HAL Id: cea-00819125
https://hal-cea.archives-ouvertes.fr/cea-00819125
Submitted on 16 Oct 2019
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Terrestrial gross carbon dioxide uptake: Global
distribution and covariation with climate
Christian Beer, Markus Reichstein, Enrico Tomelleri, Philippe Ciais, Martin
Jung, Nuno Carvalhais, Christian Rödenbeck, M. Altaf Arain, Dennis
Baldocchi, Gordon B. Bonan, et al.
To cite this version:
Christian Beer, Markus Reichstein, Enrico Tomelleri, Philippe Ciais, Martin Jung, et al.. Terres-
trial gross carbon dioxide uptake: Global distribution and covariation with climate. Science, Amer-
ican Association for the Advancement of Science (AAAS), 2010, 329 (5993), pp.834. �10.1126/sci-
ence.1184984�. �cea-00819125�
thesametime[fig.S5A(19)] are consistent with
extracellular signals.
After a relatively brief (~40-s) period of extra-
cellular signals, we observed several pronounced
changes in recorded signals (Fig. 4, B and C, II
and III) without application of external force to
the PDMS/cell support. Specifically, the initial
extracellular signals gradually dis ap pe ar ed (Fig .
4, B and C, II, pink stars). There was a con-
comitant decrease in baseline potential, and new
peaks emerged that had an opposit e sig n , s i m i l ar
frequency , much greater amplitude, and longer
duration (Fig. 4B, II, green stars). These new
peaks, which are coincident with cardiomyocyte
cell beating, rapidly reached a steady state (Fig.
4B, III) with an average calibrated peak ampli-
tude of ~80 mV and duration of ~200 ms. The
amplitude, sign, and duration are near those re-
ported for whole-cell patch clamp recordings
from cardiomyocyte s (27, 28); thus, we conclude
that these data represent a transition to steady-
state intracellular recording (Fig. 4A, right) with
the 3D nanowire probe.
Detailed analysis of the latter steady-state
peaks (Fig. 4C, III) shows five characteristic phases
of a cardiac intracellular potential (27, 28),
including (a) resting state, (b) rapid depolarization,
(c) plateau, (d) rapid repolarization, and (e) hy-
perpolarization. In addition, a sharp transient peak
(blue star) and the notch (orange star) possibly
associated with the inward sodium and outward
potassium currents (28) can be resolved. Optical
images recorded at the same time as these
intracellular peaks (fig. S5B) showed the kinked
nanowire probe tips in a possible intracellular
region of the cell (19). When the PDMS/cell
substrate was mechanically retracted from the 3D
kinked nanowire devices, the intracellular peaks
disappeared, but they reappeared when the cell
substrate was brought back into gentle contact
with the device. This process could be repeated
multiple times without degradation in the rec-
orded signal. When vertical 3D nanoprobe
devices were bent into a configuration with angle
q < ~50° with respect to the substrate, or when
kinked nanowire devices were fabricated on
planar substrates, we could record only extra-
cellular signals. These results confirm that elec-
trical recording arises from the highly localized,
pointlike nanoFET near the probe tip, which (i)
initially records only extracellular potential, (ii)
simultaneously records both extracellular and
intracellular sign als as the nanoFET spans the
cell membrane, and (iii) records only intracellular
signals when fully inside the cell.
Additional work remains to develop this new
synthetic nanoprobe as a routine tool like the
patch-clamp micropipette (10, 11), although we
believe that there are already clear advantages:
Electrical recording with kinked nanowire
probes is relatively simple without the need for
resistance or capacitance compensation (9, 11);
the nanoprobes are chemically less invasive than
pipettes, as there is no solution exchange; the
small size and biomimetic coating minimizes me-
chanical invasiveness; and the nanoFET s have high
spatial and temporal resolution for recording.
References and Notes
1. D. A. Giljohann, C. A. Mirkin, Nature 462, 461 (2009).
2. T. Cohen-Karni, B. P. Timko, L. E. Weiss, C. M. Lieber,
Proc. Natl. Acad. Sci. U.S.A. 106, 7309 (2009).
3. J. F. Eschermann et al., Appl. Phys. Lett. 95, 083703
(2009).
4. Q. Qing et al., Proc. Natl. Acad. Sci. U.S.A. 107, 1882
(2010).
5. I. Heller, W. T. T. Smaal, S. G. Lemay, C. Dekker, Small 5,
2528 (2009).
6. W. Lu, C. M. Lieber, Nat. Mater. 6, 841 (2007).
7. A. Grinvald, R. Hildesheim, Nat. Rev. Neurosc i. 5, 874
(2004).
8. M. Scanziani, M. Häusser, Nature 461, 930 (2009).
9. R. D. Purves, Microelectrode Methods for Intracellular
Recording and Ionophoresis (Academic Press, London,
1981).
10. B. Sakmann, E. Neher , Annu. Rev. Physiol. 46, 455 (1984).
11. A. Molleman, Patch Clamping: An Introductory Guide to
Patch Clamp Electrophysiology (Wiley, Chichester, UK,
2003).
12. R. M. Wightman, Science 311, 1570 (2006).
13. A. G. Ewing, T. G. Strein, Y. Y. Lau, Acc. Chem. Res. 25,
440 (1992).
14. M. G. Schrlau, N. J. Dun, H. H. Bau, ACS Nano 3, 563
(2009).
15. J. P. Donoghue, Nat. Neurosci. 5 (suppl.), 1085
(2002).
16. M. Ieong, B. Doris, J. Kedzierski, K. Rim, M. Yang,
Science 306, 2057 (2004).
17. M. Ferrari, Nat. Rev. Cancer 5, 161 (2005).
18. B. Z. Tian, P. Xie, T. J. Kempa, D. C. Bell, C. M. Lieber,
Nat. Nanotechnol. 4, 824 (2009).
19. Materials and methods are available as supporting
material on Science Online.
20. C. Conde, A. Cáceres, Nat. Rev. Neurosci. 10, 319 (2009).
21. T. G. Leong et al., Proc. Natl. Acad. Sci. U.S.A. 106, 703 (2009).
22. N. Misra et al., Proc. Natl. Acad. Sci. U.S.A. 106, 13780
(2009).
23. X. J. Zhou, J. M. Moran-Mirabal, H. G. Craighead,
P. L. McEuen, Nat. Nanotechnol. 2, 185 (2007).
24. L. V. Chernomordik, M. M. Kozlov, Nat. Str uct. Mol. Biol.
15, 675 (2008).
25. W. C. Claycomb et al., Proc. Natl. Acad. Sci. U.S.A. 95,
2979 (1998).
26. B. D. Almquist, N. A. Melosh, Proc. Natl. Acad. Sci. U.S.A.
107, 5815 (2010).
27. D. M. Bers, Nature 415, 198 (2002).
28. D. P. Zipes, J. Jalife, Cardiac Electrophysiology: From Cell
to Bedside (Saunders, Philadelphia, ed. 2, 2009).
29. WethankG.Yellen,W.C.Claycomb,B.P.Bean,
P. T. Ellinor, G. H. Yu, D. Casanova, B. P. Timko, and
T. Dvir for help with experiments and data analysis.
C.M.L. acknowledges support from a NIH Director’s Pioneer
Award (5DP1OD003900), a National Security Science and
Engineering Faculty Fellow (NSSEFF) award (N00244-09-1-
0078), and the McKnight Foun dation Neur oscience award.
Supporting Online Material
www.sciencemag.org/cgi/content/full/329/5993/830/DC1
Materials and Methods
Figs. S1 to S5
References
10 May 2010; accepted 7 July 2010
10.1126/science.1192033
Terrestrial Gross Carbon Dioxide
Uptake: Global Distribution and
Covariation with Climate
Christian Beer,
1
* Markus Reichstein,
1
Enrico Tomelleri,
1
Philippe Ciais,
2
Martin Jung,
1
Nuno Carvalhais,
1,3
Christian Rödenbeck,
4
M. Altaf Arain,
5
Dennis Baldocchi,
6
Gordon B. Bonan,
7
Alberte Bondeau,
8
Alessandro Cescatti,
9
Gitta Lasslop,
1
Anders Lindroth,
10
Mark Lomas,
11
Sebastiaan Luyssaert,
12
Hank Margolis,
13
Keith W. Oleson,
7
Olivier Roupsard,
14,15
Elmar Veenendaal,
16
Nicolas Viovy,
2
Christopher Williams,
17
F. Ian Woodward,
11
Dario Papale
18
Terrestrial gross primary production (GPP) is the largest global CO
2
flux driving several ecosystem
functions. We provide an observation-based estimate of this flux at 123 T 8 petagrams of carbon per
year (Pg C year
−1
) using eddy covariance flux data and various diagnostic models. Tropical forests and
savannahs account for 60%. GPP over 40% of the vegetated land is associated with precipitation.
State-of-the-art process-oriented biosphere models used for climate predictions exhibit a large
between-model variation of GPP’s latitudinal patterns and show higher spatial correlations between
GPP and precipitation, suggesting the existence of missing processes or feedback mechanisms which
attenuate the vegetation response to climate. Our estimates of spatially distributed GPP and its
covariation with climate can help improve coupled climate–carbon cycle process models.
T
errestrial plants fix carbon dioxide (CO
2
)
as organic comp ounds through photo-
synthesis, a carbon (C) flux also known
at the ecosystem level as gross primary produc-
tion (GPP). T errestrial GPP is the largest global
carbon flux, and it drives several ecosystem func-
tions, such as respiration and growth. GPP thus
contributes to human welfare because it is the
basis for food, fiber, and wood production. In
addition, GPP, along with respiration, is one of
the major processes controlling land-atmosphere
CO
2
exchange, providing the capacity of terres-
trial ecosystems to partly offset anthropogenic
CO
2
emissions.
Although photosynthesis at the leaf and can-
opy level are quite well understood, only tentative
observation-based estimates of global terrestrial
GPP have been possible so far . Plant- and stand-
level GPP has previously been calculated as two
times biomass production (1, 2), with substantial
13 AUGUST 2010 VOL 329 SCIENCE www.sciencemag.org
834
REPORTS
on October 16, 2019 http://science.sciencemag.org/Downloaded from
variation between biomes and sites (3–5). In the
absence of direct observations, a combined GPP
of all terrestrial ecosystems of 120 Pg C year
−1
was obtained (6) by doubling global biomass pro-
duction estima tes (7) without an emp irical basis of
spatially resolved biomass production and its rela-
tionship to GPP. A global terrestrial GPP of 100 to
150PgCyear
−1
is consistent with the observed
variation of
18
OCO in the atmosphere (8, 9). How-
ever, the ability of
18
OCO to constrain GPP de-
pends critically on the isotopic imbalance between
GPP and respiration, and large uncertainties re-
main asso ciated with isotope fractionation pro-
cesses (10). The coupled uptake of carbonyl sulfide
and CO
2
by plants (11, 12) could potentially be
used to further constrain terrestrial GPP by the
combination of atmospheric [COS] measurements
with an inversion of the atmospheric transport (13)
once the ratio of CO
2
versus COS uptake, the ad-
ditional COS deposition to soils, and the COS
efflux from oceans is more precisely quantified.
As an alternative to directly constraining at-
mospheric data to estimate GPP, local informa-
tion can be built into a process-oriented biosphere
model, which is then applied globally. Knowl-
edge of radiative transfer within vegetation can-
opies and of leaf photosynthesis has been used to
represent GPP within process-oriented biosphere
models, which explicitly simulate the behavior of
the ecosystem as an interaction of th e system com-
ponents (e.g., leafs, roots, and soil) in a reductionist
or mechanistic way . If these models are designed
to also simulate a changing state of the biosphere
(e.g., leaf area index and carbon pools), predictio ns
of ecosystem dynamics under changing environ-
mental conditions can be attempted (14). However ,
these process-oriented models are complex com-
binations of scientific hypotheses; hence, their re-
sults depen d on these embedded hypotheses. A
complementary approach is data-oriented or diag-
nostic modeling where general relationships be-
tween existing data sets are first inferred at site-level
and then applied globally by using global grids
of explanatory variables. Particularly when data-
adaptive machine learning approaches are em-
ployed (e.g., artificial neural networks), results
are much less contingent on theoretical assump-
tions and can be considered as data benchmarks
for process models. However, being essentially a
statistical approach, the diagnostic models do lack
the capacity of extrapolating to completely differ-
ent conditions and hinge on the availa bility of suf-
ficient data. With the advent of a global network of
ecosystem-level observations of CO
2
biosphere-
atmosphere exchange (15) (www .fluxdata.org) and
the development of new diagnostic modeling
approaches, a data-oriented global estimation of
GPP has become feasible. In this study , we estimate
terrestrial GPP and its spatial details by diagnostic
models and compare spatial correlations with climate
variables to results from process-oriented models.
The diagnostic modeling comprises two steps,
the parametrization of GPP in relation to explan-
atory variables at sites and the application of the
model by using gridded information about these
explanatory variables. For the first step, GPP was
estimated by partitioning continuous measurements
of net ecosystem exchange (NEE) into GPP and
ecosystem respiration at flux tower sites (16). T wo
flux partitioning methods were consid ered using
night-time or day-time NEE (16). Such site-level
GPP data was then used to calibrate five highly
diverse diagnostic models, which relate GPP to
meteorology , vegetation type, or remote sensing
indices at daily , monthly, or annual time scales
(16). T wo of these approaches are machine learn-
ing techniques: a model tree ensemble (MTE) (17)
and an artificial neural network (ANN) (18). The
Köppen-Geiger cross Biome (KGB) approach is
a look-up table of mean GPP per ecoregion. GPP
of whole river catchment areas is estimated by the
water use efficiency approach (WUE) (19, 20),
1
Biogeochemical Model-Data Integration Group, Max Planck
Institute for Biogeochemistry, 07745 Jena, Germany.
2
Labo-
ratoire des Sciences du Climat et de L’Environnement, Institut
Pier r e Simo n Laplace, CEA-CNRS-UVSQ, Gif-sur-Yvette , France.
3
Faculdade de Ciências e Tecnologia (FCT), Universidade Nova
de Lisboa, Caparica, Portugal.
4
Biogeochemical Systems, Max
Planck Institute for Biogeochemistry, 07745 Jena, Germany.
5
McMaster Centre for Climate Change, McMaster University,
Hamilton, Ontario, Canada.
6
Department of Environmental Science,
Policy and Management and Berkeley Atmospheric Science Center,
University of California, Berkeley, CA 94720, USA.
7
National Center
forAtmosphericResearch,Boulder,CO80305,USA.
8
Potsdam
Institute for Climate Impact Research (PIK), 14473 Potsdam,
Germany.
9
Climate Change Unit, Institute for Environment and
Sustainability, European Commission, DG Joint Research Centre,
Ispra, Italy.
10
Department of Earth and Ecosystem Science, Lund
University, Sweden.
11
Department of Animal and Pl ant Sciences,
University of Sheffield, Sheffield S10 2T N, UK.
12
Departement
Biologie, Universiteit Antwerpen, Belgium.
13
Centre d’étu d e de
la forêt, Faculté de foresterie, de géographie et de géomatique,
Université Laval, Quebec, Canada.
14
Cirad-Persyst, UPR80,
Fonctionnement et Pilotage des Ecos ystémes de Plantation,
Montpellier, France.
15
CATIE (Centro Agronómico Tropical de
Investigación y Enseñ anza), Turrialba, Costa Rica.
16
Nature
Conservation and Plant Ecology Group, Wageningen University,
Netherland s.
17
Graduate School of Geography, Clark University,
Worcester, MA 01610, USA.
18
Department of Forest Environ-
ment and Resources, University of Tu scia, Viterbo, Italy.
*To whom correspondence should be addressed. E-mail:
christian.beer@bgc-jena.mpg.de
Fig. 1. (A) Distributions
of global GPP (Pg C year
−1
)
for the five data-driven ap-
proaches that are most
constrained by data, their
combined global GPP dis-
tribution, and the GPP
distribution by the Mi am i
model. Shown are the me-
dian (red horizontal lines),
the quartiles (blue boxes),
and the 2.5 and 97.5 per-
centiles (vertical black lines),
indicating the 95% con-
fidence interval. MTE is
0
500
1000
1500
2000
2500
3000
3500
90
100
110
120
130
140
150
160
MTE1
MTE2
ANN
WUE
LUE
KGB
All
MIAMI
A
B
C
−50 −25 0 25 50 75
0
1000
2000
3000
4000
5000
Latitude [°]
GPP [gC/m
2
/a]
Data−driven:
Median
Process Models:
Median
Individual Process
Models
Atmospheric
Inversion
either driven by fAPAR only (MTE1) or by both fAPAR and climate data (MTE2) (16). (B)
Spatial details of the median annual GPP (gC/m
2
/a) from the spatially explicit approaches
MTE1, MTE2, ANN, LUE, and KGB. (C) Latitudinal pattern (0.5° bands) of annual GPP. The
gray area represents the range of the diagnostic models MTE1, MTE2, ANN, LUE, and KGB.
The red area represents the range of process model results (LPJ-DGVM, LPJmL, ORCHIDEE,
CLM-CN, and SDGVM). The thick lines represent the medians of both ranges. The dashed
blacklineshowstheresultfornorthernextratropical regions from an independent diagnostic
model. In this approach, we combined gridded information about the seasonal NEE am-
plitude based on atmospheric CO
2
data and an inversion of atmospheric CO
2
transport with
empirical relationships between annual GPP and the seasonal amplitude of NEE derived at
flux tower sites.
www.sciencemag.org SCIENCE VOL 329 13 AUGUST 2010
835
REPORTS
on October 16, 2019 http://science.sciencemag.org/Downloaded from
which combines recently derived global WUE
fields with the long-term averaged evapotranspi-
ration at the watershed scale. This is an important
constraint at the global scale, but the spatial res-
olution is too coarse to use the WUE approach for
estimating the spatial distribution of GPP. The
light-use efficiency approach (LUE) (21, 22)was
applied by combining in situ Bayesian calibration
with an uncertainty propagation per vegetation and
climate class. The Miami model (23) simply ex-
ploits the empirically obtained dependence of
photosynthesis on temperature and precipitation.
The second step, the mapping of flux tower GPP
to the land surface, was performed by applying
these diagnostic models to fields of remote sensing
(24–26) and climatic data (27–29), which are now
available with improved accuracy and high spa-
tial resolution. In so doing, we take into account
several sources of uncertainty, including uncer-
tainty from model parametrization and from ex-
planatory variables (16).
By making use of the new data streams and
the ensemble of five diagnostic models, we pre-
sent an observation-based estimate of an average
global terrestrial GPP of 123 Pg C year
−1
during
the period 1998 to 2005 (Fig. 1A). Uncertainties
and preprocessing of tower CO
2
flux measure-
ments, tower representativeness, flux partitioning
into GPP, uncertainties of climate and remote
sensing data sets, and structural uncertainties of
the diagnostic models propagate to a global un-
certainty with a 95% confidence range from 102
to135PgCyear
−1
or a robust estimate of standard
deviation (30) of 8 Pg C year
−1
. Results from the
LUE approach were higher when using National
Centers for Environmental Prediction (NCEP) ra-
diation. However, we do not show NCEP-driven
results because NCEP radiation and precipitation
is known to be biased (31, 32). The Miami model
overestimates GPP compared to other approaches,
particularly in sparsely vegetated areas with strong
seasonality , such as savannahs, shrublands, and tun-
dra (16) (table S5), because it does not account
for the effect of climate-independent changes in
vegetation structure (e.g., degradation) and vege-
tation type on GPP. Indeed, residuals of this model
correlate significantly with mean annual fraction
of absorbed photosynthetically active radiation
(fAPAR) from remote sensing (fig. S14). Hence,
being a classic model, it is shown only for com-
parison, but results from the Miami model were
not taken into account in the following analyses.
T ropical forests assimilate 34% of the global
terrestrial GPP (Table 1) and have the highest
GPP per unit area (table S5). Savannahs account
for 26% of the global GPP and are the second
most important biome in terms of global GPP.
The large area of savannahs (about twice the sur-
face area of tropical forests) explain their high
contribution. Moreover, the results highlight the
importance of taking into account C4 vegetation
in global GPP estimates. Based on the C4
distribution (figs. S6 and S7), more than 20% of
terrestrial GPP is conducted by C4 vegetation.
Given that there were less than 20 site-years of
flux data for C4-dominated ecosystems, our
uncertainty is lar gest for this type of vegetation.
Therefore, an expansion of observational net-
works should focus on tropical C4 ecosystems.
Boreal forests show a clear longitudinal gradient
in GPP in northern Eurasia where GPP in the
boreal zone decreases toward the east, where
Table 1. GPP for biomes of the world as defined by Prentice et al.(6). Combining the biome extent (fig.
S17) with the spatially explicit GPP distribut ions by the approaches MTE1, MTE2, ANN, LUE, WUE, and
KGB led to the respective median GPP per unit area separately for each biome (fig. S4). These medians
were then multiplied by the biome area (6, 7) (fig. S4) to estimate GPP in column 2. The estimated GPP
total of 122 Pg C year
−1
does not equal our overall median of 123 Pg C year
−1
because the biome area
definedbyfig.S17andby(6) differ slightly. The third column shows GPP as estimated by using NPP
numbers from Saugier et al.(7) under the assumption that NPP/GPP = 0.5 (6).
Biome
GPP
(Pg C year
–1
)
GPP = 2 × NPP*
(Pg C year
–1
)
Tropical forests 40.8 43.8
Temperate forests 9.9 16.2
Boreal forests 8.3 5.2
Tropical savannahs and grasslands 31.3 29.8
Temperate grasslands and shrublands 8.5 14
Deserts 6.4 7
Tundra 1.6 1
Croplands 14.8 8.2
Total 121.7 125.2
*Based on integrated numbers for biomes (6, 7)
Fig. 2. Partial correla-
tion in the spatial do-
main between GPP from
Fig. 1B and either (A)
CRU pr ecipitation, (B)
CRU air temperature, or
(C)ECMWFERA-Interim
short-wave radiation af-
ter applying a moving
4.5° by 4.5° spatial win-
dow and subsequent
median filtering. Shown
are significant correla-
Partial correlation median GPP and air temperature
−1
−0.5
0
0.5
1
Partial correlation median GPP and short−wave radiation
−1
−0.5
0
0.5
1
Partial correlation median GPP and precipitation
−1
−0.5
0
0.5
1
A
B
C
tions (P < 0.01) of which the correlation coefficient is higher/lower than T 0.2.
13 AUGUST 2010 VOL 329 SCIENCE www.sciencemag.org
836
REPORTS
on October 16, 2019 http://science.sciencemag.org/Downloaded from
photosynthesis is subject to an increasingly
continental climate (Fig. 1B).
The latitudinal pattern derived by the different
diagnostic models falls into a quite narrow range
(Fig. 1C). In contrast, there is a larger range among
an ensemble of five process-oriented biosphere
models (Fig. 1C); in comparison to our data-oriented
range, some consistently overestimate GPP, and
others underestimate tropical GPP while matching
or slightly overestimating GPP in the temperate
zone (fig. S26). A standard global parametrization
of the process-oriented models has been applied in
this study; it was not optimized against flux tower
GPP because we aimed at evaluating the process-
based GPP fields and their correlations to climatic
variables. For comparison, we show results by an
additional, completely different approach of scaling
GPP from flux tower sites to the regional scale
(fig. S16), where a reationship between the sea-
sonal NEE amplitude and annual GPP is derived
at flux tower sites and applied to the seasonal
NEE amplitude derived through atmospheric in-
version [update of (33)]. This approach leads to
values at the upper end of the range of the diag-
nostic bottom-up approaches in northern extra-
tropical regions but is still at the lower end of the
range estimated by the process-oriented models.
The differences between process-oriented and
data-oriented estimates could lie in human-induced
degradation of GPP by land use (34). However,
other reasons are possible, including insufficient
model parametrizatio n or structural model errors
that lead to an overestimation of GPP.
Partial correlation analyses between GPP and
climatic variables for 4.5° by 4.5° moving win-
dows show that spatial variation of GPP is as-
sociated with precipitation in 50 to 70% of the
area of nontundra herbaceous ecosystems (Fig.
2A and Table 2). Also, 50% of the crop pro-
duction occurs in regions where photosynthesis is
colimited by precipitation, stressing the impor-
tance of water availability for food security. Inter-
estingly, GPP in the same proportion of temperate
forest areas correlates positively with precipitation
(Table 2). In contrast, the spatial GPP variability
in only 30% of tropical and boreal forests seems
to be associated positively with precipitation, but
GPP of more than half of the boreal forests
correlates positively with air temperature (Table
2). Therefore, the GPP of these biomes seems to
be robust against a moderate climate variation in
the order of magnitude of the current spatial var-
iability of climate, given the very low probability
of a decrease in air temperature in the boreal zone.
We find negative correlations of productivity
with incoming short-wave radiation, in particular
in savannahs, the Mediterranean, and Central
Asian grasslands (Fig. 2C and tables S6 to S8).
These negative partial correlations may indicate
an additional indirect effect of radiation or tem-
perature on GPP by the water balance. Both cli-
matic variables are usually associated with higher
evapotranspiration rates, which will yield more
negative water balances with higher temperature
or radiation levels with consequent negative effects
on primary productivity in these water-limited re-
gions. This interpretation is possible notwithstanding
a direct effect of temperature on vegetation by
heat stress as well as increased levels of diff u s e
radiation associated with overall lower levels of
radiation (35).
After four decades of research on the global
magnitude of primary production of terrestrial
vegetation (23, 36), we present an observation-
based estimate of global terrestrial GPP. Although
we arrive at a global GPP of similar magnitude as
these earlier estimates, our results add confidence
and spatial details. The large range of GPP results
by process-oriented biosphere models indicates
the need for further constraining CO
2
uptake pro-
cesses in these models. Furthermore, our spatially
explicit GPP results contribute to a quantification
of the climatic control of GPP. Complementing
theoretical or process-oriented results (37, 38)
about climatic limitations of GPP, our observation-
based results now constitute empirical evidenc e
for a large effect of water availability on primary
production over 40% of the vegetated land (Fig. 3A)
and up to 70% in savannahs, shrublands, grass-
lands, and agricultural areas (Table 2). Our find-
ings imply a high susceptibility of these ecosystems’
productivity to projected changes of precipitation
over the 21st century (39), but a robustness of
tropical and boreal forests. Results of current pro-
cess models show a large range and a tendency to
overestimate precipitation-associated GPP (Fig.
3B). Most likely, the association of GPP and cli-
mate in process-oriented models can be improv ed
by including negative feedback mechanisms (e.g.,
adaptation) that might stabilize the systems. Our
high spatial resolution GPP estimates, their uncer-
tainty , and their relationship to climate drivers
should be useful for evaluating and thus improving
coupled climate–carbon cycle process models.
Fig. 3. Percentage of vegetated land
surface (A) and corresponding GPP (B)
that is controlled by precipitation, de-
pending on the chosen threshold for
the partial correlation coefficients that
signal a control of GPP by a climate fac-
tor. The blue areas represent the range
of data-driven estimates (MTE1, MTE2,
ANN, LUE, and KGB) using different cli-
mate sources [CRU, ECMWF ERA-Interim,
andGPCP(16)]. This is compared to the
range of proces s-ori ented model results
(LPJ-DGVM, LPJmL, ORCHIDEE, CLM-CN,
and SDGVM) in red. Purple shows the
overlapping area. The thick lines repre-
sent the medians of both ranges. For
instance, GPP of about 40% of the veg etated land surface is controlled by water availability by defining a water control of GPP as a partial correlation coefficient
between GPP and precipitation higher than 0.2.
Table 2. Percentage of biome area for which GPP is climatically controlled, indicated by a median partial
correlation coefficient higher than 0.2 (or 0.5 in brackets). Several climate grids (CRU, ECMWF ERA-
Interim, and GPCP precipitation) were used to perform a partial correlation between the median GPP map
(Fig. 1B) and climate variables for 4.5° by 4.5° moving windows (16). Then, the fractional area with
significant (P < 0.01) partial correlation higher than 0.2 (0.5) was calculated.
Biome P* controlled T† controlled R‡ controlled
Tropical forests 29 (12) 39 (26) 4 (1)
Temperate forests 50 (26) 41 (23) 6 (2)
Boreal forests 20 (5) 55 (31) 21 (7)
Tropical savannahs and grasslands 55 (31) 16 (5) 3 (0)
Temperate grasslands and shrublands 69 (41) 37 (18) 6 (1)
Deserts 61 (37) 18 (6) 8 (2)
Tundra 24 (13) 37 (27) 32 (12)
Croplands 51 (25) 28 (13) 5 (1)
*Precipitation †Air temperature ‡Short-wave radiation
www.sciencemag.org SCIENCE VOL 329 13 AUGUST 2010 837
REPORTS
on October 16, 2019 http://science.sciencemag.org/Downloaded from