Search

The Australian continent contributes substantially to the year-to-year variability of the global terrestrial carbon dioxide (CO2) sink. However, the scarcity of in-situ observations in remote areas prevents deciphering the processes that force the CO2 flux variability. Here, examining atmospheric CO2 measurements from satellites in the period 2009-2018, we find recurrent end-of-dry-season CO2 pulses over the Australian continent. These pulses largely control the year-to-year variability of Australia's CO2 balance, due to 2-3 times higher seasonal variations compared to previous top-down inversions and bottom-up estimates. The CO2 pulses occur shortly after the onset of rainfall and are driven by enhanced soil respiration preceding photosynthetic uptake in Australia's semi-arid regions. The suggested continental-scale relevance of soil rewetting processes has large implications for our understanding and modelling of global climate-carbon cycle feedbacks., Comment: 28 pages (including supplementary materials), 3 main figures, 7 supplementary figures; v2 changes: Last name of first author changed

We assess the land surface model JSBACHv4 (Jena Scheme for Biosphere Atmosphere Coupling in Hamburg version 4), which was recently developed at the Max Planck Institute for Meteorology as part of the effort to build the new Icosahedral Nonhydrostatic (ICON) Earth system model (ESM), ICON-ESM. We assess JSBACHv4 in simulations coupled with ICON-A, the atmosphere model of ICON-ESM, hosting JSBACHv4 as land component to provide the surface boundary conditions. The assessment is based on a comparison of simulated albedo, land surface temperature (LST), leaf area index (LAI), terrestrial water storage (TWS), fraction of absorbed photosynthetic active radiation (FAPAR), net primary production (NPP), and water use efficiency (WUE) with corresponding observational data. JSBACHv4 is the successor of JSBACHv3; therefore, another purpose of this study is to document how this step in model development has changed model biases. This is achieved by also assessing, in parallel, the results of coupled land–atmosphere simulations with the preceding model ECHAM6 hosting JSBACHv3. Large albedo biases appear in both models over ice sheets and in central Asia. The temperate to boreal warm bias observed in simulations with JSBACHv3 largely remained in JSBACHv4, despite the very good agreement with observed LST in the global mean. For the assessment of changes in land water storage, a novel procedure is suggested to compare the gravitational data from the Gravity Recovery And Climate Experiment (GRACE) satellites to simulated TWS. It turns out that the agreement of the changes in the seasonal cycle of TWS is sensitive to the representation of precipitation in the atmosphere model. The LAI is generally too high, which is partly caused by too high soil moisture and also by the parameterization of the phenology itself. The pattern of WUE is, for both models, largely as observed. In India, WUE is too high, probably because JSBACH does not incorporate irrigation in our simulations. WUE differences between the two models can be traced back to differences in precipitation patterns in the two coupled land–atmosphere simulations. For both models, most NPP biases can be associated with biases in water stress, LAI, and FAPAR. In particular, the NPP bias of the Eurasian steppes has switched from positive in JSBACHv3 to negative in JSBACHv4. This difference is mainly caused by weaker precipitation and lower FAPAR of ICON-A–JSBACHv4 in July, which is most probably caused by a feedback loop between too little soil moisture, evaporation, and clouds. While the size and patterns of biases in albedo and LST are largely similar between the two model versions, they are less well correlated for precipitation- and vegetation-related variables like FAPAR. Overall, the biases found in the different assessment variables are either already known from the previous implementation in the Max Planck Institute Earth System Model (MPI-ESM) or have changed because of the coupling with the new atmospheric component ICON-A. Accordingly, this study demonstrates the technically successful completion of the re-implementation of JSBACH into ICON-ESM-V1. As discussed, there is a good perspective on mitigating the biases by an improved representation of the processes.

In many regions of the world, frequent and continual dry spells are exacerbating drought conditions, which have severe impacts on vegetation biomes. Vegetation in southern Africa is among the most affected by drought. Here, we assessed the spatiotemporal characteristics of meteorological drought in southern Africa using the standardized precipitation evapotranspiration index (SPEI) over a 30-year period (1982–2011). The severity and the effects of droughts on vegetation productiveness were examined at different drought timescales (1- to 24-month timescales). In this study, we characterized vegetation using the leaf area index (LAI) after evaluating its relationship with the normalized difference vegetation index (NDVI). Correlating the LAI with the SPEI, we found that the LAI responds strongly (r=0.6) to drought over the central and southeastern parts of the region, with weaker impacts (r

The Amazon forests are one of the largest ecosystem carbon pools on Earth. Although more frequent and prolonged future droughts have been predicted, the impacts have remained largely uncertain, as most land surface models (LSMs) fail to capture the vegetation drought responses. In this study, the ability of the LSM JSBACH to simulate the drought responses of leaf area index (LAI) and leaf litter production in the Amazon forests is evaluated against artificial drought experiments. Based on the evaluation, improvements are implemented, including a dependency of leaf growth on leaf carbon allocation and a better representation of drought-dependent leaf shedding. The modified JSBACH is shown to capture the drought responses at two sites and across different regions of the basin. It is then coupled with an atmospheric model to simulate the carbon and biogeophysical feedbacks of drought under future climate. We separate the drought impacts into (a) the direct effect, resulting from drier soil and stomatal closure, which does not involve a change in canopy structure, and (b) the LAI effect, resulting from leaf shedding and involving canopy response. We show that the latter accounts for 35% of reduced land carbon uptake (9 ± 10 vs. 26 ± 7 g/m2/yr; mean ± 1 sd) and 12% of surface warming (0.09 ± 0.03 vs. 0.7 ± 0.07 K) during the late 21st century. A north-south dipole of precipitation change is found, which is largely attributable to the direct effect. The results highlight the importance of incorporating drought deciduousness of tropical rainforests in LSMs to better simulate land-atmosphere interactions in the future.

As the focus of climate policy shifts from pledges to implementation, there is a growing need to track progress on climate change mitigation at the country level, particularly for the land-use sector. Despite new tools and models providing unprecedented monitoring opportunities, striking differences remain in estimations of anthropogenic land-use CO2 fluxes between, on the one hand, the national greenhouse gas inventories (NGHGIs) used to assess compliance with national climate targets under the Paris Agreement and, on the other hand, the Global Carbon Budget and Intergovernmental Panel on Climate Change (IPCC) assessment reports, both based on global bookkeeping models (BMs). Recent studies have shown that these differences are mainly due to inconsistent definitions of anthropogenic CO2 fluxes in managed forests. Countries assume larger areas of forest to be managed than BMs do, due to a broader definition of managed land in NGHGIs. Additionally, the fraction of the land sink caused by indirect effects of human-induced environmental change (e.g. fertilisation effect on vegetation growth due to increased atmospheric CO2 concentration) on managed lands is treated as non-anthropogenic by BMs but as anthropogenic in most NGHGIs. We implement an approach that adds the CO2 sink caused by environmental change in countries' managed forests (estimated by 16 dynamic global vegetation models, DGVMs) to the land-use fluxes from three BMs. This sum is conceptually more comparable to NGHGIs and is thus expected to be quantitatively more similar. Our analysis uses updated and more comprehensive data from NGHGIs than previous studies and provides model results at a greater level of disaggregation in terms of regions, countries and land categories (i.e. forest land, deforestation, organic soils, other land uses). Our results confirm a large difference (6.7 GtCO2 yr−1) in global land-use CO2 fluxes between the ensemble mean of the BMs, which estimate a source of 4.8 GtCO2 yr−1 for the period 2000–2020, and NGHGIs, which estimate a sink of −1.9 GtCO2 yr−1 in the same period. Most of the gap is found on forest land (3.5 GtCO2 yr−1), with differences also for deforestation (2.4 GtCO2 yr−1), for fluxes from other land uses (1.0 GtCO2 yr−1) and to a lesser extent for fluxes from organic soils (0.2 GtCO2 yr−1). By adding the DGVM ensemble mean sink arising from environmental change in managed forests (−6.4 GtCO2 yr−1) to BM estimates, the gap between BMs and NGHGIs becomes substantially smaller both globally (residual gap: 0.3 GtCO2 yr−1) and in most regions and countries. However, some discrepancies remain and deserve further investigation. For example, the BMs generally provide higher emissions from deforestation than NGHGIs and, when adjusted with the sink in managed forests estimated by DGVMs, yield a sink that is often greater than NGHGIs. In summary, this study provides a blueprint for harmonising the estimations of anthropogenic land-use fluxes, allowing for detailed comparisons between global models and national inventories at global, regional and country levels. This is crucial to increase confidence in land-use emissions estimates, support investments in land-based mitigation strategies and assess the countries' collective progress under the Global Stocktake of the Paris Agreement. Data from this study are openly available online via the Zenodo portal (Grassi et al., 2023) at https://doi.org/10.5281/zenodo.7650360.

Satellite data reveal widespread changes in Earth's vegetation cover. Regions intensively attended to by humans are mostly greening due to land management. Natural vegetation, on the other hand, is exhibiting patterns of both greening and browning in all continents. Factors linked to anthropogenic carbon emissions, such as CO2 fertilization, climate change, and consequent disturbances such as fires and droughts, are hypothesized to be key drivers of changes in natural vegetation. A rigorous regional attribution at the biome level that can be scaled to a global picture of what is behind the observed changes is currently lacking. Here we analyze different datasets of decades-long satellite observations of global leaf area index (LAI, 1981–2017) as well as other proxies for vegetation changes and identify several clusters of significant long-term changes. Using process-based model simulations (Earth system and land surface models), we disentangle the effects of anthropogenic carbon emissions on LAI in a probabilistic setting applying causal counterfactual theory. The analysis prominently indicates the effects of climate change on many biomes – warming in northern ecosystems (greening) and rainfall anomalies in tropical biomes (browning). The probabilistic attribution method clearly identifies the CO2 fertilization effect as the dominant driver in only two biomes, the temperate forests and cool grasslands, challenging the view of a dominant global-scale effect. Altogether, our analysis reveals a slowing down of greening and strengthening of browning trends, particularly in the last 2 decades. Most models substantially underestimate the emerging vegetation browning, especially in the tropical rainforests. Leaf area loss in these productive ecosystems could be an early indicator of a slowdown in the terrestrial carbon sink. Models need to account for this effect to realize plausible climate projections of the 21st century.

Observations from the Orbiting Carbon Observatory 2 (OCO-2) satellite have been used to estimate CO2 fluxes in many regions of the globe and provide new insight into the global carbon cycle. The objective of this study is to infer the relationships between patterns in OCO-2 observations and environmental drivers (e.g., temperature, precipitation) and therefore inform a process understanding of carbon fluxes using OCO-2. We use a multiple regression and inverse model, and the regression coefficients quantify the relationships between observations from OCO-2 and environmental driver datasets within individual years for 2015–2018 and within seven global biomes. We subsequently compare these inferences to the relationships estimated from 15 terrestrial biosphere models (TBMs) that participated in the TRENDY model inter-comparison. Using OCO-2, we are able to quantify only a limited number of relationships between patterns in atmospheric CO2 observations and patterns in environmental driver datasets (i.e., 10 out of the 42 relationships examined). We further find that the ensemble of TBMs exhibits a large spread in the relationships with these key environmental driver datasets. The largest uncertainty in the models is in the relationship with precipitation, particularly in the tropics, with smaller uncertainties for temperature and photosynthetically active radiation (PAR). Using observations from OCO-2, we find that precipitation is associated with increased CO2 uptake in all tropical biomes, a result that agrees with half of the TBMs. By contrast, the relationships that we infer from OCO-2 for temperature and PAR are similar to the ensemble mean of the TBMs, though the results differ from many individual TBMs. These results point to the limitations of current space-based observations for inferring environmental relationships but also indicate the potential to help inform key relationships that are very uncertain in state-of-the-art TBMs.

International audience; Abstract Global estimates of the land carbon sink are often based on simulations by terrestrial biosphere models. The use of a large number of models that differ in their underlying hypotheses, structure and parameters is one way to assess the uncertainty in the historical land carbon sink. Here we show that the atmospheric forcing datasets used to drive these terrestrial biosphere models represent a significant source of uncertainty that is currently not systematically accounted for in land carbon cycle evaluations. We present results from three terrestrial biosphere models each forced with three different historical atmospheric forcing reconstructions over the period 1850-2015. We perform an analysis of variance to quantify the relative uncertainty in carbon fluxes arising from the models themselves, atmospheric forcing, and model-forcing interactions. We find that atmospheric forcing in this set of simulations plays a dominant role on uncertainties in global gross primary productivity (75% of variability) and autotrophic respiration (90%), and a significant but reduced role on net primary productivity and heterotrophic respiration (30%). Atmospheric forcing is the dominant driver (52%) of variability for the net ecosystem flux, defined as the difference between gross primary productivity and respiration (both autotrophic and heterotrophic respiration). In contrast, for wildfire-driven carbon emissions model uncertainties dominate and, as a result, model uncertainties dominate for net ecosystem exchange. At regional scales, the contribution of atmospheric forcing to uncertainty shows a very heterogeneous pattern and is smaller on average than at the global scale. We find that this difference in the relative importance of forcing uncertainty between global and regional scales is related to large differences in regional model flux estimates, which partially offset each other when integrated globally, while the flux differences driven by forcing are mainly consistent across world and therefore which add up to a larger fractional contribution to global uncertainty.

Accurate assessment of anthropogenic carbon dioxide (CO2) emissions and their redistribution among the atmosphere, ocean, and terrestrial biosphere in a changing climate is critical to better understand the global carbon cycle, support the development of climate policies, and project future climate change. Here we describe and synthesize data sets and methodology to quantify the five major components of the global carbon budget and their uncertainties. Fossil CO2 emissions (EFOS) are based on energy statistics and cement production data, while emissions from land-use change (ELUC), mainly deforestation, are based on land-use and land-use change data and bookkeeping models. Atmospheric CO2 concentration is measured directly, and its growth rate (GATM) is computed from the annual changes in concentration. The ocean CO2 sink (SOCEAN) is estimated with global ocean biogeochemistry models and observation-based data-products. The terrestrial CO2 sink (SLAND) is estimated with dynamic global vegetation models. The resulting carbon budget imbalance (BIM), the difference between the estimated total emissions and the estimated changes in the atmosphere, ocean, and terrestrial biosphere, is a measure of imperfect data and understanding of the contemporary carbon cycle. All uncertainties are reported as ±1σ. For the first time, an approach is shown to reconcile the difference in our ELUC estimate with the one from national greenhouse gases inventories, supporting the assessment of collective countries’ climate progress. For the year 2020, EFOS declined by 5.4 % relative to 2019, with fossil emissions at 9.5 ± 0.5 GtC yr−1 (9.3 ± 0.5 GtC yr−1 when the cement carbonation sink is included), ELUC was 0.9 ± 0.7 GtC yr−1, for a total anthropogenic CO2 emission of 10.2 ± 0.8 GtC yr−1 (37.4 ± 2.9 GtCO2). Also, for 2020, GATM was 5.0 ± 0.2 GtC yr−1 (2.4 ± 0.1 ppm yr−1), SOCEAN was 3.0 ± 0.4 GtC yr−1 and SLAND was 2.9 ± 1 GtC yr−1, with a BIM of −0.8 GtC yr−1. The global atmospheric CO2 concentration averaged over 2020 reached 412.45 ± 0.1 ppm. Preliminary data for 2021, suggest a rebound in EFOS relative to 2020 of +4.9 % (4.1 % to 5.7 %) globally. Overall, the mean and trend in the components of the global carbon budget are consistently estimated over the period 1959–2020, but discrepancies of up to 1 GtC yr−1 persist for the representation of annual to semi-decadal variability in CO2 fluxes. Comparison of estimates from multiple approaches and observations shows: (1) a persistent large uncertainty in the estimate of land-use changes emissions, (2) a low agreement between the different methods on the magnitude of the land CO2 flux in the northern extra- tropics, and (3) a discrepancy between the different methods on the strength of the ocean sink over the last decade. This living data update documents changes in the methods and data sets used in this new global carbon budget and the progress in understanding of the global carbon cycle compared with previous publications of this data set (Friedlingstein et al., 2020; Friedlingstein et al., 2019; Le Quéré et al., 2018b, 2018a, 2016, 2015b, 2015a, 2014, 2013). The data presented in this work are available at https://doi.org/10.18160/gcp-2021 (Friedlingstein et al., 2021).

Southeast Asia is a region known for active land-use changes (LUC) over the past 60 years; yet, how trends in net CO2 uptake and release resulting from LUC activities (net LUC flux) have changed through past decades remains uncertain. The level of uncertainty in net LUC flux from process-based models is so high that it cannot be concluded that newer estimates are necessarily more reliable than older ones. Here, we examined net LUC flux estimates of Southeast Asia for the 1980s−2010s from older and newer sets of Dynamic Global Vegetation Model simulations (TRENDY v2 and v7, respectively), and forcing data used for running those simulations, along with two book-keeping estimates (H&N and BLUE). These estimates yielded two contrasting historical LUC transitions, such that TRENDY v2 and H&N showed a transition from increased emissions from the 1980s to 1990s to declining emissions in the 2000s, while TRENDY v7 and BLUE showed the opposite transition. We found that these contrasting transitions originated in the update of LUC forcing data, which reduced the loss of forest area during the 1990s. Further evaluation of remote sensing studies, atmospheric inversions, and the history of forestry and environmental policies in Southeast Asia supported the occurrence of peak emissions in the 1990s and declining thereafter. However, whether LUC emissions continue to decline in Southeast Asia remains uncertain as key processes in recent years, such as conversion of peat forest to oil-palm plantation, are yet to be represented in the forcing data, suggesting a need for further revision., Global Biogeochemical Cycles, 36 (1), ISSN:0886-6236, ISSN:1944-9224

Climate change (CLI), elevated CO2 concentration (CO2), and land use change (LUC) have strongly altered land evapotranspiration (ET) during the recent decades. The fingerprints of these drivers in ET change, however, have not previously been detected due to the lack of these three scenarios from global climate models (GCMs). Here we applied an optimal fingerprint method to detect and attribute ET change by integrated utilization of state-of-the-art global ecosystem models and GCMs. Results indicate that CLI provides the greatest contribution to increasing ET, and its fingerprint is detectable at different timescales. CO2 reduces ET in most areas covered by forests. LUC decreases ET over the tropics, while increases ET over temperate and high-latitude regions. To further subdivide the impacts of CLI, we extend the Budyko framework to quantify the contribution of precipitation (P) and potential evapotranspiration (PET) and find that the dominant role of CLI mainly depends on the contribution of P.

Monitoring the implementation of emission commitments under the Paris agreement relies on accurate estimates of terrestrial carbon fluxes. Here, we assimilate a 21

Natural and anthropogenic disturbances, in particular forest management, affect forest age structures all around the globe. Forest age structures in turn influence key land surface processes, such as photosynthesis and thus the carbon cycle. Yet, many dynamic global vegetation models (DGVMs), including those used as land surface models (LSMs) in Earth system models (ESMs), do not account for subgrid forest age structures, despite being used to investigate land-use effects on the global carbon budget or simulating biogeochemical responses to climate change. In this paper we present a new scheme to introduce forest age classes in hierarchical tile-based DGVMs combining benefits of recently applied approaches the first being a computationally efficient age-dependent simulation of all relevant processes, such as photosynthesis and respiration, using a restricted number of age classes and the second being the tracking of the exact forest age, which is a prerequisite for any implementation of age-based forest management. This combination is achieved by using the tile hierarchy to track the area fraction for each age on an aggregated plant functional type level, whilst simulating the relevant processes for a set of age classes. We describe how we implemented this scheme in JSBACH4, the LSM of the ICOsahedral Non-hydrostatic Earth system model (ICON-ESM). Subsequently, we compare simulation output to global observation-based products for gross primary production, leaf area index, and above-ground biomass to assess the ability of simulations with and without age classes to reproduce the annual cycle and large-scale spatial patterns of these variables. The comparisons show decreasing differences and increasing computation costs with an increasing number of distinguished age classes. The results demonstrate the benefit of the introduction of age classes, with the optimal number of age classes being a compromise between computation costs and error reduction.

Continuous atmospheric CO2 monitoring data indicate an increase in the amplitude of seasonal CO2-cycle exchange (SCANBP) in northern high latitudes. The major drivers of enhanced SCANBP remain unclear and intensely debated, with land-use change, CO2 fertilization and warming being identified as likely contributors. We integrated CO2-flux data from two atmospheric inversions (consistent with atmospheric records) and from 11 state-of-the-art land-surface models (LSMs) to evaluate the relative importance of individual contributors to trends and drivers of the SCANBP of CO2 fluxes for 1980–2015. The LSMs generally reproduce the latitudinal increase in SCANBP trends within the inversions range. Inversions and LSMs attribute SCANBP increase to boreal Asia and Europe due to enhanced vegetation productivity (in LSMs) and point to contrasting effects of CO2 fertilization (positive) and warming (negative) on SCANBP. Our results do not support land-use change as a key contributor to the increase in SCANBP. The sensitivity of simulated microbial respiration to temperature in LSMs explained biases in SCANBP trends, which suggests that SCANBP could help to constrain model turnover times.

Despite their sparse vegetation, dryland regions exert a huge influence over global biogeochemical cycles because they cover more than 40% of the world surface (Schimel 2010 Science 327 418–9). It is thought that drylands dominate the inter-annual variability (IAV) and long-term trend in the global carbon (C) cycle (Poulter et al 2014 Nature 509 600–3, Ahlstrom et al 2015 Science 348 895–9, Zhang et al 2018 Glob. Change Biol. 24 3954–68). Projections of the global land C sink therefore rely on accurate representation of dryland C cycle processes; however, the dynamic global vegetation models (DGVMs) used in future projections have rarely been evaluated against dryland C flux data. Here, we carried out an evaluation of 14 DGVMs (TRENDY v7) against net ecosystem exchange (NEE) data from 12 dryland flux sites in the southwestern US encompassing a range of ecosystem types (forests, shrub- and grasslands). We find that all the models underestimate both mean annual C uptake/release as well as the magnitude of NEE IAV, suggesting that improvements in representing dryland regions may improve global C cycle projections. Across all models, the sensitivity and timing of ecosystem C uptake to plant available moisture was at fault. Spring biases in gross primary production (GPP) dominate the underestimate of mean annual NEE, whereas models’ lack of GPP response to water availability in both spring and summer monsoon are responsible for inability to capture NEE IAV. Errors in GPP moisture sensitivity at high elevation forested sites were more prominent during the spring, while errors at the low elevation shrub and grass-dominated sites were more important during the monsoon. We propose a range of hypotheses for why model GPP does not respond sufficiently to changing water availability that can serve as a guide for future dryland DGVM developments. Our analysis suggests that improvements in modeling C cycle processes across more than a quarter of the Earth’s land surface could be achieved by addressing the moisture sensitivity of dryland C uptake.

Fluxes from deforestation, changes in land cover, land use and management practices (FLUC for simplicity) contributed to approximately 14 % of anthropogenic CO2 emissions in 2009–2018. Estimating FLUC accurately in space and in time remains, however, challenging, due to multiple sources of uncertainty in the calculation of these fluxes. This uncertainty, in turn, is propagated to global and regional carbon budget estimates, hindering the compilation of a consistent carbon budget and preventing us from constraining other terms, such as the natural land sink. Uncertainties in FLUC estimates arise from many different sources, including differences in model structure (e.g. process based vs. bookkeeping) and model parameterisation. Quantifying the uncertainties from each source requires controlled simulations to separate their effects. Here, we analyse differences between the two bookkeeping models used regularly in the global carbon budget estimates since 2017: the model by Hansis et al. (2015) (BLUE) and that by Houghton and Nassikas (2017) (HN2017). The two models have a very similar structure and philosophy, but differ significantly both with respect to FLUC intensity and spatiotemporal variability. This is due to differences in the land-use forcing but also in the model parameterisation. We find that the larger emissions in BLUE compared to HN2017 are largely due to differences in C densities between natural and managed vegetation or primary and secondary vegetation, and higher allocation of cleared and harvested material to fast turnover pools in BLUE than in HN2017. Besides parameterisation and the use of different forcing, other model assumptions cause differences: in particular that BLUE represents gross transitions which leads to overall higher carbon losses that are also more quickly realised than HN2017.

Year-to-year variability in CO2 fluxes can yield insight into climate-carbon cycle relationships, a fundamental yet uncertain aspect of the terrestrial carbon cycle. In this study, we use global observations from NASA’s Orbiting Carbon Observatory-2 (OCO-2) satellite for years 2015–2019 and a geostatistical inverse model to evaluate 5 years of interannual variability (IAV) in CO2 fluxes and its relationships with environmental drivers. OCO-2 launched in late 2014, and we specifically evaluate IAV during the time period when OCO-2 observations are available. We then compare inferences from OCO-2 with state-of-the-art process-based models (terrestrial biosphere model, TBMs). Results from OCO-2 suggest that the tropical grasslands biome (including grasslands, savanna, and agricultural lands within the tropics) makes contributions to global IAV during the 5 year study period that are comparable to tropical forests, a result that differs from a majority of TBMs. Furthermore, existing studies disagree on the environmental variables that drive IAV during this time period, and the analysis using OCO-2 suggests that both temperature and precipitation make comparable contributions. TBMs, by contrast, tend to estimate larger IAV during this time and usually estimate larger relative contributions from the extra-tropics. With that said, TBMs show little consensus on both the magnitude and the contributions of different regions to IAV. We further find that TBMs show a wide range of responses on the relationships of CO2 fluxes with annual anomalies in temperature and precipitation, and these relationships across most of the TBMs have a larger magnitude than inferred from OCO-2. Overall, the findings of this study highlight large uncertainties in process-based estimates of IAV during recent years and provide an avenue for evaluating these processes against inferences from OCO-2.

Satellite data reveal widespread changes of Earth's vegetation cover. Regions intensively attended to by humans are mostly greening due to land management. Natural vegetation, on the other hand, is exhibiting patterns of both greening and browning in all continents. Factors linked to anthropogenic carbon emissions, such as CO2 fertilization, climate change and consequent disturbances, such as fires and droughts, are hypothesized to be key drivers of changes in natural vegetation. A rigorous regional attribution at biome-level that can be scaled into a global picture of what is behind the observed changes is currently lacking. Here we analyze different datasets of decades-long satellite observations of global leaf area index (LAI, 1981–2017) as well as other proxies of vegetation changes, and identify several clusters of significant long-term changes. Using process-based model simulations (Earth system and land surface models), we disentangle the effects of anthropogenic carbon emissions on LAI in a probabilistic setting applying Causal Counterfactual Theory. The analysis prominently indicates the effects of climate change on many biomes – warming in northern ecosystems (greening) and rainfall anomalies in tropical biomes (browning). Our results do not support previously published accounts of dominant global-scale effects of CO2 fertilization. Altogether, our analysis reveals a slowing down of greening and strengthening of browning trends, particularly in the last two decades. Most models substantially underestimate the emerging vegetation browning, especially in the tropical rainforests. Leaf area loss in these productive ecosystems could be an early indicator of a slow-down in the terrestrial carbon sink. Models need to account for this effect to realize plausible climate projections of the 21st century.

The carbon flux due to land-use and land-cover change (net LULCC flux) historically contributed to a large fraction of anthropogenic carbon emissions while at the same time being associated with large uncertainties. This study aims to compare the contribution of several sensitivities underlying the net LULCC flux by assessing their relative importance in a bookkeeping model (BLUE) based on a LULCC dataset including uncertainty estimates (the LUH2 dataset). The sensitivity experiments build upon the approach of Hurtt et al. (2011) and compare the impacts of LULCC uncertainty (a high, baseline and low land- use estimate), the starting time of the bookkeeping model simulation (850, 1700 and 1850), net area transitions versus gross area transitions (shifting cultivation) and neglecting wood harvest on estimates of the net LULCC flux. Additional factorial experiments isolate the impact of uncertainty from initial conditions and transitions on the net LULCC flux. Finally, historical simulations are extended with future land-use scenarios to assess the impact of past LULCC uncertainty in future projections. Over the period 1850–2014, baseline and low LULCC scenarios produce a comparable cumulative net LULCC flux while the high LULCC estimate initially produces a larger net LULCC flux which decreases towards the end of the period and even becomes smaller than in the baseline estimate. LULCC uncertainty leads to slightly higher sensitivity in the cumulative net LULCC flux (up to 22 %, reference are the baseline simulations) compared to the starting year of a model simulation (up to 15 %). The contribution from neglecting wood harvest activities (up to 28 % cumulative net LULCC flux) is larger than from LULCC uncertainty and the implementation of land-cover transitions (gross or net transitions) exhibits the smallest sensitivity (up to 13 %). At the end of the historical LULCC dataset in 2014, the LULCC uncertainty retains some impact on the net LULCC flux (±0.15 PgC yr−1 at an estimate of 1.7 PgC yr−1). Of the past uncertainties in LULCC, a small impact persists in 2099, mainly due to uncertainty of harvest remaining in 2014. However, compared to the uncertainty range of the LULCC flux estimated today, the estimates in 2099 appear to be indistinguishable. These results, albeit from a single model, are important for CMIP6 as they compare the relative importance of starting year, uncertainty of LULCC, applying gross transitions and wood harvest on the net LULCC flux. For the cumulative net LULCC flux over the industrial period the uncertainty of LULCC is as relevant as applying wood harvest and gross transitions. However, LULCC uncertainty matters less (by about a factor three) than the other two factors for the net LULCC flux in 2014 and historical LULCC uncertainty is negligible for estimates of future scenarios.

Fluxes from deforestation, changes in land-cover, land-use and management practices (FLUC for simplicity) contributed to circa 14 % of anthropogenic CO2 emissions in 2009–2018. Estimating FLUC accurately in space and in time remains, however, challenging, due to multiple sources of uncertainty in the calculation of these fluxes. This uncertainty, in turn, is propagated to global and regional carbon budget estimates, hindering the compilation of a consistent carbon budget and preventing us from constraining other terms, such as the natural land sink. Uncertainties in FLUC estimates arise from many different sources, including differences in model structure (e.g., process- based vs. bookkeeping) and model parameterization. Quantifying the uncertainties from each source requires controlled simulations to separate their effects. Here we analyze differences between the two bookkeeping models used regularly in the global carbon budget estimates since 2017: the model by Hansis et al. (Hansis et al., 2015) (BLUE) and that by Houghton and Nassikas (Houghton and Nassikas, 2017) (HN2017). The two models have a very similar structure and philosophy, but differ significantly both with respect to FLUC intensity and spatio-temporal variability. This is due to differences in the land-use forcing, but also in the model parameterization. We find that the larger emissions in BLUE compared to HN2017 are largely due to differences in C densities between natural and managed vegetation or primary and secondary vegetation, and higher allocation of cleared and harvested material to fast turnover pools in BLUE than in HN2017. Beside parameterization and the use of different forcing, other model assumptions cause differences, in particular that BLUE represents gross transitions which leads to overall higher carbon losses that are also more quickly realized than HN2017.

The carbon flux due to land-use and land-cover change (net LULCC flux) historically contributed to a large fraction of anthropogenic carbon emissions while at the same time being associated with large uncertainties. This study aims to compare the contribution of several sensitivities underlying the net LULCC flux by assessing their relative importance in a bookkeeping model (Bookkeeping of Land Use Emissions, BLUE) based on a LULCC dataset including uncertainty estimates (the Land-Use Harmonization 2 (LUH2) dataset). The sensitivity experiments build upon the approach of Hurtt et al. (2011) and compare the impacts of LULCC uncertainty (a high, baseline and low land-use estimate), the starting time of the bookkeeping model simulation (850, 1700 and 1850), net area transitions versus gross area transitions (shifting cultivation) and neglecting wood harvest on estimates of the net LULCC flux. Additional factorial experiments isolate the impact of uncertainty from initial conditions and transitions on the net LULCC flux. Finally, historical simulations are extended with future land-use scenarios to assess the impact of past LULCC uncertainty in future projections. Over the period 1850–2014, baseline and low LULCC scenarios produce a comparable cumulative net LULCC flux, while the high LULCC estimate initially produces a larger net LULCC flux which decreases towards the end of the period and even becomes smaller than in the baseline estimate. LULCC uncertainty leads to slightly higher sensitivity in the cumulative net LULCC flux (up to 22 %; references are the baseline simulations) compared to the starting year of a model simulation (up to 15 %). The contribution from neglecting wood harvest activities (up to 28 % cumulative net LULCC flux) is larger than that from LULCC uncertainty, and the implementation of land-cover transitions (gross or net transitions) exhibits the smallest sensitivity (up to 13 %). At the end of the historical LULCC dataset in 2014, the LULCC uncertainty retains some impact on the net LULCC flux (±0.15 PgC yr−1 at an estimate of 1.7 PgC yr−1). Of the past uncertainties in LULCC, a small impact persists in 2099, mainly due to uncertainty of harvest remaining in 2014. However, compared to the uncertainty range of the LULCC flux estimated today, the estimates in 2099 appear to be indistinguishable. These results, albeit from a single model, are important for CMIP6 as they compare the relative importance of starting year, uncertainty of LULCC, applying gross transitions and wood harvest on the net LULCC flux. For the cumulative net LULCC flux over the industrial period, the uncertainty of LULCC is as relevant as applying wood harvest and gross transitions. However, LULCC uncertainty matters less (by about a factor of 3) than the other two factors for the net LULCC flux in 2014, and historical LULCC uncertainty is negligible for estimates of future scenarios.

Quantifying the anthropogenic fluxes of CO2 is important to understand the evolution of carbon sink capacities, on which the required strength of our mitigation efforts directly depends. For the historical period, the global carbon budget (GCB) can be compiled from observations and model simulations as is done annually in the Global Carbon Project's (GCP) carbon budgets. However, the historical budget only considers a single realization of the Earth system and cannot account for internal climate variability. Understanding the distribution of internal climate variability is critical for predicting the future carbon budget terms and uncertainties. We present here a decomposition of the GCB for the historical period and the RCP4.5 scenario using single‐model large ensemble simulations from the Max Planck Institute Grand Ensemble (MPI‐GE) to capture internal variability. We calculate uncertainty ranges for the natural sinks and anthropogenic emissions that arise from internal climate variability, and by using this distribution, we investigate the likelihood of historical fluxes with respect to plausible climate states. Our results show these likelihoods have substantial fluctuations due to internal variability, which are partially related to El Niño‐Southern Oscillation (ENSO). We find that the largest internal variability in the MPI‐GE stems from the natural land sink and its increasing carbon stocks over time. The allowable fossil fuel emissions consistent with 3 C warming may be between 9 and 18 Pg C yr−1. The MPI‐GE is generally consistent with GCP's global budgets with the notable exception of land‐use change emissions in recent decades, highlighting that human action is inconsistent with climate mitigation goals., Key Points: We use a single‐model large ensemble to estimate uncertainties from internal climate variability in the global carbon budget. The land sink accounts for most internal climate uncertainty which may permit 9–18 Pg C yr−1 in allowable emissions by 2050 (for 3°C warming)., European Union's Horizon 2020

Variability in climate exerts a strong influence on vegetation productivity (gross primary productivity; GPP), and therefore has a large impact on the land carbon sink. However, no direct observations of global GPP exist, and estimates rely on models that are constrained by observations at various spatial and temporal scales. Here, we assess the consistency in GPP from global products which extend for more than three decades; two observation‐based approaches, the upscaling of FLUXNET site observations (FLUXCOM) and a remote sensing derived light use efficiency model (RS‐LUE), and from a suite of terrestrial biosphere models (TRENDYv6). At local scales, we find high correlations in annual GPP among the products, with exceptions in tropical and high northern latitudes. On longer time scales, the products agree on the direction of trends over 58% of the land, with large increases across northern latitudes driven by warming trends. Further, tropical regions exhibit the largest interannual variability in GPP, with both rainforests and savannas contributing substantially. Variability in savanna GPP is likely predominantly driven by water availability, although temperature could play a role via soil moisture‐atmosphere feedbacks. There is, however, no consensus on the magnitude and driver of variability of tropical forests, which suggest uncertainties in process representations and underlying observations remain. These results emphasize the need for more direct long‐term observations of GPP along with an extension of in situ networks in underrepresented regions (e.g., tropical forests). Such capabilities would support efforts to better validate relevant processes in models, to more accurately estimate GPP., Key Points Changes in climate over the period 1982–2016 have increased GPP at a global scale and across northern latitudesSavannas and tropical forests are hotspots for IAV in GPP, although the dominant climate drivers are not consistent among the assessed productsDGVMs systematically underestimate the IAV of GPP in tropical forests, highlighting the need for improved parameterizations/formulations

In many regions of the world, frequent and continual dry spells are exacerbating drought conditions, which have severe impacts on vegetation biomes. Vegetation in southern Africa is among the most affected by drought. Here, we assessed the spatiotemporal characteristics of meteorological drought in southern Africa using the Standardized Precipitation Evapotranspiration Index over a 30-year period (1982–2011). The severity and the effects of droughts on vegetation productiveness were examined at different drought time-scales (1- to 24-month time-scales). In this study, we characterized vegetation using the Leaf Area Index, after evaluating its relationship with the Normalized Difference Vegetation Index. We found that the LAI responds strongly (r = 0.6) to drought over the central and south eastern parts of the region, with weaker impacts (r  We also examined how well an ensemble of state of the art dynamic global vegetation models (DGVMs) simulate the LAI and its response to drought. The spatial and seasonal response of the LAI to drought is mostly overestimated in the DGVM multi-model ensemble compared to the observations. The correlation coefficient values for the multi-model ensemble are as high as 0.76 (annual) over South Africa, and 0.98 in MAM season over the temperate grassland biome. Furthermore, the DGVM model ensemble shows positive biases (3-month or longer) in the simulation of spatial distribution of drought timescales and overestimate the seasonal distribution timescales. The results of this study may highlight areas to target for further development of DGVMs in order to improve the models’ capability in simulating the drought–vegetation relationship.

Plant phenology plays a fundamental role in land-atmosphere interactions, and its variability and variations are an indicator of climate and environmental changes. For this reason, current land surface models include phenology parameterizations and related biophysical and biogeochemical processes. In this work, the climatology of beginning and end of the growing season, simulated by seven state-of-the-art European land surface models, is evaluated globally against satellite observations. The assessment is performed using the vegetation metric leaf area index and a recently-developed approach, named four growing season types. On average, the land surface models show a 0.6-month delay in the growing season start, while they are about 0.5 months earlier in the growing season end. Difference with observation tends to be higher in the Southern Hemisphere compared to the Northern Hemisphere. High agreement between land surface models and observations is exhibited in areas dominated by broad-leaf deciduous trees, while high variability is noted in regions dominated by broad-leaf deciduous shrubs. Generally, the timing of the growing season end is accurately simulated in about 25 % of global land grid points versus 16 % in the timing of growing season start. The refinement of phenology parameterization can lead to better representation of vegetation-related energy, water, and carbon cycles in land surface models, but plant phenology is also affected by plant physiology and soil hydrology processes. Consequently, phenology representation and, in general, vegetation modelling is a complex task, which still needs further improvement, evaluation, and multi-model comparison.

Emission of greenhouse gases (GHGs) and removals from land, including both anthropogenic and natural fluxes, require reliable quantification, including estimates of uncertainties, to support credible mitigation action under the Paris Agreement. This study provides a state-of-the-art scientific overview of bottom-up anthropogenic emissions data from agriculture, forestry and other land use (AFOLU) in the European Union (EU281). The data integrate recent AFOLU emission inventories with ecosystem data and land carbon models and summarize GHG emissions and removals over the period 1990-2016. This compilation of bottom-up estimates of the AFOLU GHG emissions of European national greenhouse gas inventories (NGHGIs), with those of land carbon models and observation-based estimates of large-scale GHG fluxes, aims at improving the overall estimates of the GHG balance in Europe with respect to land GHG emissions and removals. Whenever available, we present uncertainties, its propagation and role in the comparison of different estimates. While NGHGI data for the EU28 provide consistent quantification of uncertainty following the established IPCC Guidelines, uncertainty in the estimates produced with other methods needs to account for both within model uncertainty and the spread from different model results. The largest inconsistencies between EU28 estimates are mainly due to different sources of data related to human activity, referred to here as activity data (AD) and methodologies (tiers) used for calculating emissions and removals from AFOLU sectors. The referenced datasets related to figures are visualized at https://doi.org/10.5281/zenodo.3662371 (Petrescu et al., 2020).

Forests are the main source of biomass production from solar energy and take up around 2.4±0.4 PgC per year globally. Future changes in climate may affect forest growth and productivity. Currently, state-of-the-art Earth system models use prescribed wood harvest rates in future climate projections. These rates are defined by integrated assessment models (IAMs), only accounting for regional wood demand and largely ignoring the supply side from forests. Therefore, we assess how global growth and harvest potentials of forests change when they are allowed to respond to changes in environmental conditions. For this, we simulate wood harvest rates oriented towards the actual rate of forest growth. Applying this growth-based harvest rule (GB) in JSBACH, the land component of the Max Planck Institute's Earth system model, forced by several future climate scenarios, we realized a growth potential 2 to 4 times (3–9 PgC yr−1) the harvest rates prescribed by IAMs (1–3 PgC yr−1). Limiting GB to managed forest areas (MF), we simulated a harvest potential of 3–7 PgC yr−1, 2 to 3 times higher than IAMs. This highlights the need to account for the dependence of forest growth on climate. To account for the long-term effects of wood harvest as integrated in IAMs, we added a life cycle analysis, showing that the higher supply with MF as an adaptive forest harvesting rule may improve the net mitigation effects of forest harvest during the 21st century by sequestering carbon in anthropogenic wood products.

Changes in vegetation structure are expected to influence the redistribution of heat and moisture; however, how variations in the leaf area index (LAI) affect this global energy partitioning is not yet quantified. Here, we estimate that a unit change in LAI leads to 3.66 ± 0.45 and −3.26 ± 0.41 W m−2 in latent (LE) and sensible (H) fluxes, respectively, over the 1982–2016 period. Analysis of an ensemble of data-driven products shows that these sensitivities increase by about 20% over the observational period, prominently in regions with a limited water supply, probably because of an increased transpiration/evaporation ratio. Global greening has caused a decrease in the Bowen ratio (B = H/LE) of −0.010 ± 0.002 per decade, which is attributable to the increased evaporative surface. Such a direct LAI effect on energy fluxes is largely modulated by plant functional types (PFTs) and background climate conditions. Land surface models (LSMs) misrepresent this vegetation control, possibly due to underestimation of the biophysical responses to changes in the water availability and poor representation of LAI dynamics. Changes in the leaf area index alter the distribution of heat and moisture. The change in energy partitioning related to leaf area, increasing latent and decreasing sensible fluxes over the observational period 1982–2016, is moderated by plant functional type and background climate.

Climate change is shifting the phenological cycles of plants1, thereby altering the functioning of ecosystems, which in turn induces feedbacks to the climate system2. In northern (north of 30° N) ecosystems, warmer springs lead generally to an earlier onset of the growing season3,4 and increased ecosystem productivity early in the season5. In situ6 and regional7–9 studies also provide evidence for lagged effects of spring warmth on plant productivity during the subsequent summer and autumn. However, our current understanding of these lagged effects, including their direction (beneficial or adverse) and geographic distribution, is still very limited. Here we analyse satellite, field-based and modelled data for the period 1982–2011 and show that there are widespread and contrasting lagged productivity responses to spring warmth across northern ecosystems. On the basis of the observational data, we find that roughly 15 per cent of the total study area of about 41 million square kilometres exhibits adverse lagged effects and that roughly 5 per cent of the total study area exhibits beneficial lagged effects. By contrast, current-generation terrestrial carbon-cycle models predict much lower areal fractions of adverse lagged effects (ranging from 1 to 14 per cent) and much higher areal fractions of beneficial lagged effects (ranging from 9 to 54 per cent). We find that elevation and seasonal precipitation patterns largely dictate the geographic pattern and direction of the lagged effects. Inadequate consideration in current models of the effects of the seasonal build-up of water stress on seasonal vegetation growth may therefore be able to explain the differences that we found between our observation-constrained estimates and the model-constrained estimates of lagged effects associated with spring warming. Overall, our results suggest that for many northern ecosystems the benefits of warmer springs on growing-season ecosystem productivity are effectively compensated for by the accumulation of seasonal water deficits, despite the fact that northern ecosystems are thought to be largely temperature- and radiation-limited10.

El Niño has two different flavors, eastern Pacific (EP) and central Pacific (CP) El Niños, with different global teleconnections. However, their different impacts on the interannual carbon cycle variability remain unclear. Here we compared the behaviors of interannual atmospheric CO2 variability and analyzed their terrestrial mechanisms during these two types of El Niños, based on the Mauna Loa (MLO) CO2 growth rate (CGR) and the Dynamic Global Vegetation Model's (DGVM) historical simulations. The composite analysis showed that evolution of the MLO CGR anomaly during EP and CP El Niños had three clear differences: (1) negative or neutral precursors in the boreal spring during an El Niño developing year (denoted as yr0), (2) strong or weak amplitudes, and (3) durations of the peak from December (yr0) to April during an El Niño decaying year (denoted as yr1) compared to October (yr0) to January (yr1) for a CP El Niño, respectively. The global land–atmosphere carbon flux (FTA) simulated by multi-models was able to capture the essentials of these characteristics. We further found that the gross primary productivity (GPP) over the tropics and the extratropical Southern Hemisphere (Trop + SH) generally dominated the global FTA variations during both El Niño types. Regional analysis showed that during EP El Niño events significant anomalous carbon uptake caused by increased precipitation and colder temperatures, corresponding to the negative precursor, occurred between 30° S and 20° N from January (yr0) to June (yr0). The strongest anomalous carbon releases, largely due to the reduced GPP induced by low precipitation and warm temperatures, occurred between the equator and 20° N from February (yr1) to August (yr1). In contrast, during CP El Niño events, clear carbon releases existed between 10° N and 20° S from September (yr0) to September (yr1), resulting from the widespread dry and warm climate conditions. Different spatial patterns of land temperatures and precipitation in different seasons associated with EP and CP El Niños accounted for the evolutionary characteristics of GPP, terrestrial ecosystem respiration (TER), and the resultant FTA. Understanding these different behaviors of interannual atmospheric CO2 variability, along with their terrestrial mechanisms during EP and CP El Niños, is important because the CP El Niño occurrence rate might increase under global warming.

To contribute to a quantitative comparison of climate engineering (CE) methods, we assess atmosphere-, ocean-, and land-based CE measures with respect to Earth system effects consistently within one comprehensive model. We use the Max Planck Institute Earth System Model (MPI-ESM) with prognostic carbon cycle to compare solar radiation management (SRM) by stratospheric sulfur injection and two carbon dioxide removal methods: afforestation and ocean alkalinization. The CE model experiments are designed to offset the effect of fossil-fuel burning on global mean surface air temperature under the RCP8.5 scenario to follow or get closer to the RCP4.5 scenario. Our results show the importance of feedbacks in the CE effects. For example, as a response to SRM the land carbon uptake is enhanced by 92 Gt by the year 2100 compared to the reference RCP8.5 scenario due to reduced soil respiration thus reducing atmospheric CO2. Furthermore, we show that normalizations allow for a better comparability of different CE methods. For example, we find that due to compensating processes such as biogeophysical effects of afforestation more carbon needs to be removed from the atmosphere by afforestation than by alkalinization to reach the same global warming reduction. Overall, we illustrate how different CE methods affect the components of the Earth system; we identify challenges arising in a CE comparison, and thereby contribute to developing a framework for a comparative assessment of CE.

The Amazon forests are one of the largest ecosystem carbon pools on Earth. While more frequent and prolonged droughts have been predicted under future climate change there, the vulnerability of Amazon forests to drought has yet remained largely uncertain, as previous studies have shown that few land surface models succeeded in capturing the vegetation responses to drought. In this study, we present an improved version of the land surface model JSBACH, which incorporates new formulations of leaf phenology and litter production based on intensive field measurement from the artificial drought experiments in the Amazon. Coupling the new JSBACH with the atmospheric model ECHAM, we investigate the drought responses of the Amazon forests and the resulting feedbacks under RCP8.5 scenario. The climatic effects resulted from (1) direct effects including declining soil moisture and stomatal responses, and (2) soil moisture-induced canopy responses are separated to give more insights, as the latter was poorly simulated. Preliminary results show that for net primary production and soil respiration, the direct effects and canopy responses have similar spatial patterns with the magnitude of the latter being 1/5 to 1/3 of the former. In addition, declining soil moisture enhances rainfall in Northern Amazon and suppresses rainfall in the south, while canopy responses have negligible effects on rainfall. Based on our findings, we suggest cautious interpretation of results from previous studies. To address this uncertainty, better strategy in modeling leaf phenology such as implemented in this study should be adopted.