Can we reconcile atmospheric estimates of the Northern terrestrial carbon sink with land-based accounting?
Philippe Ciais,Josep G Canadell,Sebastiaan Luyssaert,Frédéric Chevallier,Anatoly Shvidenko,Zegbeu Poussi,Matthias Jonas,Philippe Peylin,Anthony Wayne King,Ernest-Detlef Schulze,Shilong Piao,Christian Rödenbeck,Wouter Peters,François-Marie Bréon
DOI: https://doi.org/10.1016/j.cosust.2010.06.008
IF: 7.964
2010-01-01
Current Opinion in Environmental Sustainability
Abstract:We estimate the northern hemisphere (NH) terrestrial carbon sink by comparing four recent atmospheric inversions with land-based C accounting data for six large northern regions. The mean NH terrestrial CO 2 sink from the inversion models is 1.7 Pg C year −1 over the period 2000–2004. The uncertainty of this estimate is based on the typical individual (1-sigma) precision of one inversion (0.9 Pg C year −1 ) and is consistent with the min–max range of the four inversion mean estimates (0.8 Pg C year −1 ). Inversions agree within their uncertainty for the distribution of the NH sink of CO 2 in longitude, with Russia being the largest sink. The land-based accounting estimate of NH carbon sink is 1.7 Pg C year −1 for the sum of the six regions studied. The 1-sigma uncertainty of the land-based estimate (0.3 Pg C year −1 ) is smaller than that of atmospheric inversions, but no independent land-based flux estimate is available to derive a ‘between accounting model’ uncertainty. Encouragingly, the top-down atmospheric and the bottom-up land-based methods converge to consistent mean estimates within their respective errors, increasing the confidence in the overall budget. These results also confirm the continued critical role of NH terrestrial ecosystems in slowing down the atmospheric accumulation of anthropogenic CO 2 . Introduction Quantifying the fate of fossil fuel carbon emissions depends critically on accurate diagnosis of spatial and temporal distribution of natural carbon (C) fluxes over land and ocean. Studies using global atmospheric transport models to infer surface fluxes from surface carbon dioxide (CO 2 ) concentration observations have estimated the northern mid-latitudes and high latitudes to be a net sink of approximately 2.0–3.5 Pg C year −1 [ 1 ]. Analyses of surface ocean partial pressure of CO 2 [ 2–4 ], atmospheric carbon isotopes [ 5,6 ], and atmospheric oxygen measurements [ 7 • ] further support that most of this sink must reside in terrestrial ecosystems. However, atmospheric inversion estimates are uncertain because of sparse atmospheric stations and transport model errors [ 8 •• ]. In parallel with atmospheric inversions, data gathering over land enables estimation of the distribution of surface fluxes [ 9–12 ]. These data include firstly, land flux measurements [ 13 ] at eddy-flux towers that can be scaled up using models and satellite and climate fields to derive spatially explicit CO 2 flux distributions [ 14,15 ] and secondly, inventories of carbon pools in biomass and soils [ 16–18 ] that can be repeated over time to deduce long-term mean flux estimates from pool changes. Despite numerous local studies, complemented by the synthesis of the C balance of some large regions such as China [ 12 ], North America [ 19 ], Europe [ 11 ], and Russia [ 20,21 ], the bottom-up terrestrial C accounting approach suffers from data gaps, which requires assumptions for upscaling local data. Land-based data also have gaps in time, and do not cover the same time period in each region. Like the atmosphere-based estimates, the land-based approach is subject to big uncertainties at continental scales. Moreover, it is difficult to reconcile inversions of surface-atmosphere CO 2 fluxes with land-based data, because upscaling of ecosystem fluxes must include the less often measured fluxes from lateral transport. Specifically, the ecosystem-atmosphere CO 2 flux is equal to the sum of a carbon stock change and of a carbon flux removed from this ecosystem and displaced by lateral transport. Lateral transport includes carbon embedded in traded products, in rivers (aquatic fluxes), and the atmospheric transport of reduced carbon compound emissions such as CO and biogenic volatile organic compounds that remains undetected by CO 2 inversions [ 22 ]. At continental scales, processing and use of products [ 23 ] and outgassing from lakes and rivers cause CO 2 sources to the atmosphere, that are accounted by inversions. The goal of this study is to establish the NH (Northern Hemisphere) terrestrial CO 2 sink and its uncertainty, based upon atmospheric CO 2 inversion fluxes, which we compare with independent bottom-up C accounting estimates from several large regions. We compile and synthesize four atmospheric inversion CO 2 fluxes and regional bottom-up estimates from recent studies [ 11,12,19,21 ]. Atmospheric CO 2 mass balance A positive pole-to-pole atmospheric CO 2 gradient is expected because fossil fuel burning occurs almost entirely in the NH (8.7 ± 0.5 Pg C year −1 in 2008) and it takes some time to mix emissions homogeneously throughout the atmosphere. Thus, the persistence of fossil fuel emissions induces a permanent accumulation of CO 2 in the NH compared to the southern hemisphere. The existence of a sink of atmospheric CO 2 in the NH has been inferred from the misfit between the observed inter-hemispheric CO 2 concentration gradient and the one obtained from fossil fuel emissions alone [ 24 •• ,25 ]. More precisely, Tans et al. [ 24 •• ] calculated that the CO 2 pole-to-pole gradient expected from fossil fuel emissions alone from 1980 to 1987 was in the range 3.8–5.6 ppm, whereas the observed gradient was only 2 ppm (see their Figure 3). Today, fossil fuel emissions remain concentrated in the NH, but the recent economic growth of China and India is adding emissions toward the Northern Tropics, closer to the Equator. Both Tans et al. [ 24 •• ] and Keeling et al. [ 26 ] deduced that, in order to match the measured CO 2 gradient, a sink was needed north of the equator. They used three-dimensional transport models with various sink scenarios in the ocean and on land. Keeling and colleagues reached the conclusion that the NH sink was dominant in the oceans. Tans et al. constrained a smaller sink in the ocean than Keeling et al. did from sparse ocean observations of the sea-to-air difference in CO 2 partial pressure (ΔpCO 2 ) and suggested that, by difference, ‘a large terrestrial sink at northern temperate latitudes is necessary’. This finding came out as a surprise to the research community, because the terrestrial sink inferred by Tans et al. was huge ‘the total terrestrial sink at high northern and temperate latitude varies between 2.0 and 2.7 Pg C yr −1 ’. There was virtually no direct bottom-up evidence at that time for such a strong uptake of carbon by northern ecosystems. Measurements of atmospheric carbon isotopes that fractionate strongly when CO 2 is absorbed on land and negligibly when CO 2 is dissolved by the ocean, have independently confirmed the existence of an NH terrestrial sink [ 6,27 ]. So did measurements of atmospheric oxygen [ 7 • ]. In the early 1990s, uncertainties were larger for bottom-up estimates than for atmospheric estimates. Virtually no or very sparse measurements of land fluxes (and of ocean fluxes) were available to confirm the results deduced from the atmospheric CO 2 mass balance. Relying on more atmospheric stations deployed over the NH during the late 1990s, and on a set of nine transport models prescribed with the same fluxes, the TRANSCOM-1 atmospheric transport models’ intercomparison study [ 28,29 ] showed that there was a significant spread between models in the CO 2 pole-to-pole gradient (range of 2 ppm) with regard to the mixing of fossil fuel emissions. This spread reflects different mixing of atmospheric transport models, in particular vertical mixing. Thus, a biased transport model will provide a biased diagnostics of the NH sink. In particular, a transport model with fast mixing results in a smaller NH sink than a model with more sluggish mixing. In addition, the co-variation between atmospheric transport and the seasonality of terrestrial CO 2 fluxes can produce a CO 2 latitudinal gradient, even with purely seasonal, annually balanced, land biospheric fluxes. In their model, Tans et al. had a very small gradient generated by purely seasonal land flux (0.05 ppm). But Rayner and Law [ 28 ] concluded from the nine TRANSCOM-1 model results that ‘The annual mean meridional response of the models to seasonal biotic forcing can be classified into two groups. Models which represent turbulent mixing in the planetary boundary layer simulate a pole-to-pole gradient in surface CO 2 that is roughly half as strong as that obtained in the fossil fuel experiment. The other models simulate a very weak meridional structure in these runs’. For instance, Denning et al. [ 30 ] using a model where transport is consistently coupled with biotic fluxes, found a strong covariance between the seasonal cycle of turbulent boundary layer mixing and CO 2 fluxes over northern lands, inducing a positive CO 2 gradient in the NH of 5 ppm at background atmospheric stations with a purely seasonal biotic forcing. Thus, for such a model, a particularly strong terrestrial sink is needed to match atmospheric CO 2 data. Later, the TRANSCOM modelers provided runs for the same inversion procedure to optimize the global CO 2 flux distribution over coarse regions. The results confirmed the existence of a large NH sink and of a significant fraction of that sink in terrestrial ecosystems [ 1,31,32 ]. The mean NH terrestrial sink estimate is 2.5 Pg C year −1 and the range is from 0.5 to 4.5 Pg C year −1 . The uncertainty is due to both sparse atmospheric stations (inversion precision) and to unresolved systematic differences between transport models (transport accuracy). Many other inversions have been published using different settings [ 33–38 ]. Their results are qualitatively similar to those of Gurney et al. Compiling results from these inversions gives a range of most likely values of the terrestrial NH sink values of 0.5–2.8 Pg C year −1 as reported in the IPCC Third Assessment Report [ 39 ] and further assessed in the Fourth Assessment report [ 40 ]. A recent study by Stephens et al. [ 8 •• ] used independent measurements of vertical profiles of CO 2 in the atmosphere to cross validate the TRANSCOM atmospheric transport models for their optimized fluxes. They found that even if prescribed with total fluxes optimized against surface atmospheric stations data, most of the transport models mixed CO 2 vertically not strongly enough in winter, and maybe too strongly in summer. Excluding the majority of biased models that do not match the observed vertical CO 2 gradient lower the NH sink magnitude down to 1.5 with range of 0.5–2.5 Pg C year −1 . Because unrealistic models were removed, this range of NH terrestrial sink is narrower than that of Ref. [ 1 ]. Top-down atmospheric CO 2 inversions Twenty years after the seminal studies of Tans et al. and Keeling et al. , we use the results from four recent inversion systems to analyze the NH land sink and its regional distribution. We choose these four inversions because they were available through the Carboscope website ( http://www.carboscope.eu/accessed November 2009) and because they are state-of-the-art inversions. These four inversions use most of surface station data available in recent versions of transport models forced by reanalyzed winds. They solve for fluxes on the model grid [ 41,42 ] (F Chevallier et al. , CO 2 surface fluxes at grid point scale estimated from a 21-year reanalysis of atmospheric measurements. J Geophys Res, unpublished data) or on a large number of regions [ 43 ], compared to older inversions using only a limited number of coarse regions [ 1 ]. The four systems are the CarbonTracker Europe [ 44 ], the JENA-CO 2 inversion Version 3.1 [ 41 ] ( http://www.bgc-jena.mpg.de/∼christian.roedenbeck/download_CO 2 / ) and two LSCE inversions [ 42 ] (F Chevallier et al. , CO 2 surface fluxes at grid point scale estimated from a 21-year reanalysis of atmospheric measurements. J Geophys Res, unpublished data). The fluxes are calculated over different periods by each inversion. We extracted results from the period 2000–2004 common to the four systems over the six countries for which summary land-based C balance estimates have been published ( Figure 1 ). We have only an estimate of the internal precision of inversion fluxes at the scale of the six countries for the system of Chevallier et al. (CO 2 surface fluxes at grid point scale estimated from a 21-year reanalysis of atmospheric measurements. J Geophys Res, unpublished data). So, this uncertainty estimate was applied by default to each inversion. In Ref. Chevallier et al. (CO 2 surface fluxes at grid point scale estimated from a 21-year reanalysis of atmospheric measurements. J Geophys Res, unpublished data), there is an error covariance among optimized fluxes in each grid point implying that errors of the mean flux in a given region are not simply obtained by the propagation of independent grid point errors. This also implies that the NH sink error is smaller than the quadratic sum of errors in the six subregions. The average of the four inversions of NH land sink is 1.7 Pg C year −1 for the six regions considered, which represent 65% of the NH land and 71% of the NH vegetated land. The range of the four inversion best estimates is 0.8 Pg C year −1 . We note that this range is compatible with the precision of a single inversion, which is of 0.9 Pg C year −1 (1-sigma). Inversion precision reflects some of the arbitrary settings, uncertain data and models, and sparse measurements. The interannual variability is on average 0.4 Pg C year −1 over the period 2000–2004. Our NH land sink estimate over 2000–2004 thus falls in good agreement with that of Stephens et al. [ 8 •• ], although the period they covered is different (1996–2002). The small between-inversion range for Russia and China shown by Figure 1 is surprising, given the scarcity of atmospheric CO 2 measurement stations over these two regions. We speculate that all the inversions consistently place their flux increment into these two most poorly observed regions, and end up with similar values there. By contrast, over regions with higher density of atmospheric stations like USA and Europe, the between-inversion range is larger than the mean sink. Finally, it is important to note that the inversions show discrepancy in their prescribed fossil fuel emissions ( Figure 1 ). There is an NH fossil fuel emission range of 0.24 Pg C year −1 , and an emission range in any of the six regions studied below 0.2 Pg C year −1 . This discrepancy between prescribed emissions is surprisingly large for the USA and Canada, where fossil fuel emissions are known from inventories within 5% [ 45 ]. In the EU-25, inconsistent system boundaries between the different countries for reporting emissions could increase the inventories uncertainty by up to 7% [ 18 ]. Here we found that the discrepancy between emissions used in the four inversions for each region is as large as the worst case inventory uncertainty of 20% reported for China [ 45 ]. Discrepancies between prescribed fossil fuel emissions explain roughly 25% of the range of inversion results for the NH terrestrial sink. Are inversions consistent with land-based accounting? The inversion CO 2 fluxes are compared with land-based C flux estimates from the SOCCR report for USA, Mexico, and Canada [ 19 ], the CARBOEUROPE-IP program for EU-25 [ 11,46 ], the International Institute for Applied System Analysis (IIASA) Russian C accounting project's most recent estimates for Russia [ 21 ] and a recent synthesis for China [ 12 ]. There is no consistent terrestrial C budget estimation for India and northern hemispheric South Asian countries. Thus, our NH sink estimate corresponds roughly to the NH mid to high latitudes north of 20°N. It is also worth noting that the six regions included do not contain Ukraine, Belarus, Mongolia, Koreas and Japan, which altogether represent 4% of the land area north of 20°N. We stress the fact that the inversion period is shorter than the period covered by C accounting studies. Therefore, part of the differences between top-down and bottom-up is caused by inconsistent time periods. It is striking to see in Figure 1 that, for each large region, the mean of the four inversions is in agreement with the land-based estimate. One region where the mean of the inversions consistently gives a smaller sink than the land-based approach is the USA. This is possibly reflecting the effect of droughts in 2002 and 2003 in the USA that may have lowered the mean CO 2 uptake in that region during the inversion period 2000–2004. Land-based estimates are not restricted or centered over the same period. In contrast, the inversions over China provide a much larger CO 2 sink than the C accounting approach. This may be due to lateral fluxes or sparse atmospheric station coverage. We speculated above that inversions might add flux increments to balance the meridional CO 2 gradient into the most uncertain and yet productive grid points, thus typically in regions like China. The large inversion uncertainty for China and the EU25 contrasts with the small uncertainty of land-based studies. The terrestrial CO 2 sink from inversions is expected to be systematically higher than the ecosystem C sequestration inferred from bottom-up accounting methods. This is because CO 2 sinks from inversions do not pick up ecosystem C losses from biogenic volatile organic compounds and wildfire CO emissions. In addition, incomplete fossil fuel combustion emits CO, so that the ‘residual’ terrestrial CO 2 sink estimated by inversions is lower than if all fossil fuel emissions were assumed to be made of CO 2 as in the four models of this study. Thirdly, ecosystem carbon losses by river conduits and harvest will be seen by inversions but not by land-based inventories of stock changes. In summary, the NH terrestrial sink estimation is consistent between the top-down and the bottom-up approaches at the scale of the regions considered, given uncertainties associated with each method. The larger range of uncertainty attributed to inversions shed doubts on the ability of this approach to accurately constrain the NH land sink. Indeed, it will take a long time until enough atmospheric sampling stations cover the NH continents to reduce the internal error of inversions. The between-model inversion uncertainty range is within the internal uncertainty of each inversion. Inversion as a method thus provides consistent uncertainty estimates for the four models used in this study. The uncertainty that we report here for inversions is more comprehensive than the one compiled from land-based studies. In particular, the uncertainty of the land-based estimate does not include between-model differences, and is based on expert judgment for some components. In addition, land C inventories are really limited to changes in above ground carbon stocks and really do not assess changes in below ground carbon stocks in a measurable sense (models are used instead to infer soil C changes). The NH land sink best estimate from this study is of 1.7 Pg C year −1 , with a 1-sigma error of 0.9 Pg C year −1 in inversions and of 0.3 Pg C year −1 in land-based accounting. In answer to the question raised in the title of this paper, we can therefore say that the reviewed atmospheric and land-based estimates of the six NH regions considered are consistent with each other. Yet, inversion biases are not estimated, and only one land-based study was available for each region. Twenty years after the Tans et al. [ 24 •• ] and Keeling et al. [ 26 ] studies, the surface atmospheric station network has become denser over the NH oceans in the late 1990s and over North America and Europe in the early 2000s, yet leaving Siberia, China, India and South Asia with insufficient coverage by the network of atmospheric stations. Stephens et al. [ 8 •• ] also raised the issue that even an ensemble of 13 models can be consistently biased for vertical transport. Here, using a smaller set of four recent inversions, we find that the spread of mean inversion fluxes is smaller than it was in the TRANSCOM [ 1 ] study. This gives some hope that between-model differences may be reducing over time. It is time to conduct an update of the TRANSCOM intercomparison program, for comparing both inversion methods and transport models, in order to better assess if the inversion precision is improving over time. An updated comparison will allow an assessment whether model differences rather than data scarcity determine inversion uncertainties. Perhaps more importantly, both the inversion approach and the land-based accounting need independent evidence to be verified. Both denser networks of atmospheric stations and more accurate transport models are needed to improve inversions. Soil carbon inventories, and more systematic data on aquatic fluxes and wood products will help to increase the accuracy of bottom accounting, as well as designing dedicated bottom-up accounting intercomparison and cross-validation studies. References and recommended reading Papers of particular interest, published within the period of review, have been highlighted as: • of special interest •• of outstanding interest Acknowledgements This manuscript is a contribution to the synthesis efforts of the Global Carbon Project and to the development of the REgional Carbon Cycle Assessment and Processes (RECCAP). The study benefited from the Seventh Research Framework Programme of the European Commission (grant agreement No 212196, COCOS). References 1 K.R. Gurney R.M. Law A.S. Denning P.J. Rayner D. Baker P. Bousquet L. Bruhwiler Y.-H. Chen P. Ciais S. Fan Towards robust regional estimates of CO 2 sources and sinks using atmospheric transport models Nature 415 2002 626 630 2 T. Takahashi R. Wanninkhof R.A. Feely R.F. Weiss D.C.N. Bates J. Olafsson C. Sabine S.C. Sutherland Net sea-air CO 2 flux over the global oceans: an improved estimate based on the sea-air pCO 2 difference Proceedings of the 2nd CO 2 in Oceans Symposium Tsukuba, Japan 1999 3 T. Takahashi S.C. Sutherland R. Wanninkhof C. Sweeney R.A. Feely D.W. Chipman B. Hales G. Friederich F. Chavez C. Sabine Climatological mean and decadal change in surface ocean pCO 2 , and net sea-air CO 2 flux over the global oceans Deep Sea Res Part II: Top Stud Oceanogr 56 2009 554 577 4 A.J. Watson U. Schuster D.C.E. Bakker N.R. Bates A. Corbiere M. Gonzalez-Davila T. Friedrich J. Hauck C. Heinze T. Johannessen Tracking the Variable North Atlantic sink for atmospheric CO 2 Science 326 2009 1391 1393 5 Trolier Monitoring the isotopic composition of atmospheric CO 2 : measurements from the NOAA global air sampling network J Geophys Res 101 1996 D20 10.1029/96JD02363 6 P. Ciais P. Tans J.W. White M. Trolier R. Francey J. Berry D. Randall P. Sellers J. Collatz D.S. Schimel Partitioning of the ocean and land uptake of CO 2 from del13C measurements from the NOAA/CMDL global air sampling network J Geophys Res 100 1995 D3 7• R. Keeling S.C. Piper M. Heimann Global and hemispheric CO 2 sinks deduced from changes in atmospheric O 2 concentration Nature 381 1996 218 221 This paper is one of the seminal studies that first put in evidence using atmospheric observations, the existence of a northern hemispheric sink. 8•• B.B. Stephens K.R. Gurney P.P. Tans C. Sweeney W. Peters L. Bruhwiler P. Ciais M. Ramonet P. Bousquet T. Nakazawa Weak northern and strong tropical land carbon uptake from vertical profiles of atmospheric CO 2 Science 316 2007 1732 1735 This paper was the first to show that the ensemble of 13 different atmospheric transport models of the TRANSCOM project was consistently biased toward a too slow vertical mixing in winter times in the Northern Hemisphere. One key implication is that the Northern sink analyzed from inversions has to be lowered and, from mass balance, the tropical sink has to be increased. 9 S.W. Pacala G.C. Hurtt D. Baker P. Peylin R.A. Houghton R.A. Birdsey L. Heath E.T. Sundquist R.F. Stallard P. Ciais Consistent land- and atmosphere-based U.S. carbon sink estimates Science 292 2001 2316 2320 10 I.A. Janssens A. Freibauer P. Ciais P. Smith G.J. Nabuurs G. Folberth B. Schlamadinger R. Hutjes R. Ceulemans E.-D. Schulze Europe's terrestrial biosphere absorbs 7 to 12% of European anthropogenic emissions Science 300 2003 1538 1542 11 E.D. Schulze S. Luyssaert P. Ciais A. Freibauer I.A. Janssens J.F. Soussana P. Smith J. Grace I. Levin B. Thiruchittampalam Importance of methane and nitrous oxide for Europe's terrestrial greenhouse-gas balance Nat Geosci 2 2009 842 850 12 S.L. Piao J.Y. Fang P. Ciais P. Peylin Y. Huang S. Sitch T. Wang The carbon balance of terrestrial ecosystems in China Nature 458 2009 U1009 U1082 13 D. Baldocchi E. Falge L. Gu R. Olson D. Hollinger S. Running P. Anthoni C. Bernhofer K. Davis R. Evans FLUXNET: a new tool to study the temporal and spatial variability of ecosystem-scale carbon dioxide, water vapor, and energy flux densities Bull Am Meteorol Soc 82 2001 2415 2434 14 M. Jung M. Verstraete N. Gobron M. Reichstein D. Papale A. Bondeau M. Robustelli B. Pinty Diagnostic assessment of European gross primary production Global Change Biol 14 2008 2349 2364 15 D. Papale A. Valentini A new assessment of European forests carbon exchanges by eddy fluxes and artificial neural network spatialization Global Change Biol 9 2003 525 535 16 P.H. Bellamy P.J. Loveland R.I. Bradley R.M. Lark G.J.D. Kirk Carbon losses from all soils across England and Wales 1978–2003 Nature 437 2005 245 248 17 C.L. Goodale M.J. Apps R.A. Birdsey C.B. Field L.S. Heath R.A. Houghton J.C. Jenkins G.H. Kohlmaier W. Kurz S.R. Liu Forest carbon sinks in the Northern Hemisphere Ecol Appl 12 2002 891 899 18 P. Ciais M. Wattenbach N. Vuichard P. Smith S.L. Piao A. Don S. Luyssaert I.A. Janssens A. Bondeau R. Dechow The European carbon balance. Part 2: croplands Global Change Biol 16 2010 1409 1428 19 SOCCR A.W. King L. Dilling G.P. Zimmerman D.M. Fairman R.A. Houghton G. Marland A.Z. Rose T.J. Wilbanks The First State of the Carbon Cycle Report (SOCCR): The North American Carbon Budget and Implications for the Global Carbon Cycle. A Report by the U. S. Climate Change Science Program and the Subcommittee on Global Change Research 2007 National Oceanic and Atmospheric Administration, National Climatic Data Center Asheville, NC, USA 242 20 M. Jonas M. Gusti S. Nilsson Assessing uncertainty in bottom-up full carbon accounting for Russia Brief project report (revised and expanded after acceptance by the Austrian Science Fund) 2008 IIASA Laxenburg, Austria 8 21 A. Shvidenko D. Schepaschenko S. Maksyutov Impact of terrestrial ecosystems of Russia on global carbon cycle in 2003–2008: an attempt of synthesis Proceedings of the International Science Conference ENVRIMIS-2010 Tomk, Russia, July 5–12 2010 22 P. Ciais A.V. Borges G. Abril M. Meybeck G. Folberth D. Hauglustaine I.A. Janssens The impact of lateral carbon fluxes on the European carbon balance Biogeosciences 5 2008 1259 1271 23 P. Ciais P. Bousquet A. Freibauer T. Naegler Horizontal displacement of carbon associated with agriculture and its impacts on atmospheric CO 2 Global Biogeochem Cycles 21 2007 GB2014 24•• P.P. Tans I.Y. Fung T. Takahashi Observational constraints on the global atmospheric CO 2 budget Science 247 1990 1431 1438 This paper is one (together with Ref. [ 7 • ]) of the seminal studies that first put in evidence using atmospheric observations, the existence of a northern hemispheric sink. 25 C.D. Keeling R.B. Bacastow A.F. Carter S.C. Piper T.P. Whorf M. Heimann W.G. Mook H.A. Roeloffzen A three-dimensional model of atmospheric CO 2 transport based on observed winds. 1. Analysis of observational data D.H. Peterson Aspects of Climate Variability in the Pacific and Western Americas, Geophysical Monogr. Ser. vol 55 1989 AGU 165 236 26 C.D. Keeling S.C. Piper M. Heimann A three-dimensional model of atmospheric CO 2 transport based on observed winds. 4. Mean annual gradients and interannual variations D.H. Peterson Aspects of Climate Variability in the Pacific and Western Americas, Geophysical Monograph 55 1989 American Geophysical Union 27 P. Bousquet P. Peylin P. Ciais M. Ramonet P. Monfray Inverse modeling of annual atmospheric CO 2 sources and sinks. 2. Sensitivity study J Geophys Res: Atmos 104 1999 26179 26193 28 P.J. Rayner R.M. Law A comparison of modelled responses to prescribed CO 2 sources Aust. Atmos. Res. Tech. Pap. 1995 CSIRO 29 R.M. Law P.J. Rayner A.S. Denning D. Erickson I.Y. Fung M. Heimann S.C. Piper M. Ramonet S. Taguchi J.A. Taylor Variations in modeled atmospheric transport of carbon dioxide and the consequences for CO 2 inversions Global Biogeochem Cycles 10 1996 783 796 30 A.S. Denning I.Y. Fung D. Randall Latitudinal gradient of atmospheric CO 2 due to seasonal exchange with land biota Nature 376 1995 240 243 31 K.R. Gurney R.M. Law A.S. Denning P.J. Rayner D. Baker P. Bousquet L. Bruhwiler Y.-H. Chen P. Ciais S. Fan TransCom 3 CO 2 inversion intercomparison. 1. Annual mean control results and sensitivity to transport and prior flux information Tellus B 55 2003 555 579 32 K.R. Gurney R.M. Law A.S. Denning P.J. Rayner B.C. Pak D. Baker P. Bousquet L. Bruhwiler Y.H. Chen P. Ciais Transcom 3 inversion intercomparison: model mean results for the estimation of seasonal carbon sources and sinks Global Biogeochem Cycles 2004 18 33 D.B. Baker R.M. Law K.R. Gurney P.J. Rayner P. Peylin A.S. Denning P. Bousquet L. Bruhwiler Y.-H. Chen P. Ciais TransCom 3 inversion intercomparison: impact of transport model errors on the interannual variability of regional CO 2 fluxes, 1988–2003 Global Biogeochem Cycles 20 2006 GB1002 10.1029/2004GB002439 34 P.K. Patra M. Ishizawa S. Makshyutov T. Nakazawa G. Inoue Role of biomass burning and climate anomalies for land-atmosphere carbon fluxes based on inverse modeling of atmospheric CO 2 Global Biogeochem Cycles 2005 10 35 S. Maksyutov T. Machida H. Mukai P.K. Patra T. Nakazawa G. Inoue Transcom3 M Effect of recent observations on Asian CO 2 flux estimates by transport model inversions Tellus B 55 2003 522 529 36 P. Bousquet P. Ciais P. Peylin M. Ramonet P. Monfray Inverse modeling of annual atmospheric CO2 sources and sinks 1. Method and control inversion J Geophys Res 104 D21 1999 26161 26178 37 P. Peylin P. Bousquet C. LeQuéré S. Sitch P. Friedlingstein G.A. McKinley N. Gruber P. Rayner P. Ciais Multiple constraints on regional CO 2 fluxes variations over land and oceans Global Biogeochem Cycles 2005 19 38 P.K. Patra K.R. Gurney A.S. Denning S. Maksyutov T. Nakazawa D. Baker P. Bousquet L. Bruhwiler Y.-H. Chen P. Ciais Sensitivity of inverse estimation of annual mean CO 2 sources and sinks to ocean-only sites versus all-sites observational networks Geophys Res Lett 33 2006 L05814 39 I.C. Prentice G.D. Farquhar M.J.R. Fasham M.L. Goulden M. Heimann V.J. Jaramillo H.S. Kheshgi C. Le Quéré R.S. Scholes D.W.R. Wallace The carbon cycle and atmospheric carbon dioxide IPCC Climate Change 2001, The Scientific Basis 2001 Cambridge University Press Chapter 3 40 K.L. Denman G. Brasseur A. Chidthaisong P. Ciais P. Cox R.E. Dickinson D. Haugustaine C. Heinze E. Holland D. Jacob Couplings between changes in the climate system and biogeochemistry S. Solomon D. Qin M. Manning Z. Chen M. Marquis K.B. Averyt M. Tignor H.L. Miller Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change 2007 Cambridge University Press 499 587 41 C. Rodenbeck S. Houweling M. Gloor M. Heimann CO 2 flux history 1982–2001 inferred from atmospheric data using a global inversion of atmospheric transport Atmos Chem Phys 3 2003 1919 1964 42 P. Peylin P.J. Rayner P. Bousquet C. Carouge F. Hourdin P. Heinrich P. Ciais Aerocarb C: daily CO 2 flux estimates over Europe from continuous atmospheric measurements: 1. Inverse methodology Atmos Chem Phys Discuss 5 2005 1647 1678 43 W. Peters M.C. Krol G.R. van der Werf S. Houweling C.D. Jones J. Hughes K. Schaefer K.A. Masarie A.R. Jacobson J.B. Miller C.H. Cho M. Ramonet Seven years of recent European net terrestrial carbon dioxide exchange constrained by atmospheric observations Global Change Biology 16 4 2010 1317 1337 44 W. Peters A.R. Jacobson C. Sweeney A.E. Andrews T.J. Conway K. Masarie J.B. Miller L.M.P. Bruhwiler G. Pétron A.I. Hirsch An atmospheric perspective on North American carbon dioxide exchange: CarbonTracker Proc Natl Acad Sci 104 2007 18925 18930 45 G. Marland K. Hamal M. Jonas How uncertain are estimates of CO 2 emissions? J Ind Ecol 13 2009 4 7 46 E.D. Schulze P. Ciais S. Luyssaert M. Schrumpf I.A. Janssens B. Thiruchittampalam J. Thelocke M. Saurat S. Bringezu J. Lelievekd The European carbon balance. Part 4. Integration of carbon and other trace-gas fluxes Global Change Biol 16 2010 1451 1469