The Fate of Marine Autotrophic Production
Author(s): Carlos M. Duarte and Just Cebrian
Reviewed work(s):
Source: Limnology and Oceanography, Vol. 41, No. 8 (Dec., 1996), pp. 1758-1766
Published by: American Society of Limnology and Oceanography
Stable URL: http://www.jstor.org/stable/2838660 .
Accessed: 31/08/2012 11:48
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].
.
American Society of Limnology and Oceanography is collaborating with JSTOR to digitize, preserve and
extend access to Limnology and Oceanography.
http://www.jstor.org
Limnol. Oceanogr, 41(8), 1996, 1758-1766
? 1996, by the American Societyof Limnologyand Oceanography,Inc.
The fateofmarineautotrophic
production
CarlosM. Duarteand JustCebrian
Centro de Estudios Avanzados de Blanes, CSIC, Cami de Santa Barbara s/n,17300 Blanes, Girona, Spain
Abstract
The fateofphotosynthetic
carbonin marineecosystemsdominatedby different
typesofprimaryproducers
was examined by compilingpublishedreportson herbivory,autotrophicrespiration,decomposition,carbon
storage,and exportratesas fractionsof net primaryproduction(NPP) in ecosystemsdominatedby different
typesofautotrophs(i.e. oceanic and coastal phytoplankton,
microphytobenthos,
coral reefalgae, macroalgae,
seagrasses,marsh plants, and mangroves). A large fraction(>40%) of the NPP of marine ecosystemsis
decomposed withinthe system,except formicrophytobenthos
(decomposition, -25% of NPP). Herbivory
tends to be highestformicroalgae(planktonicand benthic,>40% of NPP) and macroalgae (33.6 ?4.9% of
NPP) and is somewhatless forhigherplants.Microphytobenthos
exporton averagea much higherproportion
of theirNPP than do other microalgal communities,whereas marine macrophytes,except marsh plants,
exporta substantialproportion(24.3-43.5% on average)oftheirNPP. The fractionofNPP storedin sediments
is 4-foldgreaterforhigherplants (- 10-17% of NPP) than foralgae (0.4-6% of NPP). On average, -90% of
the phytoplanktonNPP is used to supportlocal heterotrophicmetabolism(i.e. grazed or decomposed). This
fractionis even higherin oceanic communities.Mangrove forests,and to a lesserextentseagrass meadows
and macroalgal beds, produce organic carbon well in excess of the ecosystemrequirements,with excess
photosyntheticcarbon (i.e. export rate plus storage)in these ecosystemsrepresenting-40% of NPP. Extrapolationof these resultsto the global ocean identifiesmarine angiosperms,which only contribute4% of
totalocean NPP, as major contributors
oftheNPP stored(30% oftotalocean carbonstorage)and subsequently
buriedin marinesediments.ConsiderationofburialofNPP frommarineangiospermsshould lead to estimates
of total burial of marine NPP that exceed currentestimatesby 15-50%.
Photosynthesisby marine organismsproduces 30-60
Pg (1 Pg = 1015g) of organiccarbon annually(e.g. Martin
et al. 1987; Charpy-Robaud and Sournia 1990; Smith
and Hollibaugh 1993), which represents -40% of the
total primaryproductionof the earth(Schlesinger1991;
Melillo et al. 1993). This photosyntheticcarbon is produced by autotrophicorganismsrangingfromcyanobacteria to trees (i.e. mangroves).Autotrophiccarbon production of the open ocean, which representsthe bulk
(90%; Smith and Hollibaugh 1993) of oceanic primary
production,is contributedbyunicellularautotrophs,with
a small contributionbeingmade by floatingmacrophytes.
In contrast,photosynthesis
by marinemacrophytes(macroalgae, seagrasses,marsh plants,and mangroves)is responsible fora significantfraction(at least 35-50%; Ryther 1963; Platt and Subba Rao 1975; Nixon et al. 1986)
ofthecarbon productionof thecoastal ocean, wherethey
represent 75% oftheautotrophicbiomass (Smith 1981).
In addition to these primaryproducers,photosynthetic
organisms associated with corals (e.g. coral reef algae)
supporta sizable production(0.7 Pg C yr-'; Crossland
et al. 1991), which is locally important.
Estimatesof primaryproductionalone do not provide
appropriateunderstandingof the ecological role of the
plants or their functionin the marine carbon cycle. A
more elaborateevaluation of the role of marineplantsin
primaryproductionrequires examination of the fate of
the carbon they produce. Communities dominated by
photosynthetic
organismsthat eitherexportor burysignificantamounts of theirprimaryproductionare likely
to be autotrophic,since thatfractionof primaryproducuse. In contrast,
tionis unavailable forlocal heterotrophic
organismsthat
communitiesdominatedbyphotosynthetic
channel most of theirproductionthroughherbivoresor
as detritusforlocal microbial use require only small inputs of allochthonouscarbon to be heterotrophic.
For example, coastal marine ecosystemsare known to
have some of the highestareal rates of primarycarbon
productionformarinesystems(Boyntonet al. 1982; Nixon et al. 1986), yetanalysesof net ecosystemmetabolism
indicatethat
based on carbon and nutrientstoichiometry
many such systemsare sources ratherthan sinks of inorganiccarbon (Smith and Mackenzie 1987; Smith and
fractionofphytoplankton
Hollibaugh 1993). A significant
productionis channeled to herbivores(55% on average;
Cebri'anand Duarte 1994) and is thereforemore directly
accessible to highertrophiclevels than is the carbon production of, for example, mangroves, which is mostly
Acknowledgments
channeled to decomposers (Lugo et al. 1988; Lee 1990).
Thisworkwas supported
bya grantfromtheSpanishInterMost organiccarbon producedby oceanic phytoplankton
ministerial
Commissionof Scienceand Technology(CICYT
grantAMB-94-0746).J.C.was supported
from is used withinthe upperocean, and onlya minorfraction
bya fellowship
accumulates in deep wateror is buried in the sediments
theCIRIT.
(< 10%;Lein 1984; Henrichsand Reeburgh1987), thereby
We thankR. Ariasforproviding
accessto literature,
and J.
Kenworthy,
J. Middelburg,
S. V. Smith,and an anonymous enteringa long-termsinkin thecarbon cycle.In contrast,
referee
forcomments
on thismanuscript.
long-livedseagrasses may buryrelativelylarge amounts
1758
Fate of marine autotrophicproduction
of the carbon theyproduce, such as the Mediterranean
seagrassPosidonia oceanica (L.) Delile, whichburies20200 g C m-2yr-' in sediments(Romero et al. 1994).
The fractionof marine photosynthetic
carbon flowing
throughdifferent
paths,such as theherbivoreand detrital
paths, has been reportedto be independentof primary
production(Cebri'anand Duarte 1994, 1995). This findingillustratestheneed forknowledgeof both thefateand
amount of photosynthetic
carbon productionin orderto
fullyunderstandthe functioning
of different
typesof marine ecosystems.Althoughour knowledgeof the magnitude of marine primarycarbon productionby different
photosynthetic
organismsis well established,our knowledge of the fateof this productionis less clear.
Here, we examine the fateof photosynthetic
carbon in
different
typesof marineecosystemsand discuss the implicationsof thesedifferences.
We do so based on a compilation of published reportson net primaryproduction
(NPP, g C m-2yr-') as well as thefateofNPP as described
by the carbon miassbalance equation
1759
values in thedata set,dependingon thetypeofecosystem
examined.Additionally,some studiesofmicroalgae,coral reefalgae, and seagrasses reportedcommunityrespiration rates. These estimates included the production
consumed by the autotrophs themselves,that decomposed by heterotrophs,and that ingestedby herbivores
contained within the respiration chamber (microzooplanktonin plankton,microfaunaand meiofaunain benthicchambers).D was obtained fromrespirationratesby
usingthe average values forthese autotrophicorganisms
(see above) and subtractingautotrophicrespiration.Additionally,herbivoreconsumptionwithinthe chambers
was also subtractedfromestimates of communityrespiration of microalgae-dominatedcommunities by assumingthis consumptionto be on average 500%of the
total herbivoryon communities of these primaryproducers(phytoplankton-e.g.Laws et al. 1988; Frost 1991;
Stromand Welschmeyer1991; microphytobenthos-Admiraal et al. 1983; McLachlan and Bate 1984; Carpenter
1986).
EstimatesofD derivedfromplanktoncommunityresNPP = D + H + E + S,
(1)
pirationratesand correctedformicroherbivory
and autowhereD is the carbon decomposed, H is the carbon controphicrespirationunderestimatethe total decomposisumed by herbivores,E is thecarbon exportedout of the
tion of phytoplanktonNPP because such estimates do
notconsiderdecompositionofthesinkingphotosynthetic
system,and S is the potentialcarbon storagein the ecocarsystems.
carbon.Decomposition ofthesinkingphotosynthetic
bon was assumed to be 17% ofNPP (Martinet al. 1987)
to estimatethe total proportionof phytoplanktonNPP
Methods
decomposed. The remainingestimatesof theannual production decomposed were derived fromthe annual proWe searched the literaturefor reportsof the compoductionofdetritus(Pd), excludingthatexportedfromthe
nents of Eq. 1. We expressedeach component as a persystem(i.e. Pd = NPP - H - E), by using average decentage of NPP to representthe relative importanceof
compositionrates(k, d- 1) forthedifferent
typeof marine
each processas a componentoftheflowofphotosynthetic phototrophs(microalgae, 0.0526?0.0049; macroalgae,
carbon. The data set comprised 154 reportson NPP and
0.0144?0.0042; seagrasses,0.0107?0.002; marshplants,
subsequent photosyntheticcarbon flow in marine eco0.0381?0.006; mangroves,0.0082?0.0035; Enriquez et
systemsthatare dominatedby different
typesof primary al. 1993) and the equation
producers,includingphytoplankton(oceanic and coastD = Pd[1 - exp(-k x 365)].
coral reefalgae, macroalgae,seaal), microphytobenthos,
grass meadows, marsh plants,and mangroves.The data
The detritusthatis not decomposed withina year acset is available fromthe authorsupon request.
cumulates on the sediment surface and representsthe
NPP was not always reportedas such. We calculated
stored carbon (S). Hence, our estimates of S represent
NPP fromgrossprimaryproduction(GPP) data for 19%
theexcesscarbonthataccumulatesannuallyin thesystem
of marsh plants and 60% of phytoplanktonaccordingto
and do not representcarbon burial, which is referredto
the equation
long-termcarbon storage.Carbon storedin thesediments
at the annual time scale is subject to remineralization
NPP = GPP(1 - Ra),
throughdiagenetic processes, which involve losses of
-70% and 97% of the carbon enteringcoastal and deep
whereRa representsthe average fractionof GPP respired
ocean sediments(Berner 1982; Lein 1984; Henrichsand
by the plants themselves.Estimates of the average proReeburgh 1987). Similarly,the carbon accumulated as
portionof GPP respiredby the autotrophswere also dedissolved organiccarbon (DOC) in deep oceanic waters,
rived from a survey of the literature,which yielded
which is >200-fold that contained withinlivingmarine
35.4?2.3% for phytoplankton,26.4?2.9% for micro14.1? 3.4% forcoral reefalgae, 50.8 ? 6.3%
organisms(Slegenthalerand Sarmiento 1993), may also
phytobenthos,
be oxidized by biotic and abiotic processes(i.e. photooxformacroalgae,57.1 ? 5.7% forseagrasses,69.4 +737/%for
idation; Mopper et al. 1991) when deep waters are upmarsh plants,and 52.1?9.4% formangroves.
welled into the upper ocean. The inputs of DOC to the
Herbivoryrepresentsthe consumptionof living plant
ocean must be balanced by these losses if the total DOC
material. The production decomposed annually represents use of detritalcarbon (dissolved and particulated) pool is to remain approximatelyconstant,so this pool
by decomposers.Direct estimatesofthisraterepresented does not representlong-termstorageof carbon (Mopper
only 10% (formacroalgae) to 50% (formangroves)of the
et al. 1991).
1760
Duarte and Cebrian
Mangroves
Marshplants
*
Seagrasses
Macroalgae
EC
Coral reefalgae
**
*
Microphytobenthos
Coastal phytoplankton
**
Oceanicphytoplankton
*
0
10
20
NPP (gCm -)d
-EI2--
Mangroves
EI
Marshplants
Seagrasses
*
m
Macroalgae
***
Coral reefalgae
Microphytobenthos
Coastalphytoplankton-11111
Oceanicphytoplankton,Zll
0
40
80
0
Decomposition
(% NPP)
40
80
Herbivory
(% NPP)
Mangroves
Marshplants
Seagrasses
*
Macroalgae
Coral reefalgae
F
Microphytobenthos
Coastalphytoplankton
Oceanicphytoplankton
0
40
Exportation
(% NPP)
80
0
40
80
Storage(% NPP)
Fig. 1. Distributionof NPP by different
marine autotrophsand the fractionof NPP that
is decomposed, consumed by herbivores,exported,and storedin the sediments.Boxes encompass 25% and 75% quartiles,the centralline representsthe median, and bars encompass
95% of the values. Asterisksand filledcirclesindicate observationsoutside the 95% limits.
Horizontal exportof oceanic phytoplanktonwas assumed to be negligible.
Export of photosyntheticcarbon representsthe horizontal transportof carbon away from the ecosystem.
Hence, our estimatesof exportedcarbon production(E)
do not include sedimentationof photosyntheticcarbon
away fromthe photic zone in the ocean, which, at the
annual scale, is eitherdecomposed, accumulated in deep
water, or is buried in the sediments. Whereas loss of
phytoplanktonicproductionby horizontaladvection in
Fate of marine autotrophicproduction
the open ocean can generallybe neglected,coastal ecosystems,wherehorizontaltransportoftenexceeds vertical
transport,may exportsubstantialphotosynthetic
carbon
(Wollast 1991; Prego 1993). However, estimationof exportedproductiondepends on the definitionof physical
boundariesforthesystemsstudied.Because thedefinition
oftheseboundariesrequiressubstantialknowledgeabout
the systems,we uncriticallyaccepted the boundaries establishedby the authorsof the estimates.
1761
40.6 ? 7.9
56.9 ? 8.7
34.2 ? 3.3
41.6 ? 2.7
Oceanic
<D
Coastal
D
phytoplankton
16.3
S
S
l
0.40 ? 0.19
V
Results
Discussion
The relativeimportanceof different
carbon pathways
among marine ecosystems(Fig. 2) reflectsboth the intrinsicpropertiesoftheproducersand thephysicalnature
of the environments.Differencesin the fractionof the
photosynthetic
productionconsumed by herbivoresmay
be relatedto the generaldecline in growthrate (Cebri'an
and Duarte 1994) and palatabilityand nutrient(N and
4.2
~~~~~~~~~~~~~~~~~~3.9il.8
29.8
42.9 ? 6.9
Marine primaryproductionrangedwidelyboth within
and among ecosystemsin the data set, with microphytobenthos and phytoplanktoncommunities supporting
relativelylow NPPs (Fig. 1). Most (-75%) of the NPP
of coral reefalgae was decomposed withinthe system.
This path comprised -40-50% of the NPP of othermarineplants,exceptformicrophytobenthos,
forwhichonly
-25% of NPP is decomposed withinthe system(Figs. 1,
2). Herbivorytendedto be highestformicroalgae(planktonic and benthic, > 40% of NPP) and macroalgae
(33.1?4.9% ofNPP) and was somewhatsmallerforhigher plants,exceptformarshplants,which may experience
substantialherbivory.Microphytobenthosexported on
average a much largerproportionof theirNPP than did
other microalgal communities,whereas marine macrophytes(exceptmarshplants)exporteda substantial(24.343.5%) proportionof theirNPP (Figs. 1, 2). The fraction
of NPP stored in the sediments(S) is 4-fold greaterfor
higherplants (- 10-17% of NPP) than foralgae (0.4-6%
of GPP), althoughS can be relativelyhigh in some microphytobenthic
communities.
On average, - 80-90% of the NPP of phytoplanktonis
used to supportlocal heterotrophic
metabolism(i.e. grazed
or decomposed), and this fractionis even higherin oceanic communities.The fractionofNPP used withincoral
reefecosystemsand marsh ecosystemsis also veryhigh
(> 80% of NPP). Mangrove forestsand, to a lesserextent,
seagrass meadows, microphytobenthos,
and macroalgal
beds produce organic carbon well in excess of the ecocarbon
systemrequirements,withexcess photosynthetic
(i.e. E + S) in these ecosystemsrepresenting-40% of
NPP (Fig. 2). Statisticalanalysis of the average carbon
budgets derived revealed significant(ANOVA on logtransformedvalues, P < 0.001) differencesamong ecosystemtypesin the fractionof NPP thatis decomposed,
eaten by herbivores,exported,and stored in the sediments.
E
phytoplankton
6
HH
7.
25.4? 3.4
.
E
28.8
l}
Coral reefalgae
D
Microphytobenthos E
<:D:4
7.9
11.4
6?4.9
S
33.6 ? 4.9
0.8?0.6
18.6 ? 5.3
50.3?2
37.3 ?14.7
<
2.3
D
E
Seagrasses
E
Macroalgae
24.3
435?8.8
?5.4
15.9 ? 6.7
0.4 ? 0.1
9.1 ? 2.4
13.3 ?7.2
40.1 ? 6.5
51.2 ? 6.6
D
Marshplants
D
E
18.6
Mangroves
E
29.5 ? 9.4
3.1
16.7?2.4S
10.4 ? 3.6
marineauFig.2. Averagecarbonbudgetsforthedifferent
The budgetsarederivedfromtheaverage(?SE) pertotrophs.
decomposedwithinthesysproduction
centageofnetprimary
(E), and stored
(H), exported
tem(D), consumedbyherbivores
these
in thesediments
(S). The widthsofthearrowsdenoting
ofnetprimary
production
areproportional
tothefraction
fluxes
100%
withthesidesof theboxesrepresenting
theyrepresent,
of NPP. Because theseestimateswerederivedby averaging
thebudgetsshownneednotadd up to
independent
estimnates,
100%.
P) concentrations(Duarte 1992, 1995) frommicroalgae
to higherplants, which have a need for nutrient-poor
structuraltissues(Enriquez et al. 1993). The thicknessof
the photosynthetictissues may also explain the differences in grazingintensity;that is, microscopicalgae can
be ingestedby a range of planktonicand benthicorgantisisms,whereasthecomparativelythickphotosynthetic
1762
Duarte and Cebrian
sues of higherplants and slow-growingmacroalgae may
diminish grazingintensity(Duarte 1995). Moreover, a
significantfractionof the productionof higherplants is
placed as roots and rhizomes in the sediments(Lipkin
1979; Prenticeand Fung 1990) and is thereforeunavailable to most herbivores.
The differencesamong ecosystemsin the fractionof
NPP that is decomposed may be attributableto similar
factors,because both tissue nutrientconcentrationsand
the thicknessand toughnessof photosynthetic
tissuesinfluencedecompositionrates (Melillo et al. 1982; Valiela
et al. 1984; Enriquez et al. 1993). Hence, microalgaldetritusand the carbon theyexcrete,withtheirhighernutrientconcentrations(Duarte 1992), are more readilydecomposed than are the tissuesof higherplants (Enriquez
et al. 1993). Additionally,the belowgroundproduction
of marine macrophytesis nutrient-poorrelativeto that
of photosynthetictissues (Walker and McComb 1988;
Short et al. 1993) and may be placed deep in the sediments,whereanaerobic conditionsmay preventefficient
decomposition (Harrison 1989). Higher plants also allocate a sizable fractionof the carbon they produce to
structuralcompounds that decompose slowly (Enriquez
et al. 1993). As a result,carbon storageis 4-foldgreater
forhigherplants than foralgae, except formicrophytobenthos,which may also accumulate substantialexcess
carbon.
Differencesin exportrates,on the otherhand, reflect
the physicalconditionsof the environmentsoccupied by
marine plants. For example, coastal plant communities
oftenoccupyareas thathave substantialhorizontaltransport. Accordingly,detritusderived fromthese plants is
likelyto be advected out ofthesystem(Nixon 1980; Lugo
et al. 1988; Hemminga et al. 1991). Similarly,coastal
phytoplanktonis oftenexportedout of the estuariesand
coastal areas where it grows(Welsh et al. 1972; Malone
and Chervin 1979).
in therelativeimportanceofdifferent
These differences
routesof photosynthetic
carbon transferhave major implicationsat the ecosystemlevel. Ecosystemsdominated
by autotrophicorganismsthatchannel most of theirproduction to herbivoresor decomposers should undergo
efficient
nutrientrecycling,withmost of theirproduction
representing,
undernear steady-statesituations,recycled
ratherthannewproduction.Similarly,thesecommunities
are likelyto show a close balance betweenphotosynthetic
productionand communityrespiration(P: R ratio - 1),
with new productionbeing very small relative to NPP
(Quinionesand Platt 1991). These communitiesare thereforelikelyto have a much smaller role as sinks for atmosphericcarbonthanis expectedfromtheirnetprimary
production.These effects
are particularly
evidentforcoral
reefalgae, wheretheclose couplingbetweenprimaryproducers and heterotrophsensures efficientnutrientrecycling and coupling between productionand respiration
(Crossland et al. 1991).
Althoughboth herbivoryand decomposition of phocarbon lead to regeneratedcarbon production
tosynthetic
and a tendencyof P: R ratiosto approach 1, theseeffects
should be even greaterwhendecompositionexceeds grazing.Ecosystemsthatchannelmostoftheirphotosynthetic
carbon to decomposers should be dominated by microbial ratherthan by metazoan food webs. Close coupling
productionand microbialgrowth
betweenphotosynthetic
of recyclingand reduce the
should increasethe efficiency
probabilityof exportof organiccarbon and nutrients,as
opposed to the greaterprobabilityof export associated
withthe motilityof herbivores.Additionally,ecosystems
dominated by autotrophsthat channel most of the productionto herbivoresare more likelyto supportefficient
productionof harvestablemarine protein(e.g. fisheries)
thanare those thatchanneltheirproductionto microbial
decomposers (Legendre and Rassoulzadegan 1995).
Whetherproductionis channeledto decomposersor herbivores is modulated, at least forplanktoniccommuniin communitystructureassociated
ties,by thedifferences
levels of turbulence.High turbulenceleads
withdifferent
to diatom-dominated phytoplankton communities
and relativelyquick(Cushing 1989), whichare efficiently
ly consumed by herbivoresor sedimentedout (cf.Legendre and Rassoulzadegan 1995). In contrast,low turbulence leads to autotrophiccommunities dominated by
nano- and picoplankton(Cushing 1989), whose producto decomposerswithinthe mition is mostlytransferred
crobial food web ratherthanto metazoan herbivores(Legendreand Rassoulzadegan 1995). Moreover,thesesmall
primaryproducershave negligiblesedimentationrates,
so thatmost of the carbon produced is used in the upper
ocean, leading to the verylow rates of carbon storageof
these ecosystems(Fig. 2).
Ecosystemsthat use most of the authochtonousprocoral
duction,suchas thosedominatedbyphytoplankton,
reefalgae, and marshes(herbivory+ respiration>80%
of NPP), must thereforeshow a close balance between
photosyntheticproduction and ecosystem respiration
(Odum 1956). Hence, small (<20% ofNPP) allocthonous
carbon inputs may drive these systemsto a net heterotrophicstatus(Odum 1956). These inputsmaybe derived
fromneighboringecosystems,such as benthicecosystems
and mangroves,which exporta sizable fraction(>20%,
Fig. 2) of theirnet primaryproduction (cf. Lugo et al.
1988; Lee 1990; Hemminga et al. 1991, 1995), or from
inputs fromland. Ecosystemsthat export a substantial
fractionof theirproductionand those in which a substantialfractionof the productionis buried in the sediments are likelyto lose substantialnutrientsalong with
the carbon. Sustained primaryproductionin these ecosystemsare thereforeclosely dependenton externalnutrientinputs.
As a result,the productionof ecosystemsthat either
export or bury an importantfractionof their primary
productionshould have a tendencyto be controlledby
nutrientsupply (i.e. bottom-up control), in contrastto
the expecteddominance of biotic control,such as microbial regenerationof nutrientsor herbivore control, in
ecosystemswhere primaryproductionis mostlygrazed
or decomposed. The relationshipbetweenexportand buried carbon and ecosystemnutrientlosses varies based on
Fate of marine autotrophicproduction
1763
Table 1. Estimates of total NPP by different
primaryproducersand the amount of this
productionthat is consumed by herbivores,decomposed, or storedin sediments.Values in
parenthesesare percentageoftotalherbivory,decomposition,or storagein theocean. Estimates
of NPP and the area covered by different
primaryproducerswere derived from different
sources,except fortotal NPP formangrovesand marsh plants,which were calculated from
our data set. The missingexportcarbon (decomposed or storedin sediments)is 5.5 Pg C
yr-I or 10% of the total NPP (see text).
Primaryproducer
Oceanic phytoplankton
Coastal phytoplankton
Microphytobenthos
Coral reefalgae
Macroalgae
Seagrasses
Marsh plants
Mangroves
Total
Area covered
Herbivory Decomposition
(106 kM2) Total NPP (Pg C yr-1)
332*
27*
6.8t
0.6?
6.8t
0.6t
0.4O
1.1**
-
43$
4.5t
0.34t
0.611
2.55t
0.49#
0.44
1.1
53.0
24.4(88)
1.8(6.5)
0.15(0.5)
0.18(0.6)
0.86(3.1)
0.09(0.3)
0.14(0.5)
0.10(0.3)
27.8(52)
14.7(77.5)
1.8(9.8)
0.09(0.4)
0.45(2.0)
0.95(4.2)
0.25(1.1)
0.23(1.0)
0.44(1.9)
19.0(36)
Storage
0.17(26.5)
0.18(27.0)
0.02(3.1)
0(0.7)
0.01(1.6)
0.08(12.0)
0.07(11.3)
0.11(17.6)
0.65(1.2)
* Sverdrupet al. 1942. tSmith and Hollibaugh 1993. tCharpy-Robaud and Soumia 1990.
?Smith 1978. 1Crossland et al. 1991. #deVooys 1979. ?Woodwell et al. 1973. **Lugo et al.
1988.
the C: N: P stoichiometryof the primaryproducersand
the detritusthey produce. Most of the organic matter
exportedand buried by highermarineplants (e.g. marsh
plants, seagrasses,mangroves) correspondsto nutrientpoor detritus(e.g. Hemminga et al. 1991, 1995), so the
nutrientinputsneeded to maintainthe NPP of these systems may be less than those expected fromthe carbon
losses. Similarly,use of the exported carbon in neighboring ecosystemsdominated by nutrient-rich
primary
producers (e.g. phytoplankton)may require additional
nutrientinputs.Conversely,the importof particulateorganic matterderivedfromnutrient-rich
primaryproducers to systemsdominated by nutrient-poorprimaryproducers(e.g. seagrasses;cf.Duarte 1992) mayincreaseboth
heterotrophic
carbon use and, particularly,
local primary
productiononce the nutrientsare remineralized.
These argumentsshow that predictionsof ecosystem
cannotbe reliablyderivedfromexaminations
functioning
of local carbon budgetsalone and that such predictions
mustconsiderthecontinuitybetweenmarineecosystems.
For example,carbonexportfrommarineecosystemsdoes
not necessarilyrepresenta largeloss of carbon,and only
the carbon storedin sedimentsmay represent,when effectivelyburied,a netcarbon sink.Based on examination
of different
ecosystemtypes,we attemptedto scale our
resultsto the global ocean by firstcalculatingthe fateof
thetotalcarbon producedby different
marineecosystems
and then compoundingthese calculations into a carbon
budgetforthe global ocean similar to that described by
Eq. 1. The productof the average NPP of different
types
of marineecosystemsby the area theycover in the global
ocean allowed an estimate of theirglobal NPP. Phytoplanktonproductionin the open ocean by fardominates
(~ 80%) marineNPP. Less than20% oftotalmarineNPP
is contributedby all otherecosystems,of which coastal
microalgae and macroalgae are the main contributors
(Table 1). The amounts of total NPP consumed by herbivoresand decomposedin different
ecosystemtypeswere
proportionalto the contributionof these ecosystemsto
total NPP (Table 1). Most (88%) herbivoryin the sea
takes place in the open ocean and representsa greater
percentageof the open ocean NPP (56.9%, Fig. 2) than
that of coastal NPP (34%, Table 1). Assuminga similar
in the transformation
of herbivoreproduction
efficiency
into harvestablebiomass forthe coastal and open ocean,
these calculationssuggestthatthe fractionof NPP available forfisheriesharvestshould be roughlyequal forthe
open and coastal ocean and thatthe harvestableproduction should be about 10-foldhigherin the open ocean.
Therefore,the reportthat the fractionof NPP required
to supportcoastal fisheriesis about 10-15 times greater
thanthatin the open ocean (Pauly and Christensen1995)
efficannot be accounted foron the basis of differential
ciencies in the conversion of NPP into harvestablebiomass. Instead, these differencessuggesta much greater
in the coastal ocean than in the open
capture efficiency
ocean, wherethefishbiomass is dilutedin a much thicker
watercolumn.
Unlike decomposition and herbivory,the total NPP
marine ecosystemsis independentof
storedby different
theirtotal production(Table 1), as is expected fromthe
generalindependencebetweendetritalcarbon storageand
ecosystemprimaryproduction(Cebri'anand Duarte 1995).
In particular,marine ecosystems dominated by angiosperms(seagrasses,marshplants,and mangroves)are responsiblefor - 30% of the totalNPP storedin the ocean,
whereastheycontributeonly - 4% ofthetotalNPP (Table
1). Moreover, because carbon burial efficiency(i.e. the
fractionofthecarbon storedthatremainsburiedover the
long term)is much greaterforcoastal than fordeep-sea
1764
Duarte and Cebrian
thesea as majorcontributors
to thecarbon sinkassociated
withmarinephotosynthesis
(- 30% ofthetotalsink).This
(52 %)
Herbivory
excess of marine photosyntheticcarbon does not imply
thatthe ocean is a net autotrophicsystem.Burial of marine photosyntheticcarbon (-0.15 Pg C yr-') is much
i
27.8
less thanthe inputof organiccarbon fromterrestrial
ecosystems,which is -4-fold greater(0.4 Pg C yr-' from
rivers,0. 15 Pg C yr-1 fromdirectsewageinputs,and 0. 10
Pg C yr-I as atmosphericdeposition of land-derivedorTotalNPP
ganic carbon; Schlesinger 1991; Smith and Hollibaugh
Decomposition 19
1993). Hence, use by marine heterotrophsof only 25%
53
(36 %)
(10 %)
1 Redistribution
oftheorganiccarbon derivedfromland would renderthe
ocean a net heterotrophicsystem,as suggestedby previous analyses (Smith and Mackenzie 1987; Smith and
Hollibaugh 1993). Of course, the input of land-derived
organiccarbon to the ocean is not uniformlydistributed,
0.65
so its importancein drivingthe systemstowardnet heterotrophymustbe greaterin thecoastal ocean (Smithand
Stored(1.2%)
Hollibaugh 1993).
The data presentedhere help us to evaluate the role of
/ 0.15
marine plant types in local carbon budgets by
different
Buried(0.25%)
establishingtrends in the average fractionof their net
Fig.3. A budgetforthetotalmarinenetprimary
production. primaryproduction that is channeled to heterotrophs
carAll valuesinsidebox and arrowsare in Pg C yr-1.Box con- throughgrazingor decompositionof photosynthetic
struction
as in Fig. 2. Dottedlinesrepresent
theredistribution bon, as well as trendsin the excess carbon storedin the
ofNPP thatis exported
fromoneecosystem
to anotherandthat sedimentsor exportedto adjacent ecosystems.From the
musteventually
be decomposedor buriedelsewhere.
resultspresented,we can identifya declinein thefractions
of the net primaryproductionused withinthe systems
and
consumed by herbivores,as well as an incrementin
sediments(-30% and 3%, respectively;cf. Berner 1982;
the fractionof NPP stored in sediments,from phytoLein 1984; Henrichsand Reeburgh1987), theNPP buried
plankton to angiosperm-dominatedecosystems. Comin the coastal ocean is likelyto be much greaterthanthat
munitiesdominatedby higherplantstendto be net autoburied in the deep ocean (-0.145 Pg C yr-' and 0.005
trophicand to act as carbon sinks, suggestingthat their
Pg C yr-', respectively,as calculated fromTable 1).
autotrophicproduction should be largelycontrolledby
The estimatespresentedmayunderestimate
by as much
resources. In contrast,planktonic communities should
as 50% the NPP of coastal plants buried if only 10% of
have a close balance betweenautotrophicand heterotrothe NPP exportedby coastal plants (i.e. 0.2 Pg C yr'-)
phic metabolism.Consumersare thereforelikelyto exert
is storedand subsequentlyburied (i.e. -0.06 Pg C yr-')
a close controlon the autotrophicproductionof plankin deep-sea sediments.The 3-folddifference
betweenNPP
toniccommunities,whichshould be largelysupportedby
storedin coastal vs. deep marinesedimentsis quite subrecyclednutrients.Our results provide a basis for forstantial,and that betweenthe NPP buried in coastal vs.
mulatingand testinghypotheseson the functioningof
deep marine sediments is even larger(- 30-fold). The
marine ecosystemsand the global ocean. Extrapolation
conservativeestimate of total NPP burial in the ocean
oftheseresultsto theglobal ocean identifiesmarinehigher
derived fromthese calculations(0.15-0.21 Pg C yr'-) is
which only contribute4% of total ocean NPP, as
plants,
close to available estimatesof this rate (0. 13 Pg C yr-1,
Berner1982). Carbon storage,however,is a small fraction major contributorsof the NPP stored in marine sediments (30% of total ocean carbon storage) and subse(1.2%) of the total marine NPP, most of which is conquentlyburied in the ocean. Futureattemptsto estimate
sumed by herbivoresor decomposed (88%; Table 1, Fig.
the potential of marine photosynthesisto sequester at3). The budget of total marine NPP obtained shows a
mosphericcarbon mustthereforeconsidermarineangiodominance of herbivoryin theuse of marineNPP (52%),
spermsas importantcontributorsto this process.
whichis greaterthan microbialuse of total NPP (36% of
NPP; Table 1, Fig. 3). However, theexcess carbon is only
-1% of the total NPP, three-fourths
of which is conReferences
sumed by diageneticprocesses in the sediments,so that
only -0.25% of the total marine NPP is buried in the
ADMIRAAL, W., L. BOUWMAN, L. HOEKSTRA, AND D. ROMEYN.
sediments(Fig. 3).
1983. Qualitative and quantitative interactions between
Burial of marine NPP representsan estimate of the
microphytobenthos and herbivorous meiofauna on a brackexcessautotrophicproductionin theocean and, therefore,
ish intertidalmudflat.Int. Rev. Gesamten Hydrobiol. 68:
the size of the carbon sink associated with marine pho175-191.
tosynthesis.Our resultsidentifyangiospermsgrowingin
BERNER,R. A. 1982. Burial oforganiccarbonand pyritesulfur
Fate of marineautotrophicproduction
in the modern ocean: Its geochemicaland environmental
significance.Am. J. Sci. 282: 451-473.
BOYNTON,W. R., W. M. KEMP,AND C. W. KEEFE. 1982. A
comparativeanalysis of nutrientsand otherfactorsinfluencing estuarinephytoplanktonproduction,p. 69-90. In
V. S. Kennedy [ed.], Estuarinecomparisons.Academic.
CARPENTER,
R. 1986. Partitioningherbivoryand its effects
on
coral reefalgal communities.Ecol. Monogr. 56: 345-363.
1765
A., AND G. BATE. 1984. Carbon budgetfora high
energysurfzone. Vie Milieu 34: 67-77.
MALONE, T., AND M. CHERVIN. 1979. The productionand fate
of phytoplankton
size in fractionsin theplume ofthe Hudson River, New York Bight. Limnol. Oceanog. 24: 683696.
MARTIN, J.H., G. A. KNAUER, D. M. KARL, AND W. W. BROENKOW. 1987. VERTEX: Carbon cyclingin the northeast
CEBRIAN J., AND C. M. DUARTE. 1994. The dependence of
Pacific.Deep-Sea Res. 34: 267-285.
herbivoryon growthrate in natural plant communities.
MELILLO, J.,J.D. ABER, AND J. M. MURATORE.
1982. Nitrogen
Funct. Ecol. 8: 518-525.
and lignincontrol of hardwood leaf litterdecomposition
, AND
. 1995. Plant growth-rate dependence of
dynamics.Ecology 63: 621-626.
detritalC-storagein ecosystems.Science 268: 1606-1608.
AND OTHERS.
1993. Global climate change and terrestrialnet primaryproduction.Nature 363: 234-239.
CHARPY-ROUBAUD, C., AND A. SOURNIA. 1990. The comparative estimation of phytoplanktonicmicrophytobenthic MOPPER, K., AND OTHERS. 1991. Photochemicaldegradation
productionin the oceans. Mar. Micr. Food Webs 4: 31-57.
of dissolved organiccarbon and its impact on the oceanic
CROSSLAND, C. J., B. G. HATCHER, AND S. V. SMITH. 1991.
carbon cycle.Nature 353: 60-62.
Role of coral reefsin global ocean production.Coral Reefs
NIXON, S. W. 1980. Betweencoastal marshesand coastal wa10: 55-64.
ters-a reviewof twentyyearsof speculationand research
in structurebetweenecoon the role of salt marshes in estuarineproductivityand
CUSHING,D. H. 1989. A difference
watersand thosethatare only
water chemistry,p. 437-525. In P. Hamilton and K. B.
systemsin stronglystratified
weaklystratified.
J. PlanktonRes. 11: 1-13.
MacDonald [eds.], Estuarine and wetland processes. PleDE VooYs, C. G. N. 1979. Primaryproduction in aquatic
num.
environments,p. 259-292. In B. Bolin and others [eds.],
, C. A. OVIATT, J. FRITHSEN, AND B. SULLIVAN. 1986.
The global carbon cycle. Wiley.
Nutrientsand theproductivity
ofestuarineand coastal ecoDUARTE,C. M. 1992. Nutrientconcentrationofaquatic plants.
systems.J. Limnol. Soc. S. Afr.12: 43-7 1.
Patternsacross species. Limnol. Oceanogr. 37: 882-889.
ODUM, H. T. 1956. Primaryproductionin flowingwaters.
. 1995. Submerged aquatic vegetationin relation to
Limnol. Oceanogr. 1: 102-117.
different
nutrientregimes.Ophelia 41: 87-112.
1995. Primaryproduction
PAULY, D., AND V. CHRISTENSEN.
requiredto sustainglobal fisheries.Nature 374: 255-257.
ENRIQUEZ,S., C. M. DUARTE, AND K. SAND-JENSEN. 1993. Patternsin decompositionratesamong photosynthetic
organPLATT, T. AND D. V. SUBBA RAO. 1975. Primaryproduction
isms: The importanceof detritusC: N: P content.Oecoof marinemacrophytes,p. 249-280. In J. P. Cooper [ed.],
Photosynthesis
and productivityin different
logia 94: 457-47 1.
environments.
areas of
FROST,B. 1991. The role of grazingin nutrient-rich
Cambridge.
the open sea. Limnol. Oceanogr. 36: 1616-1630.
PREGO, R. 1993. General aspects of carbon biogeochemistry
P. G. 1989. Detritalprocessingin seagrasssystems:
in the ria of Vigo, northwestern
HARRISON,
Spain. Geochim. CosmoA reviewof factorsaffecting
decay rates,remineralization
chim. Acta 57: 2041-2052.
and detritivory.
Aquat. Bot. 23: 263-288.
PRENTICE, K. C., AND I. FUNG.
1990. The sensitivityof terHEMMINGA,M. A., P. G. HARRISON, AND F. VAN LENT. 1991.
estrialcarbon storageto climatic change. Nature 346: 48The balance of nutrientlosses and gains in seagrassmead51.
ows. Mar. Ecol. Prog. Ser. 71: 85-96.
1991. The relationshipbeQUINONES, R. A., AND T. PLATT.
, AND OTHERS. 1995. Carbon outwelling from a mantweenthef-ratioand theP: R ratioin thepelagicecosystem.
grove forestwith adjacent seagrass beds and coral reefs
Limnol. Oceanogr. 36: 211-213.
(Gazi Bay, Kenya). Mar. Ecol. Prog. Ser. 106: 291-301.
ROMERO, J., M. PtREZ, M. A. MATEO, AND E. SALA. 1994. The
HENRICHS,S. M., AND W. S. REEBURGH. 1987. Anaerobic minbelowgroundorgans of the Mediterraneanseagrass Posieralizationof marine sedimentorganicmatter:Rates and
donia oceanica as a biogeochemical sink. Aquat. Bot. 47:
the role of anaerobic processesin the oceanic carbon econ13-19.
omy. Geomicrobiol.J. 5: 191-237.
RYTHER, J. H. 1963. Geographic variations in productivity,
LAWS,E. A., P. K. BIENFANG,D. A. ZIEMAN,AND L. D. CONQUEST.
p. 347-380. In M. N. Hill [ed.], The sea. V. 2. Interscience.
1988. Phytoplanktonpopulationsdynamicsand the fateof
An analysis of
SCHLESINGER, W. H. 1991. Biogeochemistry:
productionduringthe springbloom in Auke Bay, Alaska.
global change. Academic.
Limnol. Oceanogr. 33: 57-65.
SHORT, F. T., J. MONTGOMERY, C. F. ZIMMERMANN, AND C. A.
SHORT. 1993. Production and nutrient dynamics of a Syand particulateorganic
LEE,S. Y. 1990. Primaryproductivity
ringodiumfiliformeKutz seagrass bed in Indian River lamatterflow in an estuarine mangrove-wetlandin Honggoon, Florida. Estuaries 16: 323-334.
Kong. Mar. Biol. 106: 453-463.
1993. Atmospheric
SLEGENTHALER, U., AND J. L. SARMIENTO.
LEGENDRE,L., AND F. RASSOULZADEGAN. 1995. Plankton and
carbon dioxide and the ocean. Nature 365: 119-125.
nutrientdynamicsin marinewaters.Ophelia 41: 153-170.
SMITH, S. V. 1978. Coral-reefarea and the contributionsof
LEIN,A. Y. 1984. Anaerobic consumptionof organicmatter
reefsto processesand resourcesof theworldocean. Nature
in modernmarine sediments.Nature 312: 148-150.
273: 225-226.
LIPKIN,Y. 1979. Quantitativeaspects of seagrass communi1981. Marine macrophytesas a global carbon sink.
ties, particularlyof those dominated by Halophila stipuScience 211: 838-840.
lacea, in Sinai (NorthernRed Sea). Aquat. Bot. 7: 119-128.
1993. Coastal metabolismand
LUGO, A. E., S. BROWN,AND M. M. BRINSON. 1988. Forested
,AND J.T. HOLLIBAUGH.
theoceanic organiccarbon balance. Rev. Geophys. 31: 75wetlandsin freshwater
and salt-waterenvironments.Lim89.
nol. Oceanogr. 33: 894-909.
McLACHLAN
1766
Duarte and Cebrian
1987. The ocean as a net het, AND F. T. MACKENZIE.
erotrophicsystem:Implications fromthe carbon biogeochemical cycle. Global Biogeochem. Cycles 1: 187-198.
STROM, S., AND N. WELSCHMEYER.
1991. Pigmentspecificrates
of phytoplanktongrowthandmicrozooplanktongrazingin
the open subarcticPacific Ocean. Limnol. Oceanogr. 36:
50-63.
1942.
SVERDRUP, H. U., M. W. JOHNSON, AND R. H. FLEMING.
The oceans. Prentice-Hall.
VALIELA, I. J., AND OTHERS.
1984. Importance of chemical
compositionof salt marshlitteron decay ratesand feeding
by detritivors.Bull. Mar. Sci. 35: 261-269.
WALKER, D. I., AND A. J. MCCOMB. 1988. Seasonal variation
in the production,biomass and nutrientstatus of Amphibolls antartica(Labill.) Sonders Ex. Aschersand Posidonia
australisHook F. in Shark Bay, WesternAustralia.Aquat.
Bot. 31: 259-275.
WELSH,E. B., R. M. EMEREY,R. I. MATSUDAANDW. A. DAWSON.
1972. The relationshipof algal growthin an estuaryto
hydrographicfactors.Limnol. Oceanogr. 17: 734-737.
WOLLAST,R. 1991. The coastal organiccarbon cycle: Fluxes,
sourcesand sinks,p. 365-382. In R. Montoura [ed.], ocean
marginprocessesin global change. Wiley.
WOODWELL,G. M., P. H. RICH, AND C. S. A. HALL. 1973.
Carbon in estuaries,p. 221-240. In G. M. Woodwell and
E. V. Pecan [eds.], carbon in the biosphere.U.S. AEC.
Submitted:1 June 1995
Accepted:5 February1996
Amended: 19 June 1996