Search

The European land carbon uptake has been heavily impacted by several recent severe droughts, yet quantitative estimates of carbon uptake anomalies are uncertain. Atmospheric CO2 inverse models (AIMs) provide observation‐based estimates of the large‐scale carbon flux dynamics, but how well they capture drought impacts on the terrestrial carbon uptake is poorly known. Here we assessed the capacity of state‐of‐the‐art AIMs in monitoring drought impacts on the European carbon uptake over 2001–2015 using observations of environmental variability and vegetation function and made comparisons with bottom‐up estimates of carbon uptake anomalies. We found that global inversions with only limited surface CO2 observations give divergent estimates of drought impacts. Regional inversions assimilating denser CO2 observations over Europe demonstrated some improved consistency, with all inversions capturing a reduction in carbon uptake during the 2012 drought. However, they failed to capture the reduction caused by the 2015 drought. Finally, we found that a set of inversions that assimilated satellite XCO2 or assimilated environmental variables plus surface CO2 observations better captured carbon uptake anomalies induced by both the 2012 and 2015 droughts. In addition, the recent Orbiting Carbon Observatory—2 XCO2 inversions showed good potential in capturing drought impacts, with better performances for larger‐scale droughts like the 2018 drought. These results suggest that surface CO2 observations may still be too sparse to fully capture the impact of drought on the carbon cycle at subcontinental scales over Europe, and satellite XCO2 and ancillary environmental data can be used to improve observational constraints in atmospheric inversion systems. Plain Language Summary: Atmospheric CO2 inverse models (AIMs) are useful tools for quantifying the response of large‐scale carbon uptake to climate extremes, but their capacity for monitoring drought impacts, particularly at regional scales, is not fully explored. In this study, we assessed the capacity of state‐of‐the‐art AIMs for monitoring drought impacts on the European land carbon uptake over 2001–2015 using a large array of observational and model data sets. We found: (a) global inversions with only limited surface CO2 observations face a great challenge in monitoring drought impacts on the European carbon uptake; (b) Regional inversions assimilated denser CO2 observations over Europe, for the EUROCOM project, demonstrated some improved consistency but are still deficient, showing divergent estimates in interannual variability of carbon uptake for most years; and (c) A set of inversion systems that assimilated satellite XCO2 or assimilated environmental variables plus surface CO2 observations better captured annual and seasonal anomalies caused by droughts. Our study demonstrates that surface CO2 observations may still be too sparse to fully capture the impact of drought on the carbon cycle at subcontinental scales over Europe, whereby satellite XCO2 and ancillary environmental data can offer observational constraints for improving the estimates. Key Points: Global inversions with only limited surface CO2 observations give divergent estimates of drought impacts on the European carbon uptakeRegional inversions assimilating denser CO2 observations over Europe demonstrate some improved consistency but are still deficientThe inversions assimilating satellite XCO2 or environmental variables in addition to surface CO2 largely improve the estimates [ABSTRACT FROM AUTHOR]

Gross primary productivity (GPP) is the largest flux in the global terrestrial carbon cycle. Drought has significantly impacted global terrestrial GPP in recent decades, and has been projected to occur with increasing frequency and intensity. However, the drought risk of global terrestrial GPP has not been well investigated. In this study, global terrestrial GPP during 1981–2016 was simulated with the process‐based Boreal Ecosystem Productivity Simulator model. Then, the drought risk of GPP was quantified as the product of drought probability and reduction of GPP caused by drought, which was determined using the standardized precipitation evapotranspiration index. During the study period, the drought risk of GPP was high in the southeastern United States, most of South America, southern Europe, central and eastern Africa, eastern and southeastern Asia, and eastern Australia. It was low at some high latitudes of the Northern Hemisphere and in part of tropical South America, where terrestrial GPP increased slightly in drought years. The drought risk of terrestrial GPP was greater during 2000–2016 than during 1981–1999 in 21 out of 24 climatic zones. The global mean drought risk of GPP increased from 13.6 g C m−2 yr−1 during 1981–1999 to 19.3 g C m−2 yr−1 during 2000–2016. The increase in drought risk of GPP was mainly caused by the increase in drought vulnerability. Simulation experiments indicated that the drought vulnerability of GPP was mainly induced by climatic variability. This study advances our understanding on the impact of drought on GPP over the globe. Plain Language Summary: Drought generally affects gross primary productivity (GPP) of terrestrial ecosystems. However, the drought risk of global terrestrial GPP has not been well investigated. We assessed the drought risk of GPP based on the simulation by a process‐based model for the years 1981–2016. The drought risk of GPP was quantified in terms of the chances of drought occurring and the anticipated reduction of GPP that would be caused by the drought. The results were variable across the world, being serious in the southeastern United States, most of South America, southern Europe, central and eastern Africa, eastern and southeastern Asia, and eastern Australia. The drought risk of GPP at some high latitudes of the Northern Hemisphere and in part of tropical South America was not found to be problematic. In those areas, GPP increased slightly in drought years in comparison with that in normal years. The drought risk of terrestrial GPP was greater during 2000–2016 than during 1981–1999 in 21 out of 24 climatic zones. The global mean drought risk of GPP increased from 13.6 g C m−2 yr−1 during 1981–1999 to 19.3 g C m−2 yr−1 during 2000–2016. The increase in drought risk of GPP was mainly caused by climatic variability. Key Points: Drought vulnerability and risk of reductions in global terrestrial gross primary productivity (GPP) were assessed using a remote‐sensing, process‐based modelHigh risk located in southeastern United States, southern Europe, most of South America, central and eastern Africa, and eastern AustraliaThe risk of reductions in global GPP was greater in 2000–2016 than in 1981–1999, mainly owing to the increase in drought vulnerability [ABSTRACT FROM AUTHOR]

Abstract: Recently, severe droughts that occurred in North America are likely to have impacted its terrestrial carbon sink. However, process‐based understanding of how meteorological conditions prior to the onset of drought, for instance warm or cold springs, affect drought‐induced carbon cycle effects remains scarce. Here we assess and compare the response of terrestrial carbon fluxes to summer droughts in 2011 and 2012 characterized by contrasting spring conditions. The analysis is based on a comprehensive ensemble of carbon cycle models, including FLUXCOM, TRENDY v5, SiBCASA, CarbonTracker Europe, and CarbonTracker, and emerging Earth observations. In 2011, large reductions of net ecosystem production (NEP; −0.24 ± 0.17 Pg C/year) are due to decreased gross primary production (−0.17 ± 0.18 Pg C/year) and slightly increased ecosystem respiration (+0.07 ± 0.17 Pg C/year). Conversely, in 2012, NEP reductions (−0.17 ± 0.25 Pg C/year) are attributed to a larger increase of ecosystem respiration (+0.48 ± 0.27 Pg C/year) than gross primary production (+0.31 ± 0.29 Pg C/year), induced predominantly by an extra warmer spring prior to summer drought. Two temperate ecoregions crops/agriculture and the grass/shrubs contribute largest to these reductions and also dominate the interannual variations of NEP during 2007–2014. Moreover, the warming spring compensated largely the negative carbon anomaly due to summer drought, consistent with earlier studies; however, the compensation occurred only in some specific ecoregions. Overall, our analysis offers a refined view on recent carbon cycle variability and extremes in North America. It corroborates earlier results but also highlights differences with respect to ecoregion‐specific carbon cycle responses to drought and heat. [ABSTRACT FROM AUTHOR]