COMMENTARY

Prevalence of Heterotrophy and

Atmospheric CO
2
Emissions from
Aquatic Ecosystems
Carlos M. Duarte,
1
* and Yves T. Prairie
2
1
IMEDEA (CSIC-UIB), Instituto Mediterra neo de Estudios Avanzados, Miquel Marque s 21, 07190 Esporles, Spain;
2
De partement des
Sciences Biologiques, Universite du Que bec a` Montre al, Case postale 8888, succ. Centre-Ville, Montre al H3C 3P8, Canada
ABSTRACT
Recent, parallel developments in the study of
freshwater and marine ecosystems have provided
evidence that net heterotrophic systems (those in
which respiratory organic matter destruction exeeds
photosynthetic production) are more prevalent than
hitherto believed, including most rivers, oligo- to
mesotrophic lakes and some oligotrophic regions of
the ocean. In parallel, these aquatic ecosystems have
beenshown to act as CO
2
sources to the atmosphere,
as expected from the heterotrophic nature of the
communities they contain. The prevalence of net
heterotrophic aquatic ecosystems indicates that they
must receive significant inputs of organic carbon
from adjacent ecosystems, assigning an important
role to the lateral exchanges of carbon between land
and aquatic ecosystems, between coastal and open
ocean ecosystems, as well as internal redistribution
within large or complex aquatic ecosystems in
determining their metabolic status and the gaseous
exchange with the atmosphere. The examination of
the carbon budget of ecosystems requires, therefore,
an integrative approach that accounts for exchanges
between compartments often studied in isolation.
These recent findings conforma newparadigmof the
functioning of aquatic ecosystems, and the meta-
bolic connectivity between ecosystems in the bio-
sphere.
Key words: aquatic ecosystems; metabolism;
heterotrophy; respiration; CO
2
.
INTRODUCTION
Aquatic ecosystems, freshwater and marine, cover
77% of the surface of the Earth and contribute
more than half of the global primary production
(Field and others 1998). Although this primary
production supports highly productive food-webs,
the aquatic environment is also the site of intense
catabolism of the organic carbon originating from
this internal production and materials exported
from land (Schlesinger and Melack 1981;
Schlesinger 1991; Cole 1999; Richey and others
2002). Aquatic ecosystems are, therefore, critical
components of the metabolism of the biosphere
and the associated CO
2
exchange with the atmo-
sphere. The contribution of aquatic ecosystems to
these processes has been the subject of recent
parallel debates as to the occurrence of net het-
erotrophic metabolism in marine and freshwater
ecosystems (Smith and MacKenzie 1987; Smith
and Hollibaugh 1993; del Giorgio and others 1997;
Duarte and Agust 1998; Williams 1998; Duarte
and others 1999; 2001; Carignan and others 2000;
Prairie and others 2002; Takahashi and others
2002; Karl and others 2003; del Giorgio and
Williams in press-a), a central point of which is the
Received 7 December 2003; accepted 11 August 2004; published online
21 October 2005.
*Corresponding author; e-mail: cduarte@uib.es
Ecosystems (2005) 8: 862870
DOI: 10.1007/s10021-005-0177-4
862
importance of aquatic respiration as a source of CO
2
(del Giorgio and Duarte 2002).
Historical emphasis on primary production and,
particularly, the study of lakes and oceans as iso-
lated ecosystems led to the generalized expectation
that aquatic ecosystems ought to be net auto-
trophic, with production exceeding respiration.
However, our plea is that the weight of evidence is
now shifting considerably: whereas net autotrophy
was until recently considered the normal func-
tioning of most aquatic systems, we contend that
net heterotrophy is far more prevalent than previ-
ously believed. This contention is based on the
growing evidence that respiration exceeds primary
production in many aquatic ecosystems, which are
therefore net heterotrophic (for example, Smith
and MacKenzie 1987; Smith and Hollibaugh 1993;
del Giorgio and Peters 1994; Duarte and Agust
1998; Duarte and others 2001; del Giorgio and
Duarte 2002; Hoppe and others 2002; Agust and
others 2001; Gonza lez and others 2001; Harrison
and others 2001; Arstegui and Harrison 2002;
Prairie and others 2002; Chen and others 2003;
Hanson and others 2003; Agust and others 2004;
Caffrey 2004). The shift in our perception of the
prevalence of net heterotrophy in aquatic ecosys-
tems has important consequences for our under-
standing of their role in the global carbon cycle.
Here we integrate these parallel debates to produce
a coherent framework of the role of aquatic eco-
systems on the metabolism and the associated gas
exchange of the biosphere.
METABOLISM IN AQUATIC SYSTEMS
The metabolism of aquatic ecosystems, the pro-
duction and destruction of organic matter derived,
respectively, from the gross primary production
(GPP) and respiration (R) of the communities, was
proposed half a century ago as a powerful, syn-
thetic descriptor of ecosystem functioning (Odum
1956). Odum (1956) characterized aquatic ecosys-
tems as net autotrophic when GPP was greater than
R, and net heterotrophic when GPP was less than
R. However, H.T. Odum postulated that most
unpolluted aquatic ecosystems would show a bal-
anced or net autotrophic metabolism, and only
organic-polluted and distrophic (that is, humic)
ecosystems were expected to be net heterotrophic
(Odum 1956). The traditional view that aquatic
systems in general were fed through the photo-
synthetic activity of their primary producers has
been radically altered as evidence for the connec-
tivity between land and aquatic ecosystems in-
creases (for example, Cole and Caraco 2001; Richey
and others 2002; Caraco and Cole 2004; Pace and
others 2004; Vanni and others 2004). Upon enter-
ing aquatic ecosystems, organic carbon exported
from land undergoes chemical and photo-chemical
alterations rendering it more amenable to bacterial
degradation, which can potentially raise aquatic
respiration beyond the limits imposed by aquatic
photosynthesis (Cole 1999; Scully and others
2003). This situation is most likely met in unpro-
ductive aquatic ecosystems, which are often found
to be, consequently, net heterotrophic (del Giorgio
and others 1997; Duarte and Agust 1998; Duarte
and others 2001). Recent syntheses of the meta-
bolic balance of aquatic ecosystems have revealed
that respiration tends to dominate over production
in many rivers, estuaries, oligotrophic to meso-
trophic lakes, reservoirs, and regions of the oligo-
trophic ocean (Smith and Hollibaugh 1993; del
Giorgio and Peters 1994; Duarte and Agust 1998;
Prairie and others 2002; Caffrey 2004), whereas
littoral, particularly macrophyte-dominated eco-
systems, nutrient-rich lakes or DOC-poor lakes,
and coastal lagoons tend to be autotrophic (Duarte
and Cebria n 1996; Duarte and Agust 1998; Gattuso
and others 1998; Carignan and others 2000). The
net community metabolism, the difference be-
tween gross production and community respira-
tion, is usually a small fraction of the magnitude of
the metabolic uxes of GPP and R, which are often
in close balance (Williams 1998; Duarte and Agust
1998; Williams 2000).
The average aquatic primary production re-
quired to generate an excess organic carbon sup-
ply over respiratory losses varies considerably but
consistently across aquatic ecosystems (Figure 1;
compare Duarte and Agust 1998), depending on
the allochthonous organic loading they receive.
The aquatic primary production required to bal-
ance the community respiration demand is high-
est, on average, for rivers and marshes, which
receive large amounts of organic carbon from
their watersheds, and must reach an average level
of 2.65 g O
2
m
)3
d
)1
before primary production
balances respiratory demands (Figure 1). Aquatic
primary production is lowest in the ocean, where
a primary production of at least 0.035 g O
2
m
)3
d
)1
is required for planktonic communities to
become autotrophic (Duarte and Agust 1998).
This threshold is low, but exceeds the areal-
weighed average primary production in the open
ocean (Knauer 1993). These threshold estimates
have been validated by subsequent, independent
studies. For instance, a threshold of 0.03 g O
2
m
)3
d
)1
was found in the NE subtropical Atlantic
(assuming a 100 m mixed layer, Duarte and
Prevalence of Heterotrophy in Aquatic Ecosystems 863
others 2001), and 0.032 g O
2
m
)3
d
)1
in the N.
Atlantic Subtropical Gyre (Gonza lez and others
2001), both remarkably close to that of 0.035 g
O
2
m
-3
d
)1
derived originally (Duarte and Agust
1998). A threshold gross primary production of
0.12 and 0.13 g O
2
m
)3
d
)1
; both these estimates
derived independently of one another, was re-
ported for the Mediterranean littoral (Duarte and
others in press), and a threshold of 12.8 g O
2
m
)2
d
)1
was derived for estuaries (Hopkinson and
Smith in press), equivalent to the volumetric
threshold of 1.62 g O
2
m
)3
d
)1
derived by Duarte
and Agust (1998) for estuaries, on average, 8 m
deep. The gross primary production realized in
aquatic ecosystems is often below those thresh-
olds (Duarte and Agust 1998; Duarte and others
in press; Hopkinson and Smith in press; Robinson
and Williams in press). Accordingly, a significant
fraction of the freshwater ecosystems studied to-
date are net heterotrophic (39 % of lentic eco-
systems and 45 % of rivers in the data base
compiled by Duarte and Agust [1998]), most
coastal ecosystems dominated by macrophytes,
such as seagrass meadows, macroalgal beds, and
mangrove forests, all of which are highly pro-
ductive (Duarte and Cebria n 1996), are net
autotrophic (Duarte and Cebria n 1996; Duarte
and Agust 1998; Gattuso and others 1998;
Hemminga and Duarte 2000), and calculations
suggest that much of the planktonic communities
in 70 % of the area of the open ocean encom-
passed by the oligotrophic open ocean may be net
heterotrophic (Duarte and Agust 1998; del Gior-
gio and Duarte 2002). Indeed, there are elements
to argue that the global ocean is a heterotrophic
system (Smith and MacKenzie 1978; Duarte and
others 1999; del Giorgio and Duarte 2002;
Ducklow and McCallister in press).
Realization of the prevalence of net heterotro-
phic aquatic ecosystems (Duarte and Agust 1998)
has prompted increased attention to the quanti-
cation of aquatic respiration rates, of which there is
still a rather primitive understanding compared to
the considerable body of knowledge acquired over
the past decades on aquatic photosynthesis
(Williams 2000; del Giorgio and Duarte 2002; del
Giorgio and Williams in press-a, b). These efforts
are providing evidence that heterotrophic aquatic
ecosystems are more prevalent than hitherto rep-
resented in the available literature, a result of our
collective bias towards the study of community
metabolism of relatively productive ecosystems
(Duarte and others 1999). This situation is exem-
plied by the recent and rapid accumulation of
reports of a net heterotrophic microplankton
community in the subtropical Atlantic Ocean
(compare Duarte and others 2001; Serret and oth-
ers 2001; Hoppe and others 2002; Agust and others
2001; Gonza lez and others 2001; Harrison and
others 2001; Arstegui and Harrison 2002), which is
in stark contrast with the depiction of an overall
metabolic balance possible from the scattered data
available only ve years ago in this and other re-
gions of the ocean (Williams 1998).
Awareness of the prevalence of heterotrophy in
aquatic ecosystems has, in turn, directed attention
Figure 1. The fitted power regression equations
describing the relationship between gross primary pro-
duction (GPP) and respiration (R) in different types of
aquatic ecosystems a, and b the threshold primary pro-
duction required to render various types of aquatic eco-
systems net autotrophic (that is, GPP at GPP = R), as
derived from the intercept between the fitted regression
lines and the 1 to 1 line (that is, GPP at GPP = R). Data
from Duarte and Agust (1998).
864 C. M. Duarte and Y. T. Prairie
to the quantication of the allochthonous inputs of
organic carbon to aquatic ecosystems, which are
necessary to support the excess respiratory activity
(Duarte and Agust 1998; Duarte and others 1999;
Cole del Giorgio and Duarte 2002). These efforts
are starting to provide evidence of a general cor-
respondence between the DOC concentration and
the metabolic status of aquatic ecosystems. For in-
stance, recent results have identied a threshold of
400500 lM DOC separating autotrophic from
heterotrophic lakes in Canada and northern USA
(Hope and others 1994; Prairie and others 2002;
Hanson and others 2003). Most importantly, these
ndings are driving current research towards a
greater appreciation of the integration among
aquatic ecosystems and between these and terres-
trial ecosystems, suggesting that a substantial frac-
tion of the terrestrial primary production is
exported away and respired in aquatic ecosystems
(for example, Cole and Caraco 2001; Richey and
others 2002; Houghton 2003; Caraco and Cole
2004). Respiration in the sediments can be an
important component of this metabolism (Cole and
others 2000; Raymond and others 2000), especially
in small or shallow ecosystems (Pace and Prairie in
press). Similarly, growing evidence for the presence
of heterotrophic planktonic communities in the
open oligotrophic ocean is driving efforts to inves-
tigate of the organic carbon exports from auto-
trophic coastal ecosystems to the open ocean (for
example, Duarte and others 1999; Liu and others
2000, 2002; Ducklow and McCallister in press).
Models of the sources and fates of organic C in
aquatic systems also point to widespread net
heterotrophy in lakes and rivers. Caraco and Cole
(2004) examined the relative magnitudes of al-
lochthonous and autochthonous support of
aquatic metabolism in lakes and rivers and pro-
duced a model that suggests that net heterotro-
phy must be the most prevalent case in lakes and
rivers, despite ndings that allochthonous organic
carbon inputs are respired faster in rivers than in
lakes (Caraco and Cole 2004). Further, the model
developed by Caraco and Cole (2004) predicts
that the respiration of allochonous organic carbon
should contribute up to 50 % of total lake res-
piration in oligotrophic lakes with high organic
carbon inputs, declining as nutrient inputs in-
crease and organic carbon decrease, a prediction
substantiated by cross-lake comparative analyses
(Prairie and others 2002; Hanson and others
2003). Carbon budget models in streams also
point to widespread heterotrophy supported by
allochthonous, land-derived, carbon inputs
(Webster and Meyer 1999).
These ndings highlight the importance of lateral
transport processes, and should, therefore, promote
a shift from 1-D models to 3-D models of organic
carbon cycling in aquatic ecosystems that better
integrate ecosystems spatially. For instance, net
autotrophic coastal benthic communities may de-
liver the organic carbon needed to support net
heterotrophic planktonic communities (for exam-
ple, Lucea and others in press). Net heterotrophy
may also result from organic carbon exchanges
within large ecosystems, such as oceanic basins.
Hansell and others (2004) showed that transport of
dissolved organic carbon is sufcient to render the
North Atlantic net heterotrophic. The integration
between the carbon budgets of terrestrial and
aquatic components of the biosphere is, however,
still pending, despite evidence derived from
examination of carbon budgets in either one of
these components that terrestrial ecosystems must
export substantial amounts of carbon to aquatic
components of the biosphere (Houghton 2003),
and that aquatic components must receive sub-
stantial inputs of allochthonous carbon (for exam-
ple, Duarte and Agust 1998; Cole and Caraco 2001;
Prairie and others 2002; Richey and others 2002;
Aitkenhead and McDowell 2000; Mulholland
2003). An integrative approach to the examination
of the carbon budget of the biosphere may also help
resolve and understand apparent imbalances at
lower scales.
As indicated earlier, the growing evidence for net
heterotrophy in aquatic ecosystems has met oppo-
sition both in freshwater (Carignan and others
2000) and the open sea (Williams and Bower 1999;
Karl and others 2003; Robinson and Williams in
press). Controversy about the occurrence of net
heterotrophy in lakes was resolved by identifying
that heterotrophy was more probable in lakes with
relatively high dissolved organic carbon concen-
tration (Prairie and others 2000; Hanson and others
2003). The prevalence of net heterotrophy in open
sea communities has been postulated to be partially
an artifact derived from a possible underestimation
of oceanic GPP, due to (1) inadequacy of present
estimates of oceanic primary production, based on
14
C uptake (del Giorgio and Duarte 2002; del
Giorgio and Williams in press-a), and (2) the belief,
based on the observation of a pulse of O
2
to N
2
disequilibria off Hawaii, that much of GPP may
occur in infrequent, but significant pulses that
elude observation (Karl and others 2003). Whereas
there is consensus that the
14
C uptake underesti-
mates oceanic GPP, the evidence for sampling bias
is very limited. Frequency distributions of R and
GPP, in the Southern Ocean (Agust and others
Prevalence of Heterotrophy in Aquatic Ecosystems 865
2004), the NE Subtropical Atlantic (Duarte and
others 2001), and the Mediterranean littoral
(Duarte and others in press), do not support the
existence of a greater skeweness in GPP relative to
R, and O
2
to N
2
disequilibria concurrent with
temperature anomalies may result from their dif-
ferent solubilities rather than net autotrophy
(Compare data in Karl and others 2003). Moreover,
the imbalance between R and GPP in the ocean
grows much larger when considering the layer
extending below the photic zone (del Giorgio and
Duarte 2002), and is, consequently, more difficult
to account for methodological artifacts. Inputs of
organic carbon from the atmosphere and from the
autotrophic coastal ocean are potentially large
(Duarte and others 1999; del Giorgio and Duarte
2002), possibly driving the open ocean ecosystem
to net heterotrophy. However, these inputs are
rather poorly constrained at present (del Giorgio
and Duarte 2002). In summary, the occurrence of
net heterotrophy in open ocean communities re-
mains controversial, largely due to difficulties in
accounting for the allochthonous organic carbon
inputs required to fuel the excess respiration
(Williams and Bower 1999) as well as difficulties in
reconciling a net heterotrophic upper ocean with
the measured sinking flux of organic carbon.
CO
2
EXCHANGE IN AQUATIC ECOSYSTEMS
Unbalanced aquatic metabolic processes will gen-
erate gaseous disequilibria with respect to the
atmosphere, which can, therefore, provide indica-
tions of the prevalence of auto- or heterotrophy.
Net heterotrophic aquatic communities are net
producers of CO
2
to the atmosphere whereas net
autotrophic communities act as net CO
2
sinks. Yet,
direct demonstrations of links between net eco-
system production and CO
2
exchange are still few.
Indeed, this expected link is confounded by ther-
modynamic and hydrodynamic effects on gas ex-
change and the anthropogenic perturbation of the
CO
2
budget.
The combustion of fossil fuel in the industrial era
has greatly raised atmospheric pCO
2
(IPPC 2001),
forcing the ocean into a net sink for CO
2
by phys-
ical diffusion (Watson and Orr 2003; Sabine and
others 2004). Yet, global surveys of pCO
2
distribu-
tion in the oceans (Takahashi and others 2002)
indicate that about 66 % of the ocean surface
periodically emits CO
2
to the atmosphere, and
about 40 % of the ocean surface outgases CO
2
on
an annual basis (Figure 2). It is intriguing that,
despite the elevated pCO
2
in the atmosphere, some
large areas of the ocean are both net heterotrophic
and supersaturated in CO
2
, a situation primarily
encountered in the subtropical ocean (Takahashi
and others 2002), where planktonic communities
are often heterotrophic (Duarte and Agust 1998;
Duarte and others 2001; Agust and others 2001;
Gonza lez and others 2001; Harrison and others
2001; Serret and others 2001; Arstegui and Harri-
son 2002; Hoppe and others 2002; Chen and others
2003). The biological organic C balance of the
ocean, both pre- and post-anthropogenic pertur-
bation, is still unclear (del Giorgio and Duarte
2002), although most calculations suggest that the
ocean, taken as a whole, is slightly net hetereo-
trophic (Smith and MaKenzie 1987; Raven and
Falkowski 1999; Duarte and others 1999; del
Giorgio and Duarte 2002; Ducklow and McCallister
in press). Indeed, the ocean is believed to have
Figure 2. Frequency distribution of pCO
2
in various
aquatic ecosystems. Data from Takahashi and others
(2002) for ocean; from Cole and others (1994) for lakes
and from Cole and Caraco (2001) for rivers.
866 C. M. Duarte and Y. T. Prairie
been a weak source of CO
2
before the industrial
perturbation of the CO
2
budget (Watson and Orr
2003). Yet, the separation of physical and biological
influences, and the variability of the airsea pCO
2
gradient in time and space, mean that the ocean as
a whole can simultaneously be net heterotrophic
while acting as a net sink for atmospheric CO
2
(Ducklow and McCallister in press).
Data compilations for streams, lakes, and rivers
around the world show that the vast majority of
freshwater ecosystems investigated (90%,
Figure 2) emit CO
2
to the atmosphere (Jones and
Mulholland 1982; Kling and others 1991; Cole
and others 1994). These effluxes are particular
high in tropical systems; Amazon alone is be-
lieved to outgas about 0.5 Gt C y
)1
(Richey and
others 2002). Dillon and Molot (1997) showed,
applying a DOC and DIC mass balance model to
Canadian lakes, that CO
2
emissions to the atmo-
sphere were largely supported by allochthonous
DOC inputs to the lakes, 30 % of which were
respired within the lakes. Estuaries, which also
receive large terrestrial organic carbon inputs,
have also been reported to be important sources
of CO
2
to the atmosphere (Frankignoulle and
others 1998) at least regionally. These emissions
do not necessarily result from respiration within
aquatic ecosystems, as some of the CO
2
they emit
to the atmosphere may in fact derive from res-
piration on land transferred to aquatic ecosystems
through the loading of dissolved inorganic carbon
from groundwater. Yet, analyses of the carbon
budget of the biosphere largely ignore freshwater
and estuarine ecosystems (for example, IPCC
2001), which are considered to play a negligible
role in CO
2
exchanges, despite current knowledge
suggesting they may be important vectors of
carbon transformation and CO
2
emissions to the
atmosphere. Indeed, freshwater ecosystems are
important vectors of CO
2
emissions to the atmo-
sphere, with an estimated CO
2
efflux of more
than 1 Gt C y
)1
(Table 1). These emissions rep-
resent about half of the currently accepted oce-
anic uptake, bringing the CO
2
exchange of the
aquatic compartment of the biosphere to a mod-
est net uptake of 1.2 Gt C y
)1
(Table 1).
PROSPECT
The available information on the metabolism and
CO
2
fluxes from aquatic ecosystems brings into
question two commonly held assumptions, (1) that
aquatic ecosystems are autotrophic, depending so-
lely on their own production, which is opposed to
accumulating evidence for heterotrophy in aquatic
ecosystems; and (2) that freshwater ecosystems are
negligible components of the C budget of the bio-
sphere, which is inconsistent with the high CO
2
emission rates from freshwater ecosystems to the
atmosphere which are not much lower in magni-
tude than the CO
2
uptake by the ocean.
The prevalence of heterotrophic metabolism and
CO
2
efflux from aquatic ecosystems necessarily
requires allochthonous inputs of organic (Odum
1956; Cole and others 1994; Cole and Caraco
2001), and possibly inorganic (Raymond and Cole
2003) carbon to the aquatic systems, and therefore
emphasizes the importance of transfers between
adjacent systems. The metabolic imbalances in
some regions of the oceans may be accounted for,
by internal redistribution of organic carbon
involving, for example, export from coastal to
open-ocean ecosystems (Duarte and others 1999;
Liu and others 2000, 2002; del Giorgio and Duarte
2002; Ducklow and McCallister in press) and a
modest (0.4 Gt C y
)1
, Ludwig and others 1996)
input from land. In contrast, the CO
2
efflux from
freshwater ecosystems requires land-derived inputs
of organic as well as inorganic carbon from land
(Raymond and Cole 2003). There is therefore
growing evidence of a close link between the ter-
restrial, freshwater, and marine components of the
biospheres carbon budget (Houghton 2003), as the
important role of allochthonous, land-derived in-
puts in supporting aquatic metabolism is being
recognized both for freshwater (del Giorgio and
Peters 1994; Prairie and others 2002) and marine
(Bauer and Druffel 1998; Duarte and Agusti 1998;
Duarte and others 1999; Liu and others 2000, 2002;
del Giorgio and Duarte 2002) ecosystems. The fact
that terrestrial organic carbon may be exported and
support metabolism in aquatic environments also
suggests that the transference from the terrestrial to
the aquatic domain affects the lability of the
materials exported (for example, Bertilsson and
Tranvik 2000), in addition to enhancing transport.
The understanding of the changes in lability of
terrestrial organic carbon as it enters the aquatic
biosphere is an important challenge for the future.
Table 1. Global CO
2
Emissions ()) or Uptake (+)
by Aquatic Components of the Biosphere
Component
Flux
(Gt C y
)1
) Reference
Ocean 2.2 Takahashi and others (2002)
Rivers )0.3 Cole and Caraco (2001)
Amazon alone )0.5 Richey and others (2002)
Lakes )0.14 Cole and others (1994)
Prevalence of Heterotrophy in Aquatic Ecosystems 867
In addition to water as a vector for these transfer-
ences, the atmosphere may also be an important
vector of organic carbon transference among eco-
systems, as indicated by suggestions that atmo-
spheric carbon inputs to the ocean may be high (del
Giorgio and Duarte 2002). There are multiple
routes and forms in which allochtonous organic
carbon can be transferred from terrestrial to aquatic
ecosystems, supporting net heterotrophy and CO
2
emissions (Cole and others 1994). Despite historical
interest in cross-habitat carbon fluxes, our capacity
to integrate these still requires significant
improvement (Vanni and others 2004). In addition,
the correspondence between aquatic metabolism
and CO
2
emissions or uptake need to be explicitly
tested and not only assumed, as there are processes,
including physical and chemical processes that af-
fect CO
2
emissions and that may confound the
relationship between aquatic metabolism and CO
2
emissions, which may explain differences in the
prevalence of CO
2
emissions and net heterotrophy
in aquatic ecosystems.
In summary, the division between terrestrial,
marine, and freshwater components has been used
as a convenient way to compartmentalize the bio-
sphere, but evidence suggests that the integration
of the processes involved is much greater than
originally envisioned. The major changes in our
views of aquatic ecosystem metabolism promoted
by realization of the prevalence of heterotrophy in
aquatic ecosystems and the resulting appreciation
of the extent of the integration between adjacent
ecosystems, qualies as a potential paradigm shift
in our understanding of aquatic ecosystems.
Hopefully, this reassessment will yield a more
integrative understanding of the carbon ux in the
biosphere as well as a better integration between
aquatic and terrestrial components, which are
currently assessed independently or neglected all
together, as is the case in freshwater ecosystems,
which are still absent from current depictions of the
carbon budget in the biosphere (IPPC 2001). Proper
consideration of all aquatic ecosystems in the global
C budget of the biosphere can only lead to an im-
proved estimation of the uxes involved and in
some cases, may even alter our perception of the
very direction of these uxes.
ACKNOWLEDGEMENTS
This research is a contribution to projects EURO-
TROPH, funded by the European Commission
(project EUROTROPH, ref. REN2001-4977-E),
COCA (REN-2000-1471-C02), funded by the
Spanish I+D program, and the Integrating the
aquatic with the terrestrial component of the global
carbon budget Working Group, supported by the
National Center for Ecological Analysis and Syn-
thesis, a Center supported by NSF (Grant #DEB-94-
21535), the University of California at Santa Bar-
bara, and the State of California. Y.P. and C.M.D.
were also supported by sabbatical fellowships from
the Spanish Ministry of Education and Science. We
thank J.J. Cole and three anonymous reviewers for
helpful suggestions.
REFERENCES
Agust S, Duarte CM, Vaque D, Hein M, Gasol JM, Vidal M.
2001. Food web structure and elemental (C, N, and P) fluxes
in the Eastern tropical North Atlantic. Deep-Sea Research II
48:2295321.
Agust S, Satta MP, Mura MP. 2004. Summer community res-
piration and pelagic metabolism in upper surface Antarctic
waters. Aquat Microb Ecol 35:197205.
Aitkenhead JA, McDowell WH. 2000. Soil C:N ratio as a pre-
dictor of annual riverine DOC flux at local and global scales.
Global Biogeochem. Cycles 14:12738.
Ar stegui J, Harrison WG. 2002. Decoupling of primary pro-
duction and community respiration in the ocean: implica-
tions for regional carbon studies. Aquat Microb Ecol 29:
199209.
Bauer JE, Druffel ERM. 1998. Ocean margins as a significant
source of organic matter to the deep open ocean. Nature
392:4825.
Bertilsson S, Tranvik LJ. 2000. Photochemical transformation of
dissolved organic matter in lakes. Limnol Oceanogr 45:75362.
Caffrey JM. 2004. Factors controlling net ecosystem metabolism
in U.S. Estuaries. Estuaries 27:90101.
Caraco NF, Cole JJ. 2004. When terrestrial organic matter is sent
down the river: Importance of allochthonous C inputs to the
metabolism in lakes and rivers. In: Polis A, Power ME, Huxel
GR, Eds. Food webs at the landscape level. Press: University of
Chicago press. p 30116.
Carignan R, Planas D, Vis C. 2000. Planktonic production and
respiration in oligotrophic Shield lakes. Limnol Oceanogr
45:18999.
Chen C-C, Shiah F-K, Gong GC, Chiang K-P. 2003. Planktonic
community respiration in the East China Sea: importance of
microbial consumption of organic carbon. Deep-Sea Res. II
50:131125.
Cole JJ, Caraco NF, Kilng GW, Kratz TK. 1994. Carbon dioxide
supersaturation in the surface waters of lakes. Science
265:156870.
Cole JJ, Caraco NF. 2001. Carbon in catchments: connecting
terrestrial carbon losses with aquatic metabolism. Mar Fresh-
water Res 52:10110.
Cole JJ, Pace ML, Carpenter SR, Kitchell JF. 2000. Persistence of
net heterotrophy in lakes during nutrient addition and food
web manipulations. Limnol Oceanogr 45:171830.
Cole JJ. 1999. Aquatic microbiology for ecosystem scientists.
new and recycled paradigms in ecological microbiology. Eco-
systems 2:21525.
del Giorgio PA, Cole JJ, Cimbleris A. 1997. Respiration rates in
bacteria exceed plankton production in unproductive aquatic
systems. Nature 385:14851.
868 C. M. Duarte and Y. T. Prairie
del Giorgio PA, Duarte CM. 2002. Respiration in the open ocean.
Nature 420:37984.
del Giorgio PA, Peters RH. 1994. Patterns in planktonic P:R ratios
in lakes: Influence of lake trophy and dissolved organic car-
bon. Limnol Oceanogr 39:77287.
del Giorgio PA, Williams PjleB. The global significance of respi-
ration in aquatic ecosystems: from single cells to the bio-
sphere. In: del Giorgio PA, Williams PJleB, Eds. Respiration in
aquatic ecosystems, Oxford University Press. [in press-a].
del Giorgio PA, Williams PjleB. Respiration in aquatic ecosys-
tems, Oxford University Press [in press-b].
Dillon PJ, Molot LA. 1997. Dissolved organic and inorganic
carbon mass balances in central Ontario lakes. Biogeochem
36:2942.
Duarte CM, Agust S, del Giorgio PA, Cole JJ. 1999. Regional
carbon imbalances in the oceans. Science 284:1735b.
Duarte CM, Agust S. 1998. The CO
2
balance of unproductive
aquatic ecosystems. Science 281:2346.
Duarte CM, Ar stegui J, Gonza lez N, Agust S, Anado n R. 2001.
Evidence for a heterotrophic subtropical NE Atlantic. Limnol
Oceanogr 46:4258.
Duarte CM, Cebria n J. 1996. The fate of marine autotrophic
production. Limnol Oceanogr 41:175866.
Duarte CM, Agust S, Vaque D. 2004. Controls on planktonic
metabolism in the Bay of Blanes, north-western Mediterra-
nean litoral. Limnol Oceanogr [in press].
Ducklow HW, McCallister SL. The biogeochemistry of carbon
dioxide in the coastal oceans, In: Robinson AR, Brink K, Eds.
The sea, vol 13, The global coastal ocean-multi-scale inter-
disciplinary processes. Harvard Univ. Press [in press].
Field CB, Behrenfeld MJ, Randerson JT, Falkowski P. 1998.
Primary production of the biosphere: integrating terrestrial
and oceanic components. Science 281:23740.
Frankignoulle M, Abril G, Borges A, Bourge I, Canon C, Delille
B, Libert E, The ate JM. 1998. Carbon dioxide emisio n from
European estuaries. Science 282:4348.
Gattuso J-P, Franjignoulle M, Wollast R. 1998. Carbon and
carbonate metabolism in coastal aquatic ecosystems. Annu
Rev Ecol Syst 29:40533.
Gonza lez N, Anado n R, Mourin o B, Ferna ndez E, Sinha B,
Esca nez J, de Armas D. 2001. The metabolic balance of the
planktonic community at the N. Atlantic Subtropiucal Gyre:
the role of mesoscale instabilities. Limnol Oceanogr 46:946
52.
Hansell DA, Ducklow HW, Macdonald AM, ONeil Baringer M.
2004. Metabolic poise in the North Atlantic Ocean diagnosed
from organic matter transports. Limnol Oceanogr 49:108494.
Hanson PC, Bade DL, Carpenter SR, Kratz TK. 2003. Lake
metabolism: relationships with dissolved organic carbon and
phosphorus. Limnol Oceanogr 48:11129.
Harrison WG, Ar stegui J, Head EJH, Li WKW, Longhurst AR,
Sameoto DD. 2001. Basin-scale variability in plankton bio-
mass and community metabolism in the subtropical North
Atlantic Ocean. Deep-Sea Res II 48:224169.
Hemminga MA, Duartey CM. 2000. Seagrass ecology. Cam-
bridge Univ. Press, Cambridge.
Hope D, Kratz TK, Riera JL. 1994. Relationship between pCO
2
and dissolved organic carbon in northern Wisconsin lakes. J
Environ Qual 25:14425.
Hopkinson CS Jr., Smith EM. Estuarine respiration: An overview
of benthic, pelagic, and whole system respiration. In: del
Giorgio PA, Williams PJleB, Eds. Respiration in aquatic eco-
system. Oxford Univ. Press, NY, USA. [in press].
Hoppe H-G, Gocke K, Koppe R, Begler C. 2002. Bacterial growth
and primary production along a north-south transect in the
Atlantic Ocean. Nature 416:16871.
Houghton RA. 2003. Why are estimates of the terrestrial carbon
balance so different? Global Change Biol. 9:5009.
IPCC 2001. The carbon cycle and atmospheric carbon dioxide.
In: IPCC. Climate Change 2001. Cambridge: Cambridge Univ.
In Press. p 183237.
Karl DM, Laws EA, Morris P, Williams PleBJ, Emerson S. 2003.
Metabolic balance of the open sea. Nature 426:32.
Kling GW, Kipphut GW, Miller MC. 1991. Arctic lakes and
streams as gas conduits to the atmosphere implications for
tundra carbon budgets. Science 251:298301.
Knauer GA. 1993. In: Wollast R Eds. Interactions of C, N, P, and
S biogeochemical cycles and global change. Berlin: Springer-
Verlag, p 21131.
Liu JK, Atkinson L, Chen CTA, Gao S, Hall J, Macdonald RW,
Talaue-McManus L, Quin ones R. 2000. Exploring continental
margin carbon fluxes on a global scale. EOS 81:6414.
Liu J-K, Iseki K, Chao S-Y. 2002. Continental margin carbon
fluxes. In: Hanson RB, Ducklow HW, Field JG, Eds. The
changing ocean carbon cycle. London: Cambridge Univ. Press.
p 186239.
Lucea A, Duarte CM, Agust S, Kennedy H. Nutrient dynamics
and ecosystem metabolism in the Bay of Blanes (NW Medi-
terranean). Biogeochemistry. [in press].
Ludwig W, Probst JL, Kempe S. 1996. Predicting the oceanic
input of organic carbon by continental erosion. Global Bio-
geochemical Cycles 10:2341.
Mulholland PJ. 2003. Large-scale patterns in dissolved organic
carbon concentration, flux, and sources. In: Stuart EG, Find-
lay G, Sinsabaugh RL, Eds. Aquatic ecosystems: interactivity
of dissolved organic matter. Amsterdam: Elsevier Science. p
512.
Odum HT. 1956. Primary production in flowing waters. Limnol
Oceanogr 1:1127.
Pace ML, Cole JJ, Carpenter SR, Hodgon JR. 2004. Whole-lake
carbon-13 additions reveal terrestrial support of aquatic food
webs. Nature 423:2403.
Pace ML, Prairie YT. Respiration in lakes. In: del Giorgio PA,
Williams PJLeB, Eds. Respiration in aquatic systems Oxford
University Press, NY, USA. [in press].
Prairie Bird DF, Cole JJ. 2002. The summer metabolic balance in
the epilimnion of southeastern Quebec lakes.. Limnol Ocea-
nogr 47:31621.
Raymond PA, Bauer JE, Cole JJ. 2000. Atmospheric CO
2
eva-
sion, dissolved inorganic carbon production, and net hetero-
trophy in the york river estuary. Limnol Oceanogr 45:1707
17.
Raymond PA, Cole JJ. 2003. Increase in the export of alkalinity
from North Americas largest river. Science 301:8891.
Richey JE, Melack JM, Aufdenkampe AK, Ballester VM, Hess
LL. 2002. Outgassing from Amazonian rivers and wetlands
as a large tropical source of atmospheric CO
2
. Nature 416:
61720.
Robinson C, Williams, PJleB. Respiration and its measurement
in surface waters. In: del Giorgio PA, Williams PJleB, Eds.
Respiration in aquatic ecosystem. Oxford Univ. Press, NY,
USA. [in press].
Prevalence of Heterotrophy in Aquatic Ecosystems 869
Sabine CL, Feely RA, Gruber N, Key RM, Lee K, Bullister JL,
Wanninkhof R, Wong CS, Wallace DWR, Tilbrook B, Millero
FJ, Peng T-H, Kozyr A, Ono T, Rios AF. 2004. The oceanic sink
for Anthropogenic CO
2
. Science 305:36771.
Schelsinger WJ, Melack JM. 1981. Transport of organic carbon
in the worlds rivers. Tellus 33:17287.
Schlesinger WH. 1991. Biogeochemistry. An analysis of global
change: Academic Press, New york.
Scully NM, Cooper WJ, Tranvik LJ. 2003. Photochemical effects
on microbial activity in natural waters: the interaction of
reactive oxygen species and dissolved organic matter. Fems
Microb Ecol 46:3537.
Serret P, Robinson C, Ferna ndez E, Teira E, Tilstone G. 2001.
Latitudinal variation of the balance between plankton pho-
tosynthesis and respiration in the eastern Atlantic Ocean.
Limnol Oceanogr 46:164252.
Smith S, MacKenzie F. 1987. The ocean as a heterotrophic sys-
tem: implications for the global carbon cycle. Global Biogeo-
chem Cycles 1:18798.
Smith SV, Hollibaugh JT. 1993. Coastal metabolism and the
oceanic organic carbon balance. Rev Geophysics 3:7589.
Takahashi T, Sutherland SC, Sweeney C, Poisson A, Metzl N,
Tilbrook B, Bates N, Wanninkhof R, Feely RA, Sabine C,
Olaffson J, Nojiri Y. 2002. Global seaair CO
2
flux based on
climatological surface ocean pCO
2
, and seasonal biological and
temperature effects. Deep-Sea Res. II 49:160122.
Vanni MJ, DeAngelis DL, Schindler DE, Huxel GR. 2004.
Overview: cross-habitat flux of nutrients and detritus. In:
Polis A, Power ME, Huxel GR, Eds. Food webs at the
landscape level. Chicago, USA: University of Chicago Press.
p 311.
Watson AJ, Orr JC. 2003. Carbon dioxide fluxes in the global
ocean. In: Fasham MJR, Eds. Ocean biogeochemistry. Berlin:
Springer-Verlag. p 12343.
Webster JR, Meyer JL. 1999. Organic matter budgets for streams:
a synthesis. J North Amer Benthol Soc 16:14161.
Williams PjleB. 2000. Net production, gross production and
respiration: what are the interconnections and what controls
what? In: Hanson RB, Ducklow HW, Field JG, Eds. The
changing ocean carbon cycle. London: Cambridge Univ Press.
p 3760.
Williams PJLeB 1998. The balance of plankton respiration and
photosynthesis in the open ocean. Nature 394:557.
Williams PleBJ, Bower DG. 1999. Determination of organic
carbon balance in the oceans from field observations - A re-
evaluation. Science 284:1735a.
870 C. M. Duarte and Y. T. Prairie