is aslow in tropical rainforestsas in Siberianforests(roughly-3%o), despitethe
fact that leaf wateris moredepletedin 180 in Siberia.As
a)
Relative Humidityat the Leaf Surface Percent
Global Mean-
79.2
NP ß
ß
ß
:.
-
;....?,., ½,
3O
EQ -30 -60
8P 180
120 W
15.0 10.0
b)
60 W
25.0 20.0
35.0 30.0
0
45.0 40.0
60 E
55.0 50.0
120 E
65.0 60.0
75.0 70.0
85.0 80.0
180
95.0 90.0
•180 inWaterVaporat GroundLevel %0 V-SMOW
Global Mean-17.0
NP
c)
•180ofCO2 inLeaves %0 PDB-CO2
Global Mean-
3.2
NP
..
.
30
e
t80
120 W
-50 -6.5
-3.5
1.0 -.5
.. •.,,
....... ..?.. ......,,*
•,•
",.•...•:.:,.,•-::-'.• \ '-- ':' '
60 W
-2.0
.•,,,
0
4.0 2.5
60 E
7.0 5.5
10.0 8.5
120 E
13.0 11.5
16.0 14.5
180
19.0 17.5
Plate 2. (a) Relativehumidityat theleafsurface(annualaverage)in thephotosynthesis modelSiB2coupled with the CSU climatemodel.Becauseof planttranspiration the relativehumidityat the leaf surfaceis higher
thanin thefreeatmosphere abovethecanopy. (b) Annualmean•80 of atmospheric watervaporat ground levelin the NASA GISSisotopicmodel[afterJouzelet al., 1987].The vaporphaseis isotopically depletedby ---10%•with respectto meteoricwaterdueto the isotopicfractionation resultingfrom in-cloudcondensation
processes. (c) Annualmeant5•80of CO2in leaves, thatis,in isotopic equilibrium equilibrated withevaporatingleafwater.Thiscorresponds to the t5•80whichinfluences theatmosphere, mediated bythephotosynthesis flux.
5866
CIAIS ET AL.' STUDY OF •180 IN ATMOSPHERICCO2,1
•180 in Ocean SurfaceWater
a)
%0 V-SMOW
Global Mean = 0.3
NP 6O
3O
EQ -3O
-6O
SP 180
120 W -4.6 •, .... %,,.,,.
-5.0
60 W
-3.8
0
-3.0
-2.2
60 E
-1.4
-.6
120 E .2
180
1.0
1.8
.•.•..,•:-'
-4.2
-3.4
-2.6
-1.8
-1.0
-.2
.6
1.4
•100inCO2 attheOceanSurface
b)
%o V-PDB-CO2
Global Mean = 1.8
NP .............................................................
•...................... :--................ ?:: ..............
30
EQF-••---------------j,, "•--[-•-• -60
_[
SP 180
120 W
"
•-t..,t,,=,...•
.,, -.
-8.6
.., .x\•
l
... ...
60 W
-6.8
-5.0
i ,.
0
-3.2
60 E
-1.4
.4
120 E
2.2
180
4.0
5.8
............. ..•"'"% ;, ,....•.•.•... '"Y'•"%:•q";"'g•" :i:..,:.i•.•';,'..::
'"" '""•' 'i• ..
..........ß: ....... ,,..,
-9.5
-7.7
-5.9
-4.1
-2.3
-.5
1.3
3.1
4.9
Plate 3. (a) Oceansurface water&]aOregressed aftersalinity. The decrease nearthe icesheetsandin the riversestuaries is dueto the inputto theoceans of freshwater depletedin •80. (b) Isotopiccomposition of dissolvedCO2 emitted to the atmospherethroughair-seaexchangeprocesses, assumingfull isotopicequilibrium of CO2 with seawater.
CIAIS ET AL.' STUDY OF 8•80 IN ATMOSPHERIC CO2,1
5867
initiallyproposedbyCraigandGordon[1965] outlinedfor soils,thisisdueto theeffectof temperature onO•eq piricalregression and further update by J. C. Duplessy(personalcommunication, 1994)'
and to a lesser extent on a w L--vap'
8• = ai + a2S
5. Exchange of •80 With the Ocean The net CO2 flux betweenthe ocean and atmosphereis givenby Fo = -Fao + Foa= KexApCO2
(•3)
(17)
where S is sea surfacesalinityin gramsper kilogram [Levitus, 1982]. The empiricalvalue of the 8o w versusS linear slope,a 2 =
0.5%o g- 1kg.Thevalueof the intercept a • = - 16.75%o is
w where Fao (Foa) is the one-wayflux of CO2 from (to) the determinedsoasto yield a meanvalueof 0%0 VSMOW for 8o averaged over the world oceans between 60øS and 60øN, exatmosphere,Kex is the air-seagas exchangecoefficient,and ApCO2 is the differencein partial pressureof CO2 between cluding polar oceanswhich deviate significantlyfrom the ocean and atmosphere.
The air-seagasexchangecoefficientKex is taken from the stabilitydependenttheoreticalformulationof Erickson[1993]. The field of ApCO2 is calculatedby the oceangeneralcirculation model HAMOCC (Max Planck Institute, Hamburg) which includesa parameterizationof biologicalprocesses in the ocean[Maier-Reimer, 1993;K. Kurz andE. Maier-Reimer, Geochemicalcyclesin an ocean general circulationmodel: Planktonsuccession and seasonalpCO2, submittedto Global Biogeochemical Cycles,1996].Note that the ApCO2 fieldsare for the preindustrialera,whichis not consistent with our sim-
VSMOW value. Plate 3a indicatesthat 8o w takes lower values wherelarge amountsof freshwaterare deliveredto the ocean,
because continental freshwater isdepleted in 180withrespect to seawaterbythe isotopicdistillationof moistair movingfrom the oceansto the continents.The isotopiccompositionof the oceansurfaceis thusdepletedby about 1%o at high latitudes around Antarctica
and Greenland
because of the massive dis-
chargeof icebergsand in the estuariesof the largestrivers.
Plate3b shows8180of CO2 in isotopicequilibrium with oceanwater, 80. The temperaturedependenceof the equilib-
riumfractionation factorO•eq hastheeffectof increasing 8o at ulationof today's180 cycle.However,the isotopic fluxpro- high latitudesby a few per mil, whichopposesthe latitudinal
variation of 80. The result is an overall increaseof 8o as a atmospheric 8180value(seebelow)sothatat thisstage, using function of latitude, with maximum values of 5-6%0 near the a preindustria! ApCO2 fieldintroduces onlya verysmallbias. seaice marginaroundGreenlandand Antarctica. For consistency with our atmospherictransportmodel, the ocean-atmosphere CO2 fluxesare maskedover regionscov- 6. Anthropogenic Emissions: Fossil Fuels ered by sea ice. and Biomass Burning Regardingthe isotopicfluxes,we havemadethe assumption Carbon dioxide derived from the combustionof hydrogenthat dissolvedCO2 is in isotopic equilibrium with seawater bound carbon bears an isotopiclabel of -17%o PDB-CO2, accordingto reaction(1). We accountfor no catalyticprocess whichcorresponds to the isotopicvalueof atmosphericoxygen that could yield isotopicequilibrationduring a short contact [Kroopnick and Craig,1972].Anthropogenicfluxesof the isobetween atmosphericCO2 and ocean water. Excludingthe topicspecies CO180arethusproportional to theCO2fluxes.
portionalto the net oceanflux hasonly a smalleffecton the
possibilityof rapid hydrationof CO2 with a time constant shorter than the crossingof the diffusivefilm at the air-sea interfaceis supportedby the fact that no evidencefor CA catalysishasbeenfound so far in the ocean.The net air-sea
For fossilCO2 emissions we usedthe estimatesof Marland et al. [1985],distributedaccordingto populationdensityby Fung et al. [1987].For biomassburningemissions, we have useda compilationof observationaldata which include forest and fluxof C18OOhasanexpression similar to thatforisotope 13C savannaburning(seasonal)aswell as agriculturalwastesand [Tanset al., 1993;Ciaiset al., 1995]and is givenby fuel wood burning(annuallyconstant)[Hao and Liu, 1994]. 18Fo= _ otwR aFao+ otwRoFoa (14) We needthe grossfluxof CO2 resultingfrom biomassburning,
Fbur,for calculating 8180in the atmosphere, not thenetde-
where aw is fractionationassociated with CO2 diffusionat the
forestationfluxwhichis significantly lowersinceit includesthe
air-seainterface [Vogel etal., 1970]andRois 180/160ratioof uptakeof CO2 dueto regrowthof burnedecosystems [HoughdissolvedCO2. Here (14) can be rewrittenin the form
•SFo= awRaFo + aw(Ro- Ra)Foa
ton et al., 1987].Conceptually, regrowthshouldbe treatedfor (15)
The left-handterm of (14) is an isotopic"equilibrium"flux
8180 as an additionalcomponent of the leaf exchange flux (linkedto GPP), butwe neglectedit, first,becauseit is a small flux comparedto the natural componentsFLa and FaL and,
second,becauseit hasa minor isotopicdisequilibriumwith the 8180of leaf CO2is in the range0-4%0 in the sinceit is proportionalto the isotopicratioR a.The right-hand atmosphere: flux F o discorrespondsto an isotopic"disequilibrium"flux tropicscomparedto -17%o when plantsare burned: which can be interpretedas a tendencytoward local isotopic
Foeq,which hardly influences the8180of atmospheric CO2
balance between atmospheric CO2and8180in dissolved CO2. By definition,the isotopicratioR o of CO2 in isotopicequilibrium with water is givenby
Ro= O•eq(ro) Rj
(16)
whereToisseasurface temperature andRo wis180/160ratioof
•8Ffos = RfFfos
(18)
18Fbu r -- gfFbu r
(19)
whereRf istheisotopic ratioof CO2produced bycombustion (5 = -17%o)
surface waters.
7. GlobalBudgetof •80 in Atmospheric CO2
We have calculatedRo w in a mannersimilarto that of Farquharet al. [1993].Ro w is a functionof salinityusingthe em-
Before couplingthe fluxesas calculatedaboveto a threedimensionalatmospherictransportmodel for calculatingthe
5868
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERICCO2,1
atmospheric 8•80, it isusefulto testthevalueswe havedeter- dependent gastransfer velocities. The 8•80 of CO2emittedby minedfor the modelparametersby closingthe globalbudget. the combustion of hydrogen-bound carbonis isotopically deTheglobalmeanvalueof 8•80 in theatmosphere, 8a,isinflu- pletedwith respectto the meanatmospheric valueandhasthe encedby soils,vegetation,andair-seaexchange (anthropogen- sameisotopiccharacteras atmospheric 02 (-17%o). We ac-
ic emissions are omitted).The annualmeantrendof 8•sOin
count for fossil fuel industrial emissions in the northern hemi-
atmospheric CO2 is closeto zero [Franceyand Tans,1987]and sphere and biomassburning in the tropics.There is some it is givenby the followingexpression to a goodapproximation inconsistency in assembling fields generatedby differentcli[Farquharet al., 1993]: matemodelsto infer the terrestrialisotopicexchange. Ideally, the isotopiccompositionofwater and CO2wouldbe calculated dSa simultaneously withina singleGCM, with fullyinteractiveisod-•-= 0 (20) topic hydrology,physiology,and photosynthesis.Unfortunately,sucha modeldoesnot existat present. d8 a The oxygenisotopefluxeshave been tested againstthe --[Foa(8 o - 8a) -4-•w(Fao- Foa) dt
Ca
globaltrendin atmospheric 8•80,whichisobserved to beclose
+ Fsoi•s(8s-8a-4-Es)-1 t-ZIAa]
to zero. This conditioncanbe met if the globalaveragefractionation at the soil-air interface is of -5%o, a value smaller
where /•a standsfor the discrimination of oxygenisotopeby conversion
in leaves:
Cc
/•a----8d -4Ca_C•(SL - 8a)
than the one inferredbyFarquharet al. [1993]in an independentcalculation(-7.6%o), althoughit lieswithin a physically acceptable range.Overall,the oceanicandanthropogenic contributionsare relativelyminor comparedto the isotopicex-
(21) changewiththeterrestrialbiota.We confirmthefactthatleaf
The two mostimportanttermsin (20) are thoserelativeto leaf and to soil isotopicexchange,whereasthe oceanflux is relativelyminor. Global numbersfor GPP, discrimination,and other parametersbear a large uncertaintyso that there is not a uniquesolutionto (20) yieldingd Sa/dt = 0. Alternatively, we solved(20) for es,the diffusivefractionationof CO2 respiredby soils.Usingthe valuesgivenin Table 1 for the global quantitieswhich appear in (20), we infer es = -5%0. This working value of % is significantlylower than the value of -8.8%o corresponding to strictlymoleculardiffusion.Possibly, turbulentdiffusionplaysan importantrole in transferringCO2 from the soilsurfaceto the atmosphere, whichwouldlower Another explanationfor the relativelylow value of es that we infer by solving(20) wouldbe that the productionof CO2 in soilsoccursmainlynear the surface,whichwould diminishthe influenceof the diffusivefractionationin the isotopiccomposition of CO2 emitted by soils as shownby Hesterberg and Siegentha[er [1991]in the caseof an exponentialdecreasein the CO2 productionat depth. Farquharet al. [1993]infer that •, = -7.6%0, a valuecloser to moleculardiffusion, from a budgetequationsimilarto (20). Thisis mostlybecausetheyemploya highervalueof the global discrimination than the onewe establishin thispaper.Nevertheless,althoughitsvalueisplausible,the globalsignificance of •s = -5%0 awaitsfurtherexplanations, and it shouldbe consideredas a tuningthat we applyto the globalbudgetso as to yield a zero long-termtrend in
exchange globally enriches in •sOtheatmospheric CO2reservoir,whereassoilexchange hasthe oppositerole. Apart from the globalbudget,we expectthat the geographical differences in the isotopicfluxeshavean influenceon the spatialdistribu-
tionof 8•sOin atmospheric CO2.In a companion paper[Ciais et al., this issue]describinga three-dimensional tracer simulation,we providea moredetailedassessment of the respective role of eachreservoir,with specialemphasis onvegetationand soils.
AppendixA: Groundwater •80 InferredFrom the Soil Hydrology in the SiB2 Model We detailin the following howthe 8•sOof waterin soilsis obtained fromthe 8•80 of meteoric waterusingtheparameters of soil hydrologyin SiB2. The soil columnis dividedinto
threelayers(FigureA1). The surfacelayerreceivesprecipitation and loseswater throughsurfacerunoff,evaporation,and infiltrationto the intermediatelayer.The intermediatelayer corresponds to the rootingzoneof plants,in whichplantstake up water to transpireto the atmosphereand hasno runoff.The
deep soil layer receiveswater by infiltrationfrom the layer aboveand losesit by deeprunoff.Only evaporationfractionates the heavyisotopeof water; all other fluxesconservethe
isotopicratios.The isotopiceffectsof the interceptionof precipitationby the canopyare neglected. In Figure A1,
w•,(Rj,8•') integrated watercontent(180/160,8180)of the surfacelayer (mm);
8.
Conclusions We have calculated the surface fluxes that control the 8•80
in atmosphericCO2. The mostcrucialassumption is that CO2 exchanges isotopically to fully reachisotopicequilibriumwith water availablein leaves(becauseof the presenceof carbonic anhydrase) andin surfacesoil.The validityof thisassumption shouldbe furtherinvestigated throughlaboratoryexperiments. The calculation of terrestrial carbon fluxes is based on the SiB2
photosynthesis model coupledwith the CSU GCM. The isotopic compositionof meteoric water comesfrom the NASA
GISSclimatemodel.The air-sea8•80 exchange is calculated usingApCO2from the HAMOCC oceanmodeland stability-
wi(R•',•") integrated watercontent(18Ofi60,8•80) of the intermediate(rooting)layer;
Wd(R•',8•') integrated watercontent(18Ofi60,8•SO)of the deep(recharge)layer (mm);
a•_va p fractionation of H2180fortheliquid-vapor phasetransition, equalto (1 + (•_vap/1000); P
precipitationreachingthe groundsurface
(mmd-•); E evaporation fromtheground surface (mmd-•); T uptakeof water by rootsequalstranspiration
byplants(mmd-•); Fsi infiltration of water from surfaceto
intermediate soillayer(mmd-•);
CIAIS ET AL.' STUDY OF 8•80 IN ATMOSPHERICCO2,1
where #s is the stomatalconductanceto water vapor,A is the net assimilationrate of CO2,h and Cs are the relativehumidity and mixingratio of CO2 at the leaf surface(for simplicity,Cs is taken equal to the atmosphericvalue Ca), p is the atmosphericsurfacepressure,and m and b are empiricallyderived
i
I I
w
surface
• Ws, ,5s layer Fsi• wi'•t) layer wdeep Fid• Wd, •d layer \
. •
5869
Ms
parameters.
The net assimilationrate A is modeled as limited by the kinetics of the carboxylationenzyme Rubisco, by electron transport (a seriesof reactionsthat take place when green plant cells are illuminated with visible radiation), and by buildup of the sugarsand starchesthat are the end productsof photosynthesis. Farquhare! al. [1980]useda simpleminimum
intermediate
Vd
of the three limits to calculate
the net carbon assimilation
rate
Figure A1. Fluxesof H:O in soilsas calculatedby the SiB2 A = min (Loc,LOe, LOs) -- 'q)•d (B2) model includingthree soil compartments.The isotopiccompositionof groundwateris calculatedoff-linefrom thesefluxes: (often solid lines are fluxesthat take placewith no isotopicfraction- where % is the carbon-limitedrate of photosynthesis ation, and the dashedline includesthe specificfractionationof referred to as Rubisco-limited,sincethe rate is determinedby Rubiscoenzymekinetics), LOe is the rate limited by electron water isotopesduring evaporation. transport(light-limited),Los is the endproduct-limited(or sink-
limited)rate,and•)•dis the rate of carbonlossfromthe canopy due to "dark" respiration.SiB2 usesa similar approachbut replacesthe simpleminimumin (B2) with a smoothedfunction Fid infiltration of water from intermediateto deep to avoid abrupt transitionsfrom one limitation to another soillayer(mmd-l); [Collatz et al., 1991]. The Rubisco-limitedand light-limited Vs surface runoff(mmd-l); assimilationrates are calculatedfrom enzymekineticsmodels Vd deeprunoff(mmd-l). developedby Collatzet al. [1991]. For C3 vegetationthe sinkThe massbalanceof H:O andH:180 in eachlayeriswritten limited rate is parameterized as a simple fraction of the Rubisco activity, and for C4 vegetation, Losrefers to PEPdws= P - E - Fsi- Vs (A1) carboxylaselimitation accordingto the model of Collatzet al. [1992]. Leaf respiration•)•d is parameterizedaccordingto dwi = Fsi- T- Fid (A2) Rubiscoactivityand canopytemperature. dWd= Fid- [?d (A3) All three photosyntheticrates and the leaf respirationare scaledby nondimensional parameters(fHOT, fCOLD,f-I,, fRH, d(R•ws)= R½'P- o•_vapR•E -- R•'Fsi- R•'I/s (A4) and fFRZ) representingenvironmentalstressesdue to excessivelyhigh or low canopytemperature,drought,low relative d(R•wi) = R•Fsi- R•VT- R•VFid (A5) humidity, and frozen soils. The leaf-level assimilationrate, d(RjWd) = g•VFid- gjVd (m6) stomatalconductance,and other physiologicalparametersare scaledto the canopyusing an assumedoptimal relationship Using d(Rw) = wdR + Rdw and the definitionR = (1 + betweenleaf nitrogen(and henceRubisco)and the time-mean 8/1000)RvsMow, we substitute(A1)-(A3) into (A4)-(A6) to profile of photosynthetically active radiation in the canopy. obtain the set of differentialequations(A7)-(A9) describing Details of the parameterizationare presentedby Sellerset al. thevariationin 8180of waterin eachsoillayer. [1996a,AppendixC]. The partial pressureof CO2 in the leaf interior (Ci) is d87: w•-l[p(8•- 8•) + E(8L_vap) ] (AT)
dS•'= wi-•F•i(8• ' - 87)
(A8)
da•'= wjlFid(8•"- 8•')
(A9)
•
A-O•d
Cc CirCa
We integrate these equations numerically to calculate the monthlymeanvaluesof 8•TM and 8•vwhichare, respectively, used to expressthe isotopiccompositionof CO: in soils and in leaves(sections3 and 4).
chloroplast • •r..j surface
layer
Appendix B: Biospheric CO2 Fluxes in the SiB2 Model
B.1.
intermediate
Photosynthesis
FollowingCollatzet al. [1991,1992]andBall [1988],stomatal conductance,the carbonassimilationrate, and the CO: concentrationat the leaf surfacesare assumedto be related by Ah
g•=m-•-•-• +b
(B1)
(root) layer
Figure B1. Resistanceto the diffusionof CO: from the canopy to the leaf chloroplast.During the assimilationthere is a gradient in the CO: partial pressurefrom Ca to Co.
5870
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERICCO2,1
SiB2modelare similarto thosefoundin previousstudies,and the seasonaland diurnalcyclesof net carbonflux to the atmoCO2 efilux due to soil respiration.The partial pressurein the spherecomparefavorablywith the limited field observations chloroplastCc, is slightlysmallerthan Ci due to the resistive available[Denninge! al., 1996].When thesefluxeswereusedin path of CO2 acrossthe mesophyllcell [see,e.g.,Farquharetal., the CSU GCM to calculate the full three-dimensional concentration field of atmosphericCO2, the resultscomparedvery 1993]but we make the approximationthat favorably with the data of the NOAA/CMDL flask station Cc•Ci (g3) network [Denning,1994;Denningand Randall, 1996]. diagnosedusing a resistancenetwork (Figure B1) from the stomatal conductance,the net assimilationrate, and the rate of
B.2.
Respiration
The relativeintensityof soil respiration,denotedby •*, is References diagnosedfrom soil moisture and soil temperature at each Allison, C. E., R. J. Francey,and H. A. Meijer, Recommendationsfor the reportingof stableisotopemeasurementsof carbonand oxygen model time step followingthe method used by Raich et al. in CO2 gas,in References and Intercomparison Materialsfor Stable [1991]in the terrestrialecosystem model.The soilrespiration Isotopesof LightElements,Proceedings of a Consultants Meetingheld diagnostic•* is definedas in Vienna, 1-3 December1993 IAEA-TECDOC-825, pp. 155-162,
•* = 2.0Of(w)
(B4)
where B
f(w) = 0.2 + Wsa t
(B5)
(W ....
Int. At. Energy Agency, Vienna, 1995. Arakawa, A., and V. R. Lamb, Computationaldesignof the basic dynamicalprocesses of the UCLA generalcirculationmodel,Methods Cornput.Phys.,17, 173-265, 1977. Arakawa, A., and V. R. Lamb, A potential enstrophyand energy conservingschemefor the shallowwater equations,Mon. Weather Rev., 109, 18-36, 1981.
Baertchi,I. H., andW. C. Macklin,Absolute180 contentof standard
B= Wopt zm _ 100zm]
mean ocean water, Earth Planet. Sci. Lett., 31, 341-344, 1965.
Ball, J. T., An analysisof stomatalconductance,Ph.D. thesis,89 pp.,
In the temperatureresponsefunction,Qt = (T - 298)/10. The temperatureusedto defineQt in (B3) is thewarmerof the surfacesoil temperatureand the deep soil temperature.The variablew in (B4) is the fractionof the pore spaceoccupiedby water in the root zone (middle layer) of the soil.The param-
etersWsat, Wopt, andZmareprescribed according to soiltexture
Stanford Univ., Stanford, Calif., 1988. Bariac, T., J. Gonzalez-Dunia, F. Tardieu, D. Tessier, and A. Mariotti,
Variabilit6spatialedela composition isotopique del'eau(180,2H) au sein des organesdes plantesa6riennes,1, Approche en conditions contr616es,Chem. Geol., 115, 307-315, 1994a.
Bariac, T., J. Gonzalez-Dunia,N. Katerji, O. B6thenod,J. M. Bertolini, and A. Mariotti, Variabilit6 spatiale de la compositionisoto-
piquedel'eau(180,2H) dansle continuum sol-plante-atmosphere,
usingvaluessuggested by Raichet al. [1991].Soil respirationis
2, Approche en conditionsnaturelies,Chem. Geol., 115, 317-333,
a maximum for somevalueWopt of soilmoisture, andrespiration is lessefficientunder very dry or very wet conditions. A dimensionless monthly mean soil respirationrate is defined as
1994b.
Brenninkmeier,C. A.M., P. Kraft, and W. G. Mook, Oxygenisotope fractionationBetween CO2 and H20 , IsotopeGeosci.,1, 181-190, 1983.
Ciais, P., P. P. Tans, J. W. White, M. Trolier, R. Francey,J. A. Berry, D. Randall, P. J. Sellers,J. G. Collatz, and D. S. Schimel, Partition-
?(t) = • •*(t)At
(B6)
1 year
ingof oceanandlanduptakeof CO2asinferredby 813Cmeasurementsfrom the NOAA Climate Monitoring and DiagnosticLaboratoryglobalair samplingnetwork,J. Geophys. Res.,100, 5051-5070, 1995.
synthesis studyof 8•80 in atmowhere the overbar indicatesthe monthly mean. The flux of Ciais,P., et al.,A three-dimensional sphericCO2, 2, Simulationswith the TM2 transportmodel,J. GeoCO2 from the soil due to respirationis computedfrom this phys.Res., this issue. dimensionless rate by assuminga local steadystatefor carbon Collatz, G. J., J. T. Ball, C. Grivet, and J. A. Berry, Physiologicaland environmentalregulationof stomatalconductance,photosynthesis, storagein terrestrialecosystems on an annual basis,that is, and transpiration:A modelthat includesa laminarboundarylayer, that the annualsumof respirationlossis equal to the annual Agric.For. Meteorol.,54, 107-136, 1991. net carbonassimilation(ANA), definedas
Collatz, G. J., M. Ribas-Carbo,and J. A. Berry, Coupled photosynthesis-stomatal conductancemodel for leavesof C4 plants,Aust.J. ANA= • ,•(t)At (B7) Plant Physiol.,19, 519-538, 1992. I year Conway,T. J., P. P. Tans,L. S. Waterman,K. W. Thoning,D. R. Kitzis, K. A. Masarie,and N. Zhang,Evidencefor interannualvariabilityof whereA (t) is the monthlymean net assimilation.The flux of the carboncyclefrom the National Oceanicand AtmosphericAd(202 from the ecosystem due to respirationis calculatedfrom ministration/Climate Monitoringand DiagnosticLaboratoryGlobal air samplingnetwork,J. Geophys.Res.,99, 22,831-22,855,1994. ?(t) accordingto Craig,H., and A. Gordon,Deuteriumand oxygen-18variationsin the ocean and the marine atmosphere,in StableIsotopesin Oceanic 2•* (t) = ?(t)ANA (B8) Studiesand Paleotemperatures, Lab. Geol. and Nuclear Sci., Pisa, Italy, 1965. sothat the net annualfluxof CO2 from everygridpoint is zero. The model was integratedfor 10 yearson a low-resolution Denning, A. S., Investigationsof the transport,sources,and sinksof atmosphericCO2 usinga generalcirculationmodel,Atmos.Sci.Pap.
grid (7.2ølatitude x 9ølongitude)to allow soilmoisturefields 564, Colo. State Univ., Fort Collins, 1994. to equilibratewith the simulatedclimate.After this "spin-up" Denning,A. S., and D. A. Randall, Simulationsof terrestrialcarbon metabolismand atmosphericCO2 in a generalcirculationmodel,2, periodthe modelwasintegratedfor an additional4 yearson a SimulatedCO2 concentrations, Tellus,in press,1996. 4ø x 5øgridwith 17verticallayersand a time stepof 6 min. The
fieldsusedto drivethe•80 calculation weremonthly means for the final year of the simulation.The spatialand seasonaldistributionof net primaryproductivity(NPP) simulatedby the
Denning,A. S., J. G. Collatz, C. Zhang, D. A. Randall, J. A. Berry,P. J. Sellers,and S.C. Wofsy,Simulationsof terrestrialcarbonmetabolism and atmosphericCO2 in a generalcirculationmodel, 1, Surface carbonfluxes,Tellus,in press,1996.
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERIC CO2,1 D6rr, H., and K. O. Munnich, Annual variation in soil respirationin selectedareasof the temperatezone, Tellus,Set. B, 39, 114-121,
5871
seasonalcycle for present-day conditions,J. Geophys.Res., 92, 14,739-14,760, 1987.
Keeling, C. D., T. P. Whorf, M. Wahlen, and J. Van der Plicht, Interannual extremesin the rate of rise of atmosphericcarbon Enting, I. G., C. M. Trudinger,R. J. Francey,and H. Granek, Synthesis dioxide since 1980, Nature, 375, 666-670, 1995. inversion of atmosphericCO2 using the GISS tracer transport model,Div. Atmos.Res. Tech.Pap.Aust. C.S.I.R.O.,29, 1-44, 1993. Keeling, R., The atmosphericoxygencycle:The oxygenisotopesof atmosphericCO2 and 02 and the O2/N2 ratio, U.S. Natl. Rep. Int. Enting, I. G., C. M. Trudinger,and R. J. Francey,A synthesis inversion Union Geod. Geophys.1991-1994, Rev. Geophys.,33, 1253-1262, of theconcentration and8•3Cof atmospheric CO2,Tellus, Set.B, 47, 1987.
35-52, 1995.
1995.
Erickson,D. J., III., A stabilitydependenttheoryfor air-seaexchange, Kroopnick,P., and H. Craig, Atmosphericoxygen:Isotopiccomposition and solubilityfractionation,Science,175, 54-55, 1972. J. Geophys.Res.,98, 8471-8488, 1993. Atlas of the World Ocean,Prof Pap., 173, Farquhar,G. D., S. von Caemmerer,and J. A. Berry, A biochemical Levitus, S., Climatological U.S. Gov. Print. Off., 1982. model of photosynthetic CO2 assimilationin C3 plants,Planta, 149, Maier-Reimer, E., Geochemicalcyclesin an oceangeneralcirculation 78-90, 1980. model:Preindustrialtracer distribution,GlobalBiogeochem. Cycles, Farquhar, G. D., K. T. Hubick, A. G. Condon, and R. A. Richards, 7, 645-677, 1993. Carbonisotopefractionationand plant water-useefficiency,in StableIsotopesin EcologicalResearch,editedby P. W. Rundel et al., pp. Marland, G., R. M. Rotty, and N. L. Treat, CO2 from fossil fuel burning:Global distributionof emissions, Tellus,Set.B, 37, 243-258, 21-40, Springer-Verlag,New York, 1989. 1985. Farquhar, G. D., J. Lloyd, J. A. Taylor, L. B. Flanagan,J.P. Syvertsen, of H2•80 in gases, J. Chem.Phys., K. T. Hubick, S.C. Wong, and R. Ehleringer,Vegetation effectson Merlivat,L., Moleculardiffusivities 69, 1864-2871, 1978. the isotopecompositionof oxygenin atmosphericCO2,Nature,363, Raich, J. W., E. B. Rastetter,J. M. Melillo, D. W. Kicklighter,P. A. 439-443, 1993. Steudler,and B. J. Peterson,Potentialnet primaryproductivityof F6rstel, H., A. Putral, G. Schleser,and H. Leith, The world pattern of SouthAmerica:Applicationof a globalmodel,Ecol.Appl., 1,399oxygen-18in rainwater and its importancein understandingthe
429, 1991. biogeochemicaloxygencycle,in IsotopeRatiosas PollutantSource Randall, D. A., and D.-M. Pan, Implementation of the Arakawaand BehaviorIndicators,pp. 323-344, Int. At. Energy Agency,ViSchubertparameterizationwith a prognosticclosure,in TheRepreenna, 1975. sentationof CumulusConvectionin NumericalModels, edited by Fowler, L. A., and D. A. Randall, Liquid and ice cloudmicrophysics in K. Emanuel and D. Raymond,pp. 137-144, Am. Meteorol. Soc., the CSU generalcirculationmodel, II, Impact on cloudiness,the
Earth's radiationbudget,and the general circulationof the atmosphere,J. Clim., 9, 530-560, 1995a. Fowler,L. A., and D. A. Randall,Liquid andice cloudmicrophysics in the CSU generalcirculationmodel, III, Sensitivityto modelingassumptions,J. Clim., 9, 561-586, 1995b. Fowler,L. A., D. A. Randall, and S. A. Rutledge,Liquid and ice cloud microphysics in the CSU general circulationmodel, I, Model descriptionand simulatedmicrophysicalprocesses, J. Clim., 9, 486529, 1995.
Francey, R. J., and P. P. Tans, Latitudinal variation in oxygen-18of atmosphericCO2, Nature, 495-497, 1987. Francey,R. J., C. E. Allison, I. G. Enting, J. W. C. White, M. Trolier, and P. P. Tans,Changesin the oceanicand terrestrialcarbonuptake since 1982, Nature, 373, 326-330, 1995. Friedman,I., andJ. R. O'Neill, Compilationof stableisotopefractionation factorsof geochemicalinterest, U.S. Geol. Sum Prof. Pap.,
Boston, Mass., 1993. Randall, D. A., Harshvardhan, D. A. Dazlich, and T. G. Corsetti,
Interactionsamongradiation,convection,and large-scaledynamics in a generalcirculationmodel,J. Atmos.Sci., 46, 1943-1970, 1989. Randall, D. A., Harshvardhan,and D. A. Dazlich, Diurnal variability of the hydrologiccyclein a generalcirculationmodel,J.Atmos.Sci., 48, 40-62, 1991.
Randall, D. A., Q. Shao, and C.-H. Moeng, A second-orderbulk boundary-layermodel.J. Atmos. Sci., 49, 1903-1923, 1992. Randall,D. R., et al., A revisedland-surface parameterization(SiB2) for GCMs, 3, The greeningof the ColoradoStateUniversitygeneral circulation model, J. Clim., 9, 738-763, 1996. Sellers,P. J., Y. Mintz, Y. C. Sud,andA. Dalcher,A simplebiosphere model(SiB) for usewithingeneralcirculationmodels,J.Atmos.Sci., 43, 505-531, 1986.
Sellers,P. J., J. A. Berry, G. J. Collatz, C. B. Field, and F. G. Hall, Canopyreflectance,photosynthesis, and transpiration,III, A reanal440-KK, 1977. ysisusingenzymekinetics--Electrontransfermodelsof leaf physiFung, I. Y., C. J. Tucker, and K. C. Prentice,Applicationof advanced ology,RemoteSens.Environ.,42, 1-20, 1992a. very high resolutionradiometer vegetationindex to study atmo- Sellers, P. J., M.D. Heiser, and F. G. Hall, Relations between surface sphere-biosphere exchangeof CO2,J. Geophys. Res.,92, 2999-3013, conductance andspectral vegetation indices at intermediate (100m2 1987.
Hao, W.-M., and M.-H. Liu, Spatial and temporal distribution of tropical biomassburning,Global Biogeochem.Cycles,8, 495-503, 1994.
Harshvardhan,R., D. A. Davies, D. Randall, and T. G. Corsetti, A fast
radiationparameterizationfor generalcirculationmodels,J. Geophys.Res.,92, 1009-1016, 1987. Hesterberg,R., and U. Siegenthaler,Productionand stable isotopic compositionof CO 2 in a soil near Bern, Switzerland,Tellus,Set.B, 43, 197-205, 1991.
to 15km2) lengthscales, J. Geophys. Res.,97, 19,033-19,059, 1992b. Sellers,P. J., D. R. Randall, J. A. Collatz,J. A. Berry, C. B. Field, D. A. Dazlich, C. Zhang, and L. Bounoua,A revisedland-surfaceparameterization(SiB2) for atmosphericGCMs, 1, Model formulation, J. Clim., 9, 676-705, 1996a. Sellers,P. J., S. O. Los, C. J. Tucker, C. O. Justice,D. A. Dazlich, G. J. Collatz,and D. A. Randall,A revisedland surfaceparameterization
(SiB2) for atmospheric GCMs, 2, The generationof globalfieldsof terrestrial biophysicalparametersfrom satellite data, J. Clim., 9, 706-737, 1996b.
Houghton,R. A., et al., The flux of carbonfrom terrestrialecosystems Silverman,D. N., CarbonicAnhydrase:Oxygen-18exchangecatalyzed to the atmospherein 1980 due to changesin land use:Geographic by an enzymewith rate-contributingproton transfersteps,Methods distributionof the globalflux, Tellus,Set.B, 39, 122-139, 1987. Enzymol., 87, 733-753, 1982. Hut, G., Report to the Director General on the consultants'group Suarez, M. J., A. Arakawa, and D. A. Randall, Parameterization of the meetingon stableisotopereferencesamplesfor geochemicaland planetary boundarylayer in the UCLA general circulationmodel: hydrologicalinvestigations,report, 41 pp., Int. Atomic Energ. AsFormulation and results,Mon. WeatherRev., 111, 2224-2243, 1983. soc., Vienna, 1987. Tans, P. P., R. J. Francey,and G. I. Pearman,Large scalevariationsin InternationalAtomicEnergyAgency(IAEA), Statisticaltreatmentof 8•80 of tropospheric CO2,in Baseline 1983-1984,editedby R. J. environmentalisotopedata in precipitation,Tech.Rep.Set.I.A.E.A., Franceyand B. W. Morgan, Dep. of Sci.,Bur. of Meteorol./CSIRO 206, 1-256, 1981.
Jacob,H., and C. Sonntag,An 8-yearrecordof the seasonalvariation
of 2H and •80 in atmospheric watervaporand precipitation at Heidelberg,Germany,Tellus,Set.B, 43, 291-300, 1991. Jouzel, J., G. L. Russell, R. J. Suozzo, R. D. Koster, J. W. C. White,
and W. S. Broecker,Simulations of the HDO and H2•80 atmosphericcyclesusingthe NASA/GISS generalcirculationmodel:The
Div. of Atmos. Res., Melbourne, Victoria, Australia, 1986.
Tans,P. P., J. A. Berry,andR. F. Keeling,Oceanic•3Cdata:A new windowon CO2 uptakeby the oceans,GlobalBiogeochem. Cycles,7, 353-368, 1993.
Vogel, J. C., P.M. Grootes,and W. G. Mook, Isotopicfractionation betweengaseousand dissolvedcarbon dioxide,Z. Phys.,230, 225238, 1970.
5872
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERICCO2, 1
White, J. W. C., The climaticsignificanceof D/H ratios in white pine in the northeasternunited states,331 pp., Ph.D. thesis,Columbia
nent that controls(5180of atmospheric CO2 and 02, PlantCell
Sciences,Colorado State University, Fort Collins, CO 80523-1370. (e-mail: [email protected]) R. J. Francey, Division of AtmosphericResearch,CSIRO, PMB Aspendale,Melbourne, Victoria 3195, Australia. M. Heimann, Max-Planck Institut ffir Meteorologie,Bundestrasse 55, D-20146 Hamburg, Germany. H. A. J. Meijer, CIO, Universityof Groningen,9722 JX Groningen, Netherlands.(e-mail: [email protected]) P. Monfray, Centre des Faibles Radioactivit•s, Bfitiment 709/
Environ., 7, 73-80, 1994.
LMCE, 91191 Gif sur Yvette, France.
Univ., New York, 1983.
White, J. W. C., and S. D. Gedzelman, The isotopiccompositionof atmosphericwater vapor and the concurrentmeteorologicalconditions,J. Geophys.Res.,89, 4937-4939, 1984. Yakir, D., J. A. Berry, L. Giles, and C. B. Osmond,Isotopic heterogeneityof water in transpiringleaves:Identificationof the compo-
J. A. Berry, Department of Plant Biology, Carnegie Institution of Washington, 290 Panama Street, Stanford, CA 94305. (e-mail: [email protected]) P. Ciais, LMCE,
CEA l'Orme des Merisiers, Commissariat h
P. P. Tans, Climate Monitoring and DiagnosticLaboratory,NOAA, ERL 3, 325 Broadway,Boulder, CO 80303. M. Trolier and J. W. C. White, Institute of Arctic and Alpine Research,University of Colorado, Campus Box 450, Boulder, CO 80303.
l'Energie Atomique, Bfitiment 709, Saclay 91191 Gif sur Yvette, France.(e-mail:[email protected]) G. J. Collatz and P. J. Sellers,NASA Goddard SpaceFlight Center, MC 923, BiosphereSciencesBranch,Greenbelt,MD 20771. (ReceivedNovember 18, 1995;revisedJune 25, 1996; A. S. Denning and D. A. Randall, Department of Atmospheric acceptedJuly 12, 1996.)
OF GEOPHYSICAL
RESEARCH,
VOL. 102, NO. D5, PAGES 5857-5872, MARCH
A three-dimensional synthesis studyof 1.
Surface
20, 1997
in atmospheric COz
fluxes
PhilippeCiais,• A. Scott Denning,2 Pieter P. Tans,3 JosephA. Berry,4 David A. Randall, 2 G. James Collatz, s Piers J. Sellers,s James W. C. White, 6
Michael Trolier,3,6Harro A. J. Meijer,7 Roger J. Francey,s Patrick Monfray,9 and Martin Heimann•ø
Abstract.Theisotope•SOin CO2is of particular interestin studying theglobalcarbon cyclebecauseit is sensitiveto the processes by which the global land biosphereabsorbs and respiresCO2. Carbon dioxideand water exchangeisotopicallyboth in leavesand in
soils,andthe•SOcharacter of atmospheric CO2is strongly influenced bythelandbiota, which shouldconstrainthe grossprimary productivityand total respirationof land
ecosystems. In thisstudywecalculate theglobalsurface fluxesof •SOforvegetation and
soilsusingthe SiB2 biospheremodel coupledwith the ColoradoState Universitygeneral circulationmodel. This approachmakesit possibleto use physiologicalvariablesthat are consistentlyweightedby the carbonassimilationrate and integratedthrough the exchangeof rsO and the isotopic vegetation canopy. We alsocalculate theair-sea characterof fossilemissionsand biomassburning.Global mean valuesof the isotopic
exchange witheachreservoir areusedto closetheglobalbudgetof •SOin CO2.Our
resultsconfirm the fact that the land biota exert a dominant control on the (5•SOof the
atmosphericreservoir.At the global scale,exchangewith the canopyproducesan isotopic enrichmentof CO2, whereasexchangewith soilshas the oppositeeffect. vated atmosphericCO2 levels,changingnutrient availability, temperature,and precipitationpatterns. Increasingattentionhasbeen givenrecentlyto the terrestrial Modelsof ecosystem functioninghavebecomeprogressively biospherein controllingatmosphericCO2 levelsbecausethe more processoriented,especiallyregardingthe photosynthetic carbonstoredin the abovegroundbiomassand in soil organic uptake of CO2. A few global mechanisticmodels based on mattercanbe exchangedrapidlywith the atmosphere.It iswell external"climatic"forcing,suchas the incidentsolar flux, the recognizedthat land ecosystems take up and release large water availabilityfor plants,and the temperature,are able to quantitiesof CO2 not only on a daily and seasonaltimescale calculate the gross fluxes of CO2 exchangedbetween land but also in the long term. Severalmodelsof the global bio- ecosystemsand the atmosphere.Figure la givesa schematic sphereon land have been developedwhich simulatethe be- pictureof the cyclingof carbonbetweenplants,soils,and the havior of various ecosystemsand have been used in order to atmosphere.Of particularimportanceto the atmosphericCO2 (A is gross predictthe carbonfluxesexchangedwith the atmosphereun- budget is the uptake of CO2 by photosynthesis (GPP) minusleaf respiration•d) andthe der specificscenariosof future climate changeincludingele- primaryproductivity accompanying ecosystemtotal respiration(•). Respiratory 1.
Introduction
•Laboratoirede Mod61isationdu Climat et de l'Environnement, Commissariath l'Energie Atomique l'Orme des Merisiers, Gif sur Yvette, France.
CO2emissions includeaboveground plantrespiration ^ (•plants) andbelowground rootrespiration B (?•Plants) aswell as heterotrophicsoilrespiration(•soi•s),the total CO2 effiuxfrom soils
2Department of Atmospheric Sciences, ColoradoStateUniversity, being called Fsoils. Over the course of the year, the annual Fort Collins. mean assimilationA is almostentirely compensatedby respi3ClimateMonitoringandDiagnostic Laboratory, NOAA, Boulder, ration emissions. Colorado.
The CO2 biosphericfluxescalculatedby ecosystemmodels can be partially validated againstatmosphericobservations. For instance,a very usefulvalidationis to comparethe seaSNASAGoddardSpaceFlightCenter,Greenbelt,Maryland. 6Institute of ArcticandAlpineResearch andDepartmentof Geo- sonalvariation in atmosphericCO2 simulatedwith givenbiologicalSciences,Universityof Colorado,Boulder. sphericfluxesto the well-documentedobservationalrecordat 7Centrum voorIsotopenOnderzoek, University of Groningen, Gronumeroussitesaroundthe world [Conwayet al., 1994].This is ningen,Netherlands. 8Division of Atmospheric Research, Commonwealth Scientific and commonly done by coupling the calculated field of the net IndustrialResearchOrganisation,Melbourne,Victoria, Australia. ecosystem flux of CO2, the differencebetweenA and •, to an 9CentredesFaiblesRadioactivit6s, Laboratoirede Mod61isation du atmospherictransport model, and comparingthe results to Climat et de l'Environnement, Gif sur Yvette, France. observations. This approachhasprovenvery valuablefor test•øMax-Planck-Institut ffir Meteorologie, Hamburg,Germany. ing the seasonality of net CO2 fluxes[e.g.,Funget al., 1987]but Copyright1997 by the American GeophysicalUnion. doesnot constrainthe grossfluxesof CO2 separately. The globalbudgetof atmosphericCO2 hasalsobeenstudied Paper number 96JD02360. 4Department of PlantBiology,Carnegie Institution of Washington,
Stanford, California.
0148-0227/97/96JD-02360509.00
usingmeasurements of the /5•3Cof atmospheric CO2.The
5857
5858
CIAIS ET AL.: STUDY OF &•80IN ATMOSPHERICCO2,1
FALingoing CO2
FLA retrodiffusion ofCO2
A
•Plants
stem, twigs '•' respiration
Fso,s
•
totalsoil • total respiration
respiration
•lant
!plant belowsoilheterotrophic :ground respiration respiration Figure la. (a) The cyclingof carbonbetweenthe land biosphereand the atmosphere.The ecosystem representedis in equilibriumsincethe annualmeanuptakeof CO2by photosynthesis (A) compensates exactly the total respiratoryloss(•).
methodrelieson theinterpretation of atmospheric 8•3Cvariationsas indicatingnet biosphericfluxes[Tans et al., 1993; Franceyet al., 1995;Keelingetal., 1995;Entinget al., 1993,1995; Ciais et al., 1995]. The methodis limited by uncertaintyconcerningthe influenceof isotopicdisequilibriabetweenatmosphereand surfacereservoirs.Suchdisequilibriacan be transferred to the atmosphereby grossexchangefluxes,evenin the absenceof net exchange. The Earth'svegetationlikely exertsa major influenceon the
For CO> all isotopicvalues are given relative to the standardisotopicratio Vienna Pee Dee belemnite(VPDB)-CO: = 0.002088349077 as recommendedbyAllisonet al. [1995].For H:O we expressisotopicabundancerelative to the standard Vienna SMOW (VSMOW) = 0.00200520[Baertchi andMackfin, 1965]. We must subtract41.47%o to expressVSMOW valuesin the VPDB-CO: scale.This includesa differenceof -30.9%0 between VSMOW and VPDB-calcite [Hut, 1987]
andaccounts forthe•gOfractionation duringCO: evolution at •80/•60 ratioof atmospheric CO2[Keeling, 1995].Francey and 25øCwith 100%phosphoricacid [Friedmanand O'Neill,1977]
Tans [1987] first pointedout that the isotopicexchangewith water in leaves (and possiblysoils) may determine the ob-
betweenVPDB-calcite and VPDB-CO2.
served persistent north-south differences in 1gOof atmospheric 1.2. Climate Variables Used in This Study:
CO2.Farquharet al. [1993]furtherquantifiedthe globalrole of leaf exchangeand calculateda globalatmosphericbudgetof
CSU GCM
and SiB2 Model
The Colorado State University (CSU) generalcirculation model(GCM) is derivedfrom the Universityof California,Los is controlledby the fluxesA and •. We presenthere a syn- Angeles,(UCLA) GCM, whichwasdevelopedat UCLA over thesissimulation of •gOin CO2thatwe compare with atmo- a period of 20 yearsby A. Arakawa and collaborators.A copy sphericmeasurements. In the presentpaper we focuson the of the model was brought to the Goddard Laboratory for mechanisms thatgovernthe•gO/•60ratioin CO:. Specifically,Atmospheresin 1982 and from there to CSU in 1988. Many we havecalculatedon a 4øby 5ø grid the isotopicfluxesasso- changeshavebeen madesincethe modelleft UCLA, including ciated with the terrestrial and oceanic reservoirs, as well as revised parameterizationsof solar and terrestrial radiation with anthropogenicCO: emissions. In the companionpaperby [Harshvardhanet al., 1987], the planetary boundary layer Ciaiset al. [this issue]we haveprescribedthesefluxesin the (PBL) [Randallet al., 1992],cumulusconvection[Randalland processes [Fowleret al., 1995], three-dimensionalatmospherictracer model TM2 and com- Pan, 1993],cloudmicrophysical [Sellerset al., 1986,1992a,b, 1996a, paredthesimulated 8•80 valuesto atmospheric observations.and land-surfaceprocesses b]. Somerecent resultsare presentedby Randall et al. [1989, 1.1. Conventions and Units 1991, 1996], Fowler et al. [1995], and Fowler and Randall [1995a, b]. In this paper, sinkscorrespondto a negativenet flux of The prognosticvariablesof the CSU GCM are potential carbon(CO2 is removedfrom the atmosphere)and sources correspondto a positivenet flux (CO2 is releasedto the atmo- temperature;the horizontal wind components;the surface sphere).Isotopicratiosare expressed in per mil (%0), definedas pressure;the PBL's depth and turbulencekinetic energy;the mixingratio of three phasesof water plus rain and snow;the ( O/ O)sampl e -( )standard temperaturesof the plant canopy,the groundsurface,and the 8180 = 1000 •J/ •JJstandard 18t•/16t•\ deep soil; the water contentsof four abovegroundand three
180in CO2.Specifically, thelgo/•60ratioof atmospheric CO:
1816
180/160
CIAIS ET AL.: STUDY OF 8•80 IN ATMOSPHERICCO2,1 belowgroundmoisturestores;the stomatalconductanceof the plant canopy;and the ice temperatureat land ice and sea ice points.The governingequationsare finite-differenced,using highlyconservative schemes[Arakawaand Lamb, 1977,1981]. The modelis formulatedin termsof a modifiedsigmacoordinate, in which the PBL top is a coordinatesurface,and the PBL itself is identifiedwith the lowestmodel layer [Suarezet al., 1983].The masssourcesand sinksfor the PBL consistof large-scaleconvergence or divergence,turbulententrainment, and the cumulus
mass flux. Turbulent
entrainment
can be
drivenbypositivebuoyancyfluxesor by shearof the meanwind in the surfacelayer or at the PBL top. For vegetatedland pointsthe surfacefluxesof sensibleand
5859
Direct isotopicexchangebetweenCO2 and H20 vapor is excludedbecausethe rate of hydrationis slow (severalminutes)andonlya verysmallfractionof CO2 is dissolved in liquid water at any time [Franceyand Tans, 1987]. However, the enzymecarbonicanhydrase(CA), ubiquitousin plant tissues, catalyzesthe hydration and stronglyacceleratesthe rate of reaction(1) [Silverman,1982].In livingplant tissuesthe isotopic equilibrium between CO2 and H20 is reached quasiinstantaneously. Little is knownof CA activityin soils,but CO2 producedfrom decayingplant tissuesremainsin contactwith soil water for sufficienttime (see below) for reaction(1) to occur and most likely yield full isotopicequilibrationof CO2
withwater,evenin the absence of CA. The •80 of CO2in
latent heat, radiation, moisture, and momentum are deter-
leavesand in soilscan thereforebe predictedby reaction(1),
minedusingthe simplebiosphere(SiB) parameterizationdeveloped by Sellerset al. [1986]. SiB has recentlyundergone substantialmodification[Sellerset al., 1996a,b; Randall et al., 1996] and is now referred to as SiB2. The numberof biomespecificparametershas been reduced,and most are now derived directly from processedsatellite data rather than prescribedfrom the literature. The vegetationcanopyhas been reduced to a single layer. Another major change is in the parameterizationof stomataland canopyconductance [Collatz et al., 1991, 1992;Sellerset al., 1992a,b, 1996a] used in the calculationof the surfaceenergybudget over land. This pa-
provided weknowthe•80 ofwaterreacting withCO2andthe
rameterization
involves
the direct
calculation
of the rate of
carbon assimilationby photosynthesis, making possiblethe calculationof CO2 exchangebetweenthe atmosphereand the terrestrialbiota at the dynamictime step(6 min) of the CSU GCM [Denning,1994;Denninget al., 1996;Dennyand Randall, 1996]. Details of the carbonflux calculationsand their use in isotopicexchangecalculationsare presentedin Appendix B.
2. OxygenIsotope Fractionation BetweenCOz and HzO Of major importancefor the isotopiccompositionof CO2 in the atmosphereis the fact that dissolvedCO2 mayexchangean
•80 atomwith water according to the isotopicequilibrium reaction(1):
COO + H2180 transition, equalto 1 + (•_vap/1000)----R•/RvWap; usedto distinguish a•' kineticfractionation of H2180versus H2160in the {sv•p •tm) inthelimith -->0 andplanttranspired vapor ({5180 --> w
diffusionof water vapor acrossthe stomatalcavity and leaf boundarylayer,equalto 1 + (•'/1000);
Rv•p 180/160 ratioofwatervapor in theairoutside the
{5)")in the limit h --> 1. We did not extrapolatethis empirical
regression to the globallevelto correct{5180of vaporin canopy, however.By usingthe GISS model fieldswe insteadas-
leaf;
sumea lowerboundary for {5180 in watervaporandhencefor {5•80of CO2in leaves. roots. Plate2c shows{5•80of leaf CO2,whichdecreases toward Overdryareas,{5180 of leafCO2islargerthan The firstimportantparameter in (12') is {5•',the {5180of highlatitudes. R•' •80/160ratioof groundwater whichistakenup by
groundwaterdeliveredto the leaf. At steadystate an equivalent amountof water deliveredto the leaf and lost by transpiration mustbe pumpedfrom the soil by the root system.Following the hypothesisof the SiB2 model soil hydrology,we considerthat the rootspumpgroundwaterfrom an intermediate soillayerbeneaththe surface(FigureA1 in AppendixA). Assumingthat no isotopicfractionationoccursduringthe root uptakeof water [Bariacet al., 1994b],we calculate{5•"from the
6%0, with a maximum over the Sahara Desert of 15%o. This is
mostlydue to low relative humiditywhich increasesthe value of {SLin (12). Note, however,that the maximumvaluesob-
tainedin deserts isnotlikelyto influence theatmospheric {5180 becauseit is associated with a negligibleexchangeof CO2. The simulatedisotopiccompositionof leaf CO2 is aslow in tropical rainforestsas in Siberianforests(roughly-3%o), despitethe
fact that leaf wateris moredepletedin 180 in Siberia.As
a)
Relative Humidityat the Leaf Surface Percent
Global Mean-
79.2
NP ß
ß
ß
:.
-
;....?,., ½,
3O
EQ -30 -60
8P 180
120 W
15.0 10.0
b)
60 W
25.0 20.0
35.0 30.0
0
45.0 40.0
60 E
55.0 50.0
120 E
65.0 60.0
75.0 70.0
85.0 80.0
180
95.0 90.0
•180 inWaterVaporat GroundLevel %0 V-SMOW
Global Mean-17.0
NP
c)
•180ofCO2 inLeaves %0 PDB-CO2
Global Mean-
3.2
NP
..
.
30
e
t80
120 W
-50 -6.5
-3.5
1.0 -.5
.. •.,,
....... ..?.. ......,,*
•,•
",.•...•:.:,.,•-::-'.• \ '-- ':' '
60 W
-2.0
.•,,,
0
4.0 2.5
60 E
7.0 5.5
10.0 8.5
120 E
13.0 11.5
16.0 14.5
180
19.0 17.5
Plate 2. (a) Relativehumidityat theleafsurface(annualaverage)in thephotosynthesis modelSiB2coupled with the CSU climatemodel.Becauseof planttranspiration the relativehumidityat the leaf surfaceis higher
thanin thefreeatmosphere abovethecanopy. (b) Annualmean•80 of atmospheric watervaporat ground levelin the NASA GISSisotopicmodel[afterJouzelet al., 1987].The vaporphaseis isotopically depletedby ---10%•with respectto meteoricwaterdueto the isotopicfractionation resultingfrom in-cloudcondensation
processes. (c) Annualmeant5•80of CO2in leaves, thatis,in isotopic equilibrium equilibrated withevaporatingleafwater.Thiscorresponds to the t5•80whichinfluences theatmosphere, mediated bythephotosynthesis flux.
5866
CIAIS ET AL.' STUDY OF •180 IN ATMOSPHERICCO2,1
•180 in Ocean SurfaceWater
a)
%0 V-SMOW
Global Mean = 0.3
NP 6O
3O
EQ -3O
-6O
SP 180
120 W -4.6 •, .... %,,.,,.
-5.0
60 W
-3.8
0
-3.0
-2.2
60 E
-1.4
-.6
120 E .2
180
1.0
1.8
.•.•..,•:-'
-4.2
-3.4
-2.6
-1.8
-1.0
-.2
.6
1.4
•100inCO2 attheOceanSurface
b)
%o V-PDB-CO2
Global Mean = 1.8
NP .............................................................
•...................... :--................ ?:: ..............
30
EQF-••---------------j,, "•--[-•-• -60
_[
SP 180
120 W
"
•-t..,t,,=,...•
.,, -.
-8.6
.., .x\•
l
... ...
60 W
-6.8
-5.0
i ,.
0
-3.2
60 E
-1.4
.4
120 E
2.2
180
4.0
5.8
............. ..•"'"% ;, ,....•.•.•... '"Y'•"%:•q";"'g•" :i:..,:.i•.•';,'..::
'"" '""•' 'i• ..
..........ß: ....... ,,..,
-9.5
-7.7
-5.9
-4.1
-2.3
-.5
1.3
3.1
4.9
Plate 3. (a) Oceansurface water&]aOregressed aftersalinity. The decrease nearthe icesheetsandin the riversestuaries is dueto the inputto theoceans of freshwater depletedin •80. (b) Isotopiccomposition of dissolvedCO2 emitted to the atmospherethroughair-seaexchangeprocesses, assumingfull isotopicequilibrium of CO2 with seawater.
CIAIS ET AL.' STUDY OF 8•80 IN ATMOSPHERIC CO2,1
5867
initiallyproposedbyCraigandGordon[1965] outlinedfor soils,thisisdueto theeffectof temperature onO•eq piricalregression and further update by J. C. Duplessy(personalcommunication, 1994)'
and to a lesser extent on a w L--vap'
8• = ai + a2S
5. Exchange of •80 With the Ocean The net CO2 flux betweenthe ocean and atmosphereis givenby Fo = -Fao + Foa= KexApCO2
(•3)
(17)
where S is sea surfacesalinityin gramsper kilogram [Levitus, 1982]. The empiricalvalue of the 8o w versusS linear slope,a 2 =
0.5%o g- 1kg.Thevalueof the intercept a • = - 16.75%o is
w where Fao (Foa) is the one-wayflux of CO2 from (to) the determinedsoasto yield a meanvalueof 0%0 VSMOW for 8o averaged over the world oceans between 60øS and 60øN, exatmosphere,Kex is the air-seagas exchangecoefficient,and ApCO2 is the differencein partial pressureof CO2 between cluding polar oceanswhich deviate significantlyfrom the ocean and atmosphere.
The air-seagasexchangecoefficientKex is taken from the stabilitydependenttheoreticalformulationof Erickson[1993]. The field of ApCO2 is calculatedby the oceangeneralcirculation model HAMOCC (Max Planck Institute, Hamburg) which includesa parameterizationof biologicalprocesses in the ocean[Maier-Reimer, 1993;K. Kurz andE. Maier-Reimer, Geochemicalcyclesin an ocean general circulationmodel: Planktonsuccession and seasonalpCO2, submittedto Global Biogeochemical Cycles,1996].Note that the ApCO2 fieldsare for the preindustrialera,whichis not consistent with our sim-
VSMOW value. Plate 3a indicatesthat 8o w takes lower values wherelarge amountsof freshwaterare deliveredto the ocean,
because continental freshwater isdepleted in 180withrespect to seawaterbythe isotopicdistillationof moistair movingfrom the oceansto the continents.The isotopiccompositionof the oceansurfaceis thusdepletedby about 1%o at high latitudes around Antarctica
and Greenland
because of the massive dis-
chargeof icebergsand in the estuariesof the largestrivers.
Plate3b shows8180of CO2 in isotopicequilibrium with oceanwater, 80. The temperaturedependenceof the equilib-
riumfractionation factorO•eq hastheeffectof increasing 8o at ulationof today's180 cycle.However,the isotopic fluxpro- high latitudesby a few per mil, whichopposesthe latitudinal
variation of 80. The result is an overall increaseof 8o as a atmospheric 8180value(seebelow)sothatat thisstage, using function of latitude, with maximum values of 5-6%0 near the a preindustria! ApCO2 fieldintroduces onlya verysmallbias. seaice marginaroundGreenlandand Antarctica. For consistency with our atmospherictransportmodel, the ocean-atmosphere CO2 fluxesare maskedover regionscov- 6. Anthropogenic Emissions: Fossil Fuels ered by sea ice. and Biomass Burning Regardingthe isotopicfluxes,we havemadethe assumption Carbon dioxide derived from the combustionof hydrogenthat dissolvedCO2 is in isotopic equilibrium with seawater bound carbon bears an isotopiclabel of -17%o PDB-CO2, accordingto reaction(1). We accountfor no catalyticprocess whichcorresponds to the isotopicvalueof atmosphericoxygen that could yield isotopicequilibrationduring a short contact [Kroopnick and Craig,1972].Anthropogenicfluxesof the isobetween atmosphericCO2 and ocean water. Excludingthe topicspecies CO180arethusproportional to theCO2fluxes.
portionalto the net oceanflux hasonly a smalleffecton the
possibilityof rapid hydrationof CO2 with a time constant shorter than the crossingof the diffusivefilm at the air-sea interfaceis supportedby the fact that no evidencefor CA catalysishasbeenfound so far in the ocean.The net air-sea
For fossilCO2 emissions we usedthe estimatesof Marland et al. [1985],distributedaccordingto populationdensityby Fung et al. [1987].For biomassburningemissions, we have useda compilationof observationaldata which include forest and fluxof C18OOhasanexpression similar to thatforisotope 13C savannaburning(seasonal)aswell as agriculturalwastesand [Tanset al., 1993;Ciaiset al., 1995]and is givenby fuel wood burning(annuallyconstant)[Hao and Liu, 1994]. 18Fo= _ otwR aFao+ otwRoFoa (14) We needthe grossfluxof CO2 resultingfrom biomassburning,
Fbur,for calculating 8180in the atmosphere, not thenetde-
where aw is fractionationassociated with CO2 diffusionat the
forestationfluxwhichis significantly lowersinceit includesthe
air-seainterface [Vogel etal., 1970]andRois 180/160ratioof uptakeof CO2 dueto regrowthof burnedecosystems [HoughdissolvedCO2. Here (14) can be rewrittenin the form
•SFo= awRaFo + aw(Ro- Ra)Foa
ton et al., 1987].Conceptually, regrowthshouldbe treatedfor (15)
The left-handterm of (14) is an isotopic"equilibrium"flux
8180 as an additionalcomponent of the leaf exchange flux (linkedto GPP), butwe neglectedit, first,becauseit is a small flux comparedto the natural componentsFLa and FaL and,
second,becauseit hasa minor isotopicdisequilibriumwith the 8180of leaf CO2is in the range0-4%0 in the sinceit is proportionalto the isotopicratioR a.The right-hand atmosphere: flux F o discorrespondsto an isotopic"disequilibrium"flux tropicscomparedto -17%o when plantsare burned: which can be interpretedas a tendencytoward local isotopic
Foeq,which hardly influences the8180of atmospheric CO2
balance between atmospheric CO2and8180in dissolved CO2. By definition,the isotopicratioR o of CO2 in isotopicequilibrium with water is givenby
Ro= O•eq(ro) Rj
(16)
whereToisseasurface temperature andRo wis180/160ratioof
•8Ffos = RfFfos
(18)
18Fbu r -- gfFbu r
(19)
whereRf istheisotopic ratioof CO2produced bycombustion (5 = -17%o)
surface waters.
7. GlobalBudgetof •80 in Atmospheric CO2
We have calculatedRo w in a mannersimilarto that of Farquharet al. [1993].Ro w is a functionof salinityusingthe em-
Before couplingthe fluxesas calculatedaboveto a threedimensionalatmospherictransportmodel for calculatingthe
5868
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERICCO2,1
atmospheric 8•80, it isusefulto testthevalueswe havedeter- dependent gastransfer velocities. The 8•80 of CO2emittedby minedfor the modelparametersby closingthe globalbudget. the combustion of hydrogen-bound carbonis isotopically deTheglobalmeanvalueof 8•80 in theatmosphere, 8a,isinflu- pletedwith respectto the meanatmospheric valueandhasthe encedby soils,vegetation,andair-seaexchange (anthropogen- sameisotopiccharacteras atmospheric 02 (-17%o). We ac-
ic emissions are omitted).The annualmeantrendof 8•sOin
count for fossil fuel industrial emissions in the northern hemi-
atmospheric CO2 is closeto zero [Franceyand Tans,1987]and sphere and biomassburning in the tropics.There is some it is givenby the followingexpression to a goodapproximation inconsistency in assembling fields generatedby differentcli[Farquharet al., 1993]: matemodelsto infer the terrestrialisotopicexchange. Ideally, the isotopiccompositionofwater and CO2wouldbe calculated dSa simultaneously withina singleGCM, with fullyinteractiveisod-•-= 0 (20) topic hydrology,physiology,and photosynthesis.Unfortunately,sucha modeldoesnot existat present. d8 a The oxygenisotopefluxeshave been tested againstthe --[Foa(8 o - 8a) -4-•w(Fao- Foa) dt
Ca
globaltrendin atmospheric 8•80,whichisobserved to beclose
+ Fsoi•s(8s-8a-4-Es)-1 t-ZIAa]
to zero. This conditioncanbe met if the globalaveragefractionation at the soil-air interface is of -5%o, a value smaller
where /•a standsfor the discrimination of oxygenisotopeby conversion
in leaves:
Cc
/•a----8d -4Ca_C•(SL - 8a)
than the one inferredbyFarquharet al. [1993]in an independentcalculation(-7.6%o), althoughit lieswithin a physically acceptable range.Overall,the oceanicandanthropogenic contributionsare relativelyminor comparedto the isotopicex-
(21) changewiththeterrestrialbiota.We confirmthefactthatleaf
The two mostimportanttermsin (20) are thoserelativeto leaf and to soil isotopicexchange,whereasthe oceanflux is relativelyminor. Global numbersfor GPP, discrimination,and other parametersbear a large uncertaintyso that there is not a uniquesolutionto (20) yieldingd Sa/dt = 0. Alternatively, we solved(20) for es,the diffusivefractionationof CO2 respiredby soils.Usingthe valuesgivenin Table 1 for the global quantitieswhich appear in (20), we infer es = -5%0. This working value of % is significantlylower than the value of -8.8%o corresponding to strictlymoleculardiffusion.Possibly, turbulentdiffusionplaysan importantrole in transferringCO2 from the soilsurfaceto the atmosphere, whichwouldlower Another explanationfor the relativelylow value of es that we infer by solving(20) wouldbe that the productionof CO2 in soilsoccursmainlynear the surface,whichwould diminishthe influenceof the diffusivefractionationin the isotopiccomposition of CO2 emitted by soils as shownby Hesterberg and Siegentha[er [1991]in the caseof an exponentialdecreasein the CO2 productionat depth. Farquharet al. [1993]infer that •, = -7.6%0, a valuecloser to moleculardiffusion, from a budgetequationsimilarto (20). Thisis mostlybecausetheyemploya highervalueof the global discrimination than the onewe establishin thispaper.Nevertheless,althoughitsvalueisplausible,the globalsignificance of •s = -5%0 awaitsfurtherexplanations, and it shouldbe consideredas a tuningthat we applyto the globalbudgetso as to yield a zero long-termtrend in
exchange globally enriches in •sOtheatmospheric CO2reservoir,whereassoilexchange hasthe oppositerole. Apart from the globalbudget,we expectthat the geographical differences in the isotopicfluxeshavean influenceon the spatialdistribu-
tionof 8•sOin atmospheric CO2.In a companion paper[Ciais et al., this issue]describinga three-dimensional tracer simulation,we providea moredetailedassessment of the respective role of eachreservoir,with specialemphasis onvegetationand soils.
AppendixA: Groundwater •80 InferredFrom the Soil Hydrology in the SiB2 Model We detailin the following howthe 8•sOof waterin soilsis obtained fromthe 8•80 of meteoric waterusingtheparameters of soil hydrologyin SiB2. The soil columnis dividedinto
threelayers(FigureA1). The surfacelayerreceivesprecipitation and loseswater throughsurfacerunoff,evaporation,and infiltrationto the intermediatelayer.The intermediatelayer corresponds to the rootingzoneof plants,in whichplantstake up water to transpireto the atmosphereand hasno runoff.The
deep soil layer receiveswater by infiltrationfrom the layer aboveand losesit by deeprunoff.Only evaporationfractionates the heavyisotopeof water; all other fluxesconservethe
isotopicratios.The isotopiceffectsof the interceptionof precipitationby the canopyare neglected. In Figure A1,
w•,(Rj,8•') integrated watercontent(180/160,8180)of the surfacelayer (mm);
8.
Conclusions We have calculated the surface fluxes that control the 8•80
in atmosphericCO2. The mostcrucialassumption is that CO2 exchanges isotopically to fully reachisotopicequilibriumwith water availablein leaves(becauseof the presenceof carbonic anhydrase) andin surfacesoil.The validityof thisassumption shouldbe furtherinvestigated throughlaboratoryexperiments. The calculation of terrestrial carbon fluxes is based on the SiB2
photosynthesis model coupledwith the CSU GCM. The isotopic compositionof meteoric water comesfrom the NASA
GISSclimatemodel.The air-sea8•80 exchange is calculated usingApCO2from the HAMOCC oceanmodeland stability-
wi(R•',•") integrated watercontent(18Ofi60,8•80) of the intermediate(rooting)layer;
Wd(R•',8•') integrated watercontent(18Ofi60,8•SO)of the deep(recharge)layer (mm);
a•_va p fractionation of H2180fortheliquid-vapor phasetransition, equalto (1 + (•_vap/1000); P
precipitationreachingthe groundsurface
(mmd-•); E evaporation fromtheground surface (mmd-•); T uptakeof water by rootsequalstranspiration
byplants(mmd-•); Fsi infiltration of water from surfaceto
intermediate soillayer(mmd-•);
CIAIS ET AL.' STUDY OF 8•80 IN ATMOSPHERICCO2,1
where #s is the stomatalconductanceto water vapor,A is the net assimilationrate of CO2,h and Cs are the relativehumidity and mixingratio of CO2 at the leaf surface(for simplicity,Cs is taken equal to the atmosphericvalue Ca), p is the atmosphericsurfacepressure,and m and b are empiricallyderived
i
I I
w
surface
• Ws, ,5s layer Fsi• wi'•t) layer wdeep Fid• Wd, •d layer \
. •
5869
Ms
parameters.
The net assimilationrate A is modeled as limited by the kinetics of the carboxylationenzyme Rubisco, by electron transport (a seriesof reactionsthat take place when green plant cells are illuminated with visible radiation), and by buildup of the sugarsand starchesthat are the end productsof photosynthesis. Farquhare! al. [1980]useda simpleminimum
intermediate
Vd
of the three limits to calculate
the net carbon assimilation
rate
Figure A1. Fluxesof H:O in soilsas calculatedby the SiB2 A = min (Loc,LOe, LOs) -- 'q)•d (B2) model includingthree soil compartments.The isotopiccompositionof groundwateris calculatedoff-linefrom thesefluxes: (often solid lines are fluxesthat take placewith no isotopicfraction- where % is the carbon-limitedrate of photosynthesis ation, and the dashedline includesthe specificfractionationof referred to as Rubisco-limited,sincethe rate is determinedby Rubiscoenzymekinetics), LOe is the rate limited by electron water isotopesduring evaporation. transport(light-limited),Los is the endproduct-limited(or sink-
limited)rate,and•)•dis the rate of carbonlossfromthe canopy due to "dark" respiration.SiB2 usesa similar approachbut replacesthe simpleminimumin (B2) with a smoothedfunction Fid infiltration of water from intermediateto deep to avoid abrupt transitionsfrom one limitation to another soillayer(mmd-l); [Collatz et al., 1991]. The Rubisco-limitedand light-limited Vs surface runoff(mmd-l); assimilationrates are calculatedfrom enzymekineticsmodels Vd deeprunoff(mmd-l). developedby Collatzet al. [1991]. For C3 vegetationthe sinkThe massbalanceof H:O andH:180 in eachlayeriswritten limited rate is parameterized as a simple fraction of the Rubisco activity, and for C4 vegetation, Losrefers to PEPdws= P - E - Fsi- Vs (A1) carboxylaselimitation accordingto the model of Collatzet al. [1992]. Leaf respiration•)•d is parameterizedaccordingto dwi = Fsi- T- Fid (A2) Rubiscoactivityand canopytemperature. dWd= Fid- [?d (A3) All three photosyntheticrates and the leaf respirationare scaledby nondimensional parameters(fHOT, fCOLD,f-I,, fRH, d(R•ws)= R½'P- o•_vapR•E -- R•'Fsi- R•'I/s (A4) and fFRZ) representingenvironmentalstressesdue to excessivelyhigh or low canopytemperature,drought,low relative d(R•wi) = R•Fsi- R•VT- R•VFid (A5) humidity, and frozen soils. The leaf-level assimilationrate, d(RjWd) = g•VFid- gjVd (m6) stomatalconductance,and other physiologicalparametersare scaledto the canopyusing an assumedoptimal relationship Using d(Rw) = wdR + Rdw and the definitionR = (1 + betweenleaf nitrogen(and henceRubisco)and the time-mean 8/1000)RvsMow, we substitute(A1)-(A3) into (A4)-(A6) to profile of photosynthetically active radiation in the canopy. obtain the set of differentialequations(A7)-(A9) describing Details of the parameterizationare presentedby Sellerset al. thevariationin 8180of waterin eachsoillayer. [1996a,AppendixC]. The partial pressureof CO2 in the leaf interior (Ci) is d87: w•-l[p(8•- 8•) + E(8L_vap) ] (AT)
dS•'= wi-•F•i(8• ' - 87)
(A8)
da•'= wjlFid(8•"- 8•')
(A9)
•
A-O•d
Cc CirCa
We integrate these equations numerically to calculate the monthlymeanvaluesof 8•TM and 8•vwhichare, respectively, used to expressthe isotopiccompositionof CO: in soils and in leaves(sections3 and 4).
chloroplast • •r..j surface
layer
Appendix B: Biospheric CO2 Fluxes in the SiB2 Model
B.1.
intermediate
Photosynthesis
FollowingCollatzet al. [1991,1992]andBall [1988],stomatal conductance,the carbonassimilationrate, and the CO: concentrationat the leaf surfacesare assumedto be related by Ah
g•=m-•-•-• +b
(B1)
(root) layer
Figure B1. Resistanceto the diffusionof CO: from the canopy to the leaf chloroplast.During the assimilationthere is a gradient in the CO: partial pressurefrom Ca to Co.
5870
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERICCO2,1
SiB2modelare similarto thosefoundin previousstudies,and the seasonaland diurnalcyclesof net carbonflux to the atmoCO2 efilux due to soil respiration.The partial pressurein the spherecomparefavorablywith the limited field observations chloroplastCc, is slightlysmallerthan Ci due to the resistive available[Denninge! al., 1996].When thesefluxeswereusedin path of CO2 acrossthe mesophyllcell [see,e.g.,Farquharetal., the CSU GCM to calculate the full three-dimensional concentration field of atmosphericCO2, the resultscomparedvery 1993]but we make the approximationthat favorably with the data of the NOAA/CMDL flask station Cc•Ci (g3) network [Denning,1994;Denningand Randall, 1996]. diagnosedusing a resistancenetwork (Figure B1) from the stomatal conductance,the net assimilationrate, and the rate of
B.2.
Respiration
The relativeintensityof soil respiration,denotedby •*, is References diagnosedfrom soil moisture and soil temperature at each Allison, C. E., R. J. Francey,and H. A. Meijer, Recommendationsfor the reportingof stableisotopemeasurementsof carbonand oxygen model time step followingthe method used by Raich et al. in CO2 gas,in References and Intercomparison Materialsfor Stable [1991]in the terrestrialecosystem model.The soilrespiration Isotopesof LightElements,Proceedings of a Consultants Meetingheld diagnostic•* is definedas in Vienna, 1-3 December1993 IAEA-TECDOC-825, pp. 155-162,
•* = 2.0Of(w)
(B4)
where B
f(w) = 0.2 + Wsa t
(B5)
(W ....
Int. At. Energy Agency, Vienna, 1995. Arakawa, A., and V. R. Lamb, Computationaldesignof the basic dynamicalprocesses of the UCLA generalcirculationmodel,Methods Cornput.Phys.,17, 173-265, 1977. Arakawa, A., and V. R. Lamb, A potential enstrophyand energy conservingschemefor the shallowwater equations,Mon. Weather Rev., 109, 18-36, 1981.
Baertchi,I. H., andW. C. Macklin,Absolute180 contentof standard
B= Wopt zm _ 100zm]
mean ocean water, Earth Planet. Sci. Lett., 31, 341-344, 1965.
Ball, J. T., An analysisof stomatalconductance,Ph.D. thesis,89 pp.,
In the temperatureresponsefunction,Qt = (T - 298)/10. The temperatureusedto defineQt in (B3) is thewarmerof the surfacesoil temperatureand the deep soil temperature.The variablew in (B4) is the fractionof the pore spaceoccupiedby water in the root zone (middle layer) of the soil.The param-
etersWsat, Wopt, andZmareprescribed according to soiltexture
Stanford Univ., Stanford, Calif., 1988. Bariac, T., J. Gonzalez-Dunia, F. Tardieu, D. Tessier, and A. Mariotti,
Variabilit6spatialedela composition isotopique del'eau(180,2H) au sein des organesdes plantesa6riennes,1, Approche en conditions contr616es,Chem. Geol., 115, 307-315, 1994a.
Bariac, T., J. Gonzalez-Dunia,N. Katerji, O. B6thenod,J. M. Bertolini, and A. Mariotti, Variabilit6 spatiale de la compositionisoto-
piquedel'eau(180,2H) dansle continuum sol-plante-atmosphere,
usingvaluessuggested by Raichet al. [1991].Soil respirationis
2, Approche en conditionsnaturelies,Chem. Geol., 115, 317-333,
a maximum for somevalueWopt of soilmoisture, andrespiration is lessefficientunder very dry or very wet conditions. A dimensionless monthly mean soil respirationrate is defined as
1994b.
Brenninkmeier,C. A.M., P. Kraft, and W. G. Mook, Oxygenisotope fractionationBetween CO2 and H20 , IsotopeGeosci.,1, 181-190, 1983.
Ciais, P., P. P. Tans, J. W. White, M. Trolier, R. Francey,J. A. Berry, D. Randall, P. J. Sellers,J. G. Collatz, and D. S. Schimel, Partition-
?(t) = • •*(t)At
(B6)
1 year
ingof oceanandlanduptakeof CO2asinferredby 813Cmeasurementsfrom the NOAA Climate Monitoring and DiagnosticLaboratoryglobalair samplingnetwork,J. Geophys. Res.,100, 5051-5070, 1995.
synthesis studyof 8•80 in atmowhere the overbar indicatesthe monthly mean. The flux of Ciais,P., et al.,A three-dimensional sphericCO2, 2, Simulationswith the TM2 transportmodel,J. GeoCO2 from the soil due to respirationis computedfrom this phys.Res., this issue. dimensionless rate by assuminga local steadystatefor carbon Collatz, G. J., J. T. Ball, C. Grivet, and J. A. Berry, Physiologicaland environmentalregulationof stomatalconductance,photosynthesis, storagein terrestrialecosystems on an annual basis,that is, and transpiration:A modelthat includesa laminarboundarylayer, that the annualsumof respirationlossis equal to the annual Agric.For. Meteorol.,54, 107-136, 1991. net carbonassimilation(ANA), definedas
Collatz, G. J., M. Ribas-Carbo,and J. A. Berry, Coupled photosynthesis-stomatal conductancemodel for leavesof C4 plants,Aust.J. ANA= • ,•(t)At (B7) Plant Physiol.,19, 519-538, 1992. I year Conway,T. J., P. P. Tans,L. S. Waterman,K. W. Thoning,D. R. Kitzis, K. A. Masarie,and N. Zhang,Evidencefor interannualvariabilityof whereA (t) is the monthlymean net assimilation.The flux of the carboncyclefrom the National Oceanicand AtmosphericAd(202 from the ecosystem due to respirationis calculatedfrom ministration/Climate Monitoringand DiagnosticLaboratoryGlobal air samplingnetwork,J. Geophys.Res.,99, 22,831-22,855,1994. ?(t) accordingto Craig,H., and A. Gordon,Deuteriumand oxygen-18variationsin the ocean and the marine atmosphere,in StableIsotopesin Oceanic 2•* (t) = ?(t)ANA (B8) Studiesand Paleotemperatures, Lab. Geol. and Nuclear Sci., Pisa, Italy, 1965. sothat the net annualfluxof CO2 from everygridpoint is zero. The model was integratedfor 10 yearson a low-resolution Denning, A. S., Investigationsof the transport,sources,and sinksof atmosphericCO2 usinga generalcirculationmodel,Atmos.Sci.Pap.
grid (7.2ølatitude x 9ølongitude)to allow soilmoisturefields 564, Colo. State Univ., Fort Collins, 1994. to equilibratewith the simulatedclimate.After this "spin-up" Denning,A. S., and D. A. Randall, Simulationsof terrestrialcarbon metabolismand atmosphericCO2 in a generalcirculationmodel,2, periodthe modelwasintegratedfor an additional4 yearson a SimulatedCO2 concentrations, Tellus,in press,1996. 4ø x 5øgridwith 17verticallayersand a time stepof 6 min. The
fieldsusedto drivethe•80 calculation weremonthly means for the final year of the simulation.The spatialand seasonaldistributionof net primaryproductivity(NPP) simulatedby the
Denning,A. S., J. G. Collatz, C. Zhang, D. A. Randall, J. A. Berry,P. J. Sellers,and S.C. Wofsy,Simulationsof terrestrialcarbonmetabolism and atmosphericCO2 in a generalcirculationmodel, 1, Surface carbonfluxes,Tellus,in press,1996.
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERIC CO2,1 D6rr, H., and K. O. Munnich, Annual variation in soil respirationin selectedareasof the temperatezone, Tellus,Set. B, 39, 114-121,
5871
seasonalcycle for present-day conditions,J. Geophys.Res., 92, 14,739-14,760, 1987.
Keeling, C. D., T. P. Whorf, M. Wahlen, and J. Van der Plicht, Interannual extremesin the rate of rise of atmosphericcarbon Enting, I. G., C. M. Trudinger,R. J. Francey,and H. Granek, Synthesis dioxide since 1980, Nature, 375, 666-670, 1995. inversion of atmosphericCO2 using the GISS tracer transport model,Div. Atmos.Res. Tech.Pap.Aust. C.S.I.R.O.,29, 1-44, 1993. Keeling, R., The atmosphericoxygencycle:The oxygenisotopesof atmosphericCO2 and 02 and the O2/N2 ratio, U.S. Natl. Rep. Int. Enting, I. G., C. M. Trudinger,and R. J. Francey,A synthesis inversion Union Geod. Geophys.1991-1994, Rev. Geophys.,33, 1253-1262, of theconcentration and8•3Cof atmospheric CO2,Tellus, Set.B, 47, 1987.
35-52, 1995.
1995.
Erickson,D. J., III., A stabilitydependenttheoryfor air-seaexchange, Kroopnick,P., and H. Craig, Atmosphericoxygen:Isotopiccomposition and solubilityfractionation,Science,175, 54-55, 1972. J. Geophys.Res.,98, 8471-8488, 1993. Atlas of the World Ocean,Prof Pap., 173, Farquhar,G. D., S. von Caemmerer,and J. A. Berry, A biochemical Levitus, S., Climatological U.S. Gov. Print. Off., 1982. model of photosynthetic CO2 assimilationin C3 plants,Planta, 149, Maier-Reimer, E., Geochemicalcyclesin an oceangeneralcirculation 78-90, 1980. model:Preindustrialtracer distribution,GlobalBiogeochem. Cycles, Farquhar, G. D., K. T. Hubick, A. G. Condon, and R. A. Richards, 7, 645-677, 1993. Carbonisotopefractionationand plant water-useefficiency,in StableIsotopesin EcologicalResearch,editedby P. W. Rundel et al., pp. Marland, G., R. M. Rotty, and N. L. Treat, CO2 from fossil fuel burning:Global distributionof emissions, Tellus,Set.B, 37, 243-258, 21-40, Springer-Verlag,New York, 1989. 1985. Farquhar, G. D., J. Lloyd, J. A. Taylor, L. B. Flanagan,J.P. Syvertsen, of H2•80 in gases, J. Chem.Phys., K. T. Hubick, S.C. Wong, and R. Ehleringer,Vegetation effectson Merlivat,L., Moleculardiffusivities 69, 1864-2871, 1978. the isotopecompositionof oxygenin atmosphericCO2,Nature,363, Raich, J. W., E. B. Rastetter,J. M. Melillo, D. W. Kicklighter,P. A. 439-443, 1993. Steudler,and B. J. Peterson,Potentialnet primaryproductivityof F6rstel, H., A. Putral, G. Schleser,and H. Leith, The world pattern of SouthAmerica:Applicationof a globalmodel,Ecol.Appl., 1,399oxygen-18in rainwater and its importancein understandingthe
429, 1991. biogeochemicaloxygencycle,in IsotopeRatiosas PollutantSource Randall, D. A., and D.-M. Pan, Implementation of the Arakawaand BehaviorIndicators,pp. 323-344, Int. At. Energy Agency,ViSchubertparameterizationwith a prognosticclosure,in TheRepreenna, 1975. sentationof CumulusConvectionin NumericalModels, edited by Fowler, L. A., and D. A. Randall, Liquid and ice cloudmicrophysics in K. Emanuel and D. Raymond,pp. 137-144, Am. Meteorol. Soc., the CSU generalcirculationmodel, II, Impact on cloudiness,the
Earth's radiationbudget,and the general circulationof the atmosphere,J. Clim., 9, 530-560, 1995a. Fowler,L. A., and D. A. Randall,Liquid andice cloudmicrophysics in the CSU generalcirculationmodel, III, Sensitivityto modelingassumptions,J. Clim., 9, 561-586, 1995b. Fowler,L. A., D. A. Randall, and S. A. Rutledge,Liquid and ice cloud microphysics in the CSU general circulationmodel, I, Model descriptionand simulatedmicrophysicalprocesses, J. Clim., 9, 486529, 1995.
Francey, R. J., and P. P. Tans, Latitudinal variation in oxygen-18of atmosphericCO2, Nature, 495-497, 1987. Francey,R. J., C. E. Allison, I. G. Enting, J. W. C. White, M. Trolier, and P. P. Tans,Changesin the oceanicand terrestrialcarbonuptake since 1982, Nature, 373, 326-330, 1995. Friedman,I., andJ. R. O'Neill, Compilationof stableisotopefractionation factorsof geochemicalinterest, U.S. Geol. Sum Prof. Pap.,
Boston, Mass., 1993. Randall, D. A., Harshvardhan, D. A. Dazlich, and T. G. Corsetti,
Interactionsamongradiation,convection,and large-scaledynamics in a generalcirculationmodel,J. Atmos.Sci., 46, 1943-1970, 1989. Randall, D. A., Harshvardhan,and D. A. Dazlich, Diurnal variability of the hydrologiccyclein a generalcirculationmodel,J.Atmos.Sci., 48, 40-62, 1991.
Randall, D. A., Q. Shao, and C.-H. Moeng, A second-orderbulk boundary-layermodel.J. Atmos. Sci., 49, 1903-1923, 1992. Randall,D. R., et al., A revisedland-surface parameterization(SiB2) for GCMs, 3, The greeningof the ColoradoStateUniversitygeneral circulation model, J. Clim., 9, 738-763, 1996. Sellers,P. J., Y. Mintz, Y. C. Sud,andA. Dalcher,A simplebiosphere model(SiB) for usewithingeneralcirculationmodels,J.Atmos.Sci., 43, 505-531, 1986.
Sellers,P. J., J. A. Berry, G. J. Collatz, C. B. Field, and F. G. Hall, Canopyreflectance,photosynthesis, and transpiration,III, A reanal440-KK, 1977. ysisusingenzymekinetics--Electrontransfermodelsof leaf physiFung, I. Y., C. J. Tucker, and K. C. Prentice,Applicationof advanced ology,RemoteSens.Environ.,42, 1-20, 1992a. very high resolutionradiometer vegetationindex to study atmo- Sellers, P. J., M.D. Heiser, and F. G. Hall, Relations between surface sphere-biosphere exchangeof CO2,J. Geophys. Res.,92, 2999-3013, conductance andspectral vegetation indices at intermediate (100m2 1987.
Hao, W.-M., and M.-H. Liu, Spatial and temporal distribution of tropical biomassburning,Global Biogeochem.Cycles,8, 495-503, 1994.
Harshvardhan,R., D. A. Davies, D. Randall, and T. G. Corsetti, A fast
radiationparameterizationfor generalcirculationmodels,J. Geophys.Res.,92, 1009-1016, 1987. Hesterberg,R., and U. Siegenthaler,Productionand stable isotopic compositionof CO 2 in a soil near Bern, Switzerland,Tellus,Set.B, 43, 197-205, 1991.
to 15km2) lengthscales, J. Geophys. Res.,97, 19,033-19,059, 1992b. Sellers,P. J., D. R. Randall, J. A. Collatz,J. A. Berry, C. B. Field, D. A. Dazlich, C. Zhang, and L. Bounoua,A revisedland-surfaceparameterization(SiB2) for atmosphericGCMs, 1, Model formulation, J. Clim., 9, 676-705, 1996a. Sellers,P. J., S. O. Los, C. J. Tucker, C. O. Justice,D. A. Dazlich, G. J. Collatz,and D. A. Randall,A revisedland surfaceparameterization
(SiB2) for atmospheric GCMs, 2, The generationof globalfieldsof terrestrial biophysicalparametersfrom satellite data, J. Clim., 9, 706-737, 1996b.
Houghton,R. A., et al., The flux of carbonfrom terrestrialecosystems Silverman,D. N., CarbonicAnhydrase:Oxygen-18exchangecatalyzed to the atmospherein 1980 due to changesin land use:Geographic by an enzymewith rate-contributingproton transfersteps,Methods distributionof the globalflux, Tellus,Set.B, 39, 122-139, 1987. Enzymol., 87, 733-753, 1982. Hut, G., Report to the Director General on the consultants'group Suarez, M. J., A. Arakawa, and D. A. Randall, Parameterization of the meetingon stableisotopereferencesamplesfor geochemicaland planetary boundarylayer in the UCLA general circulationmodel: hydrologicalinvestigations,report, 41 pp., Int. Atomic Energ. AsFormulation and results,Mon. WeatherRev., 111, 2224-2243, 1983. soc., Vienna, 1987. Tans, P. P., R. J. Francey,and G. I. Pearman,Large scalevariationsin InternationalAtomicEnergyAgency(IAEA), Statisticaltreatmentof 8•80 of tropospheric CO2,in Baseline 1983-1984,editedby R. J. environmentalisotopedata in precipitation,Tech.Rep.Set.I.A.E.A., Franceyand B. W. Morgan, Dep. of Sci.,Bur. of Meteorol./CSIRO 206, 1-256, 1981.
Jacob,H., and C. Sonntag,An 8-yearrecordof the seasonalvariation
of 2H and •80 in atmospheric watervaporand precipitation at Heidelberg,Germany,Tellus,Set.B, 43, 291-300, 1991. Jouzel, J., G. L. Russell, R. J. Suozzo, R. D. Koster, J. W. C. White,
and W. S. Broecker,Simulations of the HDO and H2•80 atmosphericcyclesusingthe NASA/GISS generalcirculationmodel:The
Div. of Atmos. Res., Melbourne, Victoria, Australia, 1986.
Tans,P. P., J. A. Berry,andR. F. Keeling,Oceanic•3Cdata:A new windowon CO2 uptakeby the oceans,GlobalBiogeochem. Cycles,7, 353-368, 1993.
Vogel, J. C., P.M. Grootes,and W. G. Mook, Isotopicfractionation betweengaseousand dissolvedcarbon dioxide,Z. Phys.,230, 225238, 1970.
5872
CIAIS ET AL.: STUDY OF 8180IN ATMOSPHERICCO2, 1
White, J. W. C., The climaticsignificanceof D/H ratios in white pine in the northeasternunited states,331 pp., Ph.D. thesis,Columbia
nent that controls(5180of atmospheric CO2 and 02, PlantCell
Sciences,Colorado State University, Fort Collins, CO 80523-1370. (e-mail: [email protected]) R. J. Francey, Division of AtmosphericResearch,CSIRO, PMB Aspendale,Melbourne, Victoria 3195, Australia. M. Heimann, Max-Planck Institut ffir Meteorologie,Bundestrasse 55, D-20146 Hamburg, Germany. H. A. J. Meijer, CIO, Universityof Groningen,9722 JX Groningen, Netherlands.(e-mail: [email protected]) P. Monfray, Centre des Faibles Radioactivit•s, Bfitiment 709/
Environ., 7, 73-80, 1994.
LMCE, 91191 Gif sur Yvette, France.
Univ., New York, 1983.
White, J. W. C., and S. D. Gedzelman, The isotopiccompositionof atmosphericwater vapor and the concurrentmeteorologicalconditions,J. Geophys.Res.,89, 4937-4939, 1984. Yakir, D., J. A. Berry, L. Giles, and C. B. Osmond,Isotopic heterogeneityof water in transpiringleaves:Identificationof the compo-
J. A. Berry, Department of Plant Biology, Carnegie Institution of Washington, 290 Panama Street, Stanford, CA 94305. (e-mail: [email protected]) P. Ciais, LMCE,
CEA l'Orme des Merisiers, Commissariat h
P. P. Tans, Climate Monitoring and DiagnosticLaboratory,NOAA, ERL 3, 325 Broadway,Boulder, CO 80303. M. Trolier and J. W. C. White, Institute of Arctic and Alpine Research,University of Colorado, Campus Box 450, Boulder, CO 80303.
l'Energie Atomique, Bfitiment 709, Saclay 91191 Gif sur Yvette, France.(e-mail:[email protected]) G. J. Collatz and P. J. Sellers,NASA Goddard SpaceFlight Center, MC 923, BiosphereSciencesBranch,Greenbelt,MD 20771. (ReceivedNovember 18, 1995;revisedJune 25, 1996; A. S. Denning and D. A. Randall, Department of Atmospheric acceptedJuly 12, 1996.)
