The estimates of global and national fossil CO2 emissions (EFOS) include the oxidation of fossil fuels through both combustion (e.g. transport, heating) and chemical oxidation (e.g. carbon anode decomposition in aluminium refining) activities, and the decomposition of carbonates in industrial processes (e.g. the production of cement). We also include CO2 uptake from the cement carbonation process. Several emissions sources are not estimated or not fully covered: coverage of emissions from lime production are not global, and decomposition of carbonates in glass and ceramic production are included only for the “Annex 1” countries of the United Nations Framework Convention on Climate Change (UNFCCC) due to a lack of activity data. These omissions are considered to be minor. Short-cycle carbon emissions – for example from combustion of biomass – are not included here but are accounted for in the CO2 emissions from land use (see Sect. 2.2).

Our estimates of fossil CO2 emissions are derived using the standard approach of activity data and emission factors, relying on data collection by many other parties. Our goal is to produce the best estimate of this flux, and we therefore use a prioritization framework to combine data from different sources that have used different methods, while being careful to avoid double counting and undercounting of emissions sources. The CDIAC-FF emissions dataset, derived largely from UN energy data, forms the foundation, and we extend emissions to year Y-1 using energy growth rates reported by BP. We then proceed to replace estimates using data from what we consider to be superior sources, for example Annex 1 countries' official submissions to the UNFCCC. All data points are potentially subject to revision, not just the latest year. For full details see Andrew and Peters (2021).

Other estimates of global fossil CO2 emissions exist, and these are compared by Andrew (2020a). The most common reason for differences in estimates of global fossil CO2 emissions is a difference in which emissions sources are included in the datasets. Datasets such as those published by BP energy company, the US Energy Information Administration, and the International Energy Agency's “CO2 emissions from fuel combustion” are all generally limited to emissions from combustion of fossil fuels. In contrast, datasets such as PRIMAP-hist, CEDS, EDGAR, and GCP's dataset aim to include all sources of fossil CO2 emissions. See Andrew (2020a) for detailed comparisons and discussion.

Cement absorbs CO2 from the atmosphere over its lifetime, a process known as “cement carbonation”. We estimate this CO2 sink as the average of two studies in the literature (Cao et al., 2020; Guo et al., 2021). Both studies use the same model, developed by Xi et al. (2016), with different parameterizations and input data. Since carbonation is a function of both current and previous cement production, we extend these estimates by 1 year to 2020 by using the growth rate derived from the smoothed cement emissions (10-year smoothing) fitted to the carbonation data.

We use the Kaya identity for a simple decomposition of CO2 emissions into the key drivers (Raupach et al., 2007). While there are variations (Peters et al., 2017), we focus here on a decomposition of CO2 emissions into population, GDP per person, energy use per GDP, and CO2 emissions per energy use. Multiplying these individual components together returns the CO2 emissions. Using the decomposition, it is possible to attribute the change in CO2 emissions to the change in each of the drivers. This method gives a first-order understanding of what causes CO2 emissions to change each year.

We provide a projection of global CO2 emissions in 2021 by combining separate projections for China, the USA, the EU, India, and for all other countries combined. The methods are different for each of these. For China we combine monthly fossil fuel production data from the National Bureau of Statistics, import and export data from the Customs Administration, and monthly coal consumption estimates from SX Coal (2021), giving us partial data for the growth rates to date of natural gas, petroleum, and cement, and of the consumption itself for raw coal. We then use a regression model to project full-year emissions based on historical observations. For the USA our projection is taken directly from the Energy Information Administration's (EIA) Short-Term Energy Outlook (EIA, 2022), combined with the year-to-date growth rate of cement production. For the EU we use monthly energy data from Eurostat to derive estimates of monthly CO2 emissions through July, with coal emissions extended first through September using a statistical relationship with reported electricity generation from coal and other factors, then through December assuming normal seasonal patterns. EU emissions from natural gas – a strongly seasonal cycle – are extended through December using bias-adjusted Holt–Winters exponential smoothing (Chatfield, 1978). EU emissions from oil are derived using the EIA's projection of oil consumption for Europe. EU cement emissions are based on available year-to-date data from two of the largest producers, Germany and Poland. India's projected emissions are derived from estimates through August (September for coal) using the methods of Andrew (2020b) and extrapolated assuming normal seasonal patterns. Emissions for the rest of the world are derived using projected growth in economic production from the IMF (2022) combined with extrapolated changes in emissions intensity of economic production. More details on the EFOS methodology and its 2021 projection can be found in Appendix C1.

The rate of growth of the atmospheric CO2 concentration is provided for years 1959–2020 by the US National Oceanic and Atmospheric Administration Earth System Research Laboratory (NOAA/ESRL; Dlugokencky and Tans, 2022), which is updated from Ballantyne et al. (2012) and includes recent revisions to the calibration scale of atmospheric CO2 measurements (Hall et al., 2021). For the 1959–1979 period, the global growth rate is based on measurements of atmospheric CO2 concentration averaged from the Mauna Loa and South Pole stations, as observed by the CO2 Program at Scripps Institution of Oceanography (Keeling et al., 1976). For the 1980–2020 time period, the global growth rate is based on the average of multiple stations selected from the marine boundary layer sites with well-mixed background air (Ballantyne et al., 2012), after fitting each station with a smoothed curve as a function of time, and averaging by latitude band (Masarie and Tans, 1995). The annual growth rate is estimated by Dlugokencky and Tans (2022) from atmospheric CO2 concentration by taking the average of the most recent December–January months corrected for the average seasonal cycle and subtracting this same average 1 year earlier. The growth rate in units of ppm yr-1 is converted to units of GtC yr-1 by multiplying by a factor of 2.124 GtC ppm-1, assuming instantaneous mixing of CO2 throughout the atmosphere (Ballantyne et al., 2012).

Starting in 2020, NOAA/ESRL now provides estimates of atmospheric CO2 concentrations with respect to a new calibration scale, referred to as WMO-CO2-X2019, in line with the recommendation of the World Meteorological Organization (WMO) Global Atmosphere Watch (GAW) community (Hall et al., 2021). The WMO-CO2-X2019 scale improves upon the earlier WMO-CO2-X2007 scale by including a broader set of standards, which contain CO2 in a wider range of concentrations that span the range 250–800 ppm (versus 250–520 ppm for WMO-CO2-X2007). In addition, NOAA/ESRL made two minor corrections to the analytical procedure used to quantify CO2 concentrations, fixing an error in the second virial coefficient of CO2 and accounting for loss of a small amount of CO2 to materials in the manometer during the measurement process. The difference in concentrations measured using WMO-CO2-X2019 versus WMO-CO2-X2007 is ∼+0.18 ppm at 400 ppm and the observational record of atmospheric CO2 concentrations have been revised accordingly. The revisions have been applied retrospectively in all cases where the calibrations were performed by NOAA/ESRL, thus affecting measurements made by members of the WMO-GAW programme and other regionally coordinated programmes (e.g. Integrated Carbon Observing System, ICOS). Changes to the CO2 concentrations measured across these networks propagate to the global mean CO2 concentrations. Comparing the estimates of GATM made by Dlugokencky and Tans (2020), used in the Global Carbon Budget 2020 (Friedlingstein et al., 2020), with updated estimates from Dlugokencky and Tans (2022), used here, we find that GATM reduced on average by -0.06 GtC yr-1 during 2010–2019 and by -0.01 GtC yr-1 during 1959–2019 due to the new calibration. These changes are well within the uncertainty ranges reported below. Hence the change in analytical procedures made by NOAA/ESRL has a negligible impact on the atmospheric growth rate GATM.

The uncertainty around the atmospheric growth rate is due to four main factors. First, the long-term reproducibility of reference gas standards (around 0.03 ppm for 1σ from the 1980s; Dlugokencky and Tans, 2022). Second, small unexplained systematic analytical errors that may have a duration of several months to 2 years come and go. They have been simulated by randomizing both the duration and the magnitude (determined from the existing evidence) in a Monte Carlo procedure. Third, the network composition of the marine boundary layer with some sites coming or going, gaps in the time series at each site, etc. (Dlugokencky and Tans, 2022). The latter uncertainty was estimated by NOAA/ESRL with a Monte Carlo method by constructing 100 “alternative” networks (Masarie and Tans, 1995; NOAA/ESRL, 2019). The second and third uncertainties, summed in quadrature, add up to 0.085 ppm on average (Dlugokencky and Tans, 2022). Fourth, the uncertainty associated with using the average CO2 concentration from a surface network to approximate the true atmospheric average CO2 concentration (mass-weighted, in three dimensions) as needed to assess the total atmospheric CO2 burden. In reality, CO2 variations measured at the stations will not exactly track changes in total atmospheric burden, with offsets in magnitude and phasing due to vertical and horizontal mixing. This effect must be very small on decadal and longer timescales, when the atmosphere can be considered well mixed. Preliminary estimates suggest this effect would increase the annual uncertainty, but a full analysis is not yet available. We therefore maintain an uncertainty around the annual growth rate based on the multiple stations' dataset ranges between 0.11 and 0.72 GtC yr-1, with a mean of 0.61 GtC yr-1 for 1959–1979 and 0.17 GtC yr-1 for 1980–2020, when a larger set of stations were available as provided by Dlugokencky and Tans (2022), but recognize further exploration of this uncertainty is required. At this time, we estimate the uncertainty of the decadal averaged growth rate after 1980 at 0.02 GtC yr-1 based on the calibration and the annual growth rate uncertainty but stretched over a 10-year interval. For years prior to 1980, we estimate the decadal averaged uncertainty to be 0.07 GtC yr-1 based on a factor proportional to the annual uncertainty prior and after 1980 (0.02×[0.61/0.17] GtC yr-1).

We assign a high confidence to the annual estimates of GATM because they are based on direct measurements from multiple and consistent instruments and stations distributed around the world (Ballantyne et al., 2012; Hall et al., 2021).

To estimate the total carbon accumulated in the atmosphere since 1750 or 1850, we use an atmospheric CO2 concentration of 277 ± 3 ppm or 286 ± 3 ppm, respectively, based on a cubic spline fit to ice core data (Joos and Spahni, 2008). For the construction of the cumulative budget shown in Fig. 3, we use the fitted estimates of CO2 concentration from Joos and Spahni (2008) to estimate the annual atmospheric growth rate using the conversion factors shown in Table 1. The uncertainty of ±3 ppm (converted to ±1σ) is taken directly from the IPCC's AR5 assessment (Ciais et al., 2013). Typical uncertainties in the growth rate in atmospheric CO2 concentration from ice core data are equivalent to ±0.1–0.15 GtC yr-1 as evaluated from the Law Dome data (Etheridge et al., 1996) for individual 20-year intervals over the period from 1850 to 1960 (Bruno and Joos, 1997).

We provide an assessment of GATM for 2021 based on the monthly calculated global atmospheric CO2 concentration (GLO) through August (Dlugokencky and Tans, 2022), and bias-adjusted Holt–Winters exponential smoothing with additive seasonality (Chatfield, 1978) to project to January 2022. Additional analysis suggests that the first half of the year (the boreal winter–spring–summer transition) shows more interannual variability than the second half of the year (the boreal summer–autumn–winter transition), so that the exact projection method applied to the second half of the year has a relatively smaller impact on the projection of the full year. Uncertainty is estimated from past variability using the standard deviation of the last 5 years' monthly growth rates.

Cumulative fossil CO2 emissions for 1850–2020 were 455 ± 25 GtC, including the cement carbonation sink (Fig. 3, Table 8).

In this period, 46 % of fossil CO2 emissions came from coal, 35 % from oil, 14 % from natural gas, 3 % from decomposition of carbonates, and 1 % from flaring.

In 1850, the UK accounted for 62 % of global fossil CO2 emissions. In 1891 the combined cumulative emissions of the current members of the European Union reached and subsequently surpassed the level of the UK. Since 1917 US cumulative emissions have been the largest. Over the entire period 1850–2020, US cumulative emissions amounted to 110 GtC (25 % of world total), the EU's to 80 GtC (18 %), and China's to 60 GtC (14 %).

There are three additional global datasets that include all sources of fossil CO2 emissions: CDIAC-FF (Gilfillan and Marland, 2021), CEDS version v_2021_04_21 (Hoesly et al., 2018; O'Rourke et al., 2021), and PRIMAP-hist version 2.3.1 (Gütschow et al., 2016, 2021), although these datasets are not independent. CDIAC-FF has the lowest cumulative emissions over 1750–2018 at 437 GtC, GCP has 443 GtC, CEDS 445 GtC, PRIMAP-hist TP 453 GtC, and PRIMAP-hist CR 455 GtC. CDIAC-FF excludes emissions from lime production, while neither CDIAC-FF nor GCP explicitly include emissions from international bunker fuels prior to 1950. CEDS has higher emissions from international shipping in recent years, while PRIMAP-hist has higher fugitive emissions than the other datasets. However, in general these four datasets are in relative agreement with total historical global emissions of fossil CO2.

Global fossil CO2 emissions, EFOS (including the cement carbonation sink), have increased every decade from an average of 3.0 ± 0.2 GtC yr-1 for the decade of the 1960s to an average of 9.5 ± 0.5 GtC yr-1 during 2011–2020 (Table 6, Figs. 2, 4 and 5). The growth rate in these emissions decreased between the 1960s and the 1990s, from 4.3 % yr-1 in the 1960s (1960–1969), 3.2 % yr-1 in the 1970s (1970–1979), and 1.6 % yr-1 in the 1980s (1980–1989), to 0.9 % yr-1 in the 1990s (1990–1999). After this period, the growth rate began increasing again in the 2000s at an average growth rate of 3.0 % yr-1, decreasing to 0.6 % yr-1 for the last decade (2011–2020). China's emissions increased by +1.0 % yr-1 on average over the last 10 years, dominating the global trend, followed by India's emissions increase by +3.9 % yr-1, while emissions decreased in EU27 by -1.9 % yr-1, and in the USA by -1.1 % yr-1. Figure 6 illustrates the spatial distribution of fossil fuel emissions for the 2011–2020 period.

EFOS includes the uptake of CO2 by cement via carbonation which has increased with increasing stocks of cement products, from an average of 20 MtC yr-1 (0.02 GtC yr-1) in the 1960s to an average of 200 MtC yr-1 (0.2 GtC yr-1) during 2011–2020 (Fig. 5).

Global fossil CO2 emissions were 5.4 % lower in 2020 than in 2019, because of the COVID-19 pandemic, with a decline of 0.5 GtC to reach 9.5 ± 0.5 GtC (9.3 ± 0.5 GtC when including the cement carbonation sink) in 2020 (Fig. 5), distributed among coal (40 %), oil (32 %), natural gas (21 %), cement (5 %), and others (2 %). Compared to the previous year, 2020 emissions from coal, oil, and gas declined by 4.4 %, 9.7 %, and 2.3 % respectively, while emissions from cement increased by 0.8 %. All growth rates presented are adjusted for the leap year, unless stated otherwise.

In 2020, the largest absolute contributions to global fossil CO2 emissions were from China (31 %), the USA (14 %), the EU27 (7 %), and India (7 %). These four regions account for 59 % of global CO2 emissions, while the rest of the world contributed 41 %, including international aviation and marine bunker fuels (2.9 % of the total). Growth rates for these countries from 2019 to 2020 were +1.4 % (China), -10.6 % (USA), -10.9 % (EU27), and -7.3 % (India), with -7.0 % for the rest of the world. The per capita fossil CO2 emissions in 2020 were 1.2 tC person-1 yr-1 for the globe, and were 3.9 (USA), 2.0 (China), 1.6 (EU27) and 0.5 (India) tC per person per year for the four highest emitting countries (Fig. 5).

The COVID-19-induced decline in emissions of -5.4 % in 2020 is close to the projected decline of -6.7 %, which was the median of four approaches, published in Friedlingstein et al. (2020) (Table 7). Of the four approaches, the “GCP” method was closest at -5.8 %. That method was based on national emissions projections for China, the USA, the EU27, and India using reported monthly activity data when available and projections of gross domestic product corrected for trends in fossil fuel intensity (IFOS) for the rest of the world. Of the regions, the projection for the EU27 was the least accurate, and the reasons for this are discussed by Andrew (2021).

Globally, we estimate that global fossil CO2 emissions will rebound 4.8 % in 2021 (4.2 % to 5.4 %) to 9.9 GtC (36.4 GtCO2), returning to near their 2019 emission levels of 10.0 GtC (36.7 GtCO2). Global increases in 2021 emissions per fuel types are +6.3 % (range 5.5 % to 7.0 %) for coal, +4.0 % (range 2.6 % to 5.4 %) for oil, +3.8 % (range 2.8 % to 4.8 %) for natural gas, and +3.2 % (range 1.7 % to 4.6 %) for cement.

For China, projected fossil emissions in 2021 are expected to increase by 4.3 % (range 3.0 % to 5.4 %) compared with 2020 emissions, bringing 2021 emissions for China to around 3.0 GtC yr-1 (11.1 GtCO2 yr-1). Chinese emissions appear to have risen in both 2020 and 2021 despite the economic disruptions of COVID-19. Increases in fuel-specific projections for China are +4.1 % for coal, +4.4 % for oil, +12.8 % natural gas, and a decrease of 0.1 % for cement.

For the USA, the Energy Information Administration (EIA) emissions projection for 2021 combined with cement clinker data from USGS gives an increase of 6.8 % (range 6.6 % to 7.0 %) compared to 2020, bringing USA 2021 emissions to around 1.4 GtC yr-1 (5.0 GtCO2 yr-1). This is based on separate projections for coal of +17.1 %, oil +9.0 %, natural gas -0.8 %, and cement +0.3 %.

For the European Union, our projection for 2021 is for an increase of 6.3 % (range 4.3 % to 8.3 %) over 2020, with 2021 emissions around 0.8 GtC yr-1 (2.8 GtCO2 yr-1). This is based on separate projections for coal of +14.6 %, oil +3.7 %, natural gas +4.6 %, and cement +0.3 %.

For India, our projection for 2021 is an increase of 11.2 % (range of 10.7 % to 11.7 %) over 2020, with 2021 emissions around 0.7 GtC yr-1 (2.7 GtCO2 yr-1). This is based on separate projections for coal of +13.9 %, oil +3.4 %, natural gas +4.8 %, and cement +21.6 %.

For the rest of the world, the expected growth rate for 2021 is 3.2 % (range 2.0 % to 4.3 %). This is computed using the GDP projection for the world (excluding China, the USA, the EU, and India) of 4.4 % made by the IMF (2022) and a decrease in IFOS of -1.7 % yr-1, which is the average over 2011–2020. The uncertainty range is based on the standard deviation of the interannual variability in IFOS during 2011–2020 of 0.6 % yr-1 and our estimates of uncertainty in the IMF's GDP forecast of 0.6 %. The methodology allows independent projections for coal, oil, natural gas, cement, and other components, which add to the total emissions in the rest of the world. The fuel-specific projected 2021 growth rates for the rest of the world are: +3.2 % (range 0.7 % to 5.8 %) for coal, +2.3 % (-0.3 % to +4.9 %) for oil, +4.1 % (2.6 % to 5.7 %) for natural gas, and +4.8 % (+2.7 % to +6.9 %) for cement.

Independently, the IEA has published two forecasts of global fossil energy CO2 emissions (i.e. a subset of fossil CO2 emissions), the first in April (4.8 %; IEA, 2021a) which was then revised in October to 4 % (IEA, 2021b). In March 2022 they also published a new, preliminary estimate of 6 % growth (IEA, 2021a). Carbon Monitor produces estimates of global emissions with low temporal lag, and their estimates suggest that emissions were 5.1 % higher than in 2020 (Carbon Monitor, 2022).

Cumulative CO2 emissions from land-use changes (ELUC) for 1850–2020 were 200 ± 65 GtC (Table 8; Figs. 3, 13). The cumulative emissions from ELUC are particularly uncertain, with large spread among individual estimates of 140 GtC (updated H&N2017), 270 GtC (BLUE), and 195 GtC (OSCAR) for the three bookkeeping models and a similar wide estimate of 190 ± 60 GtC for the DGVMs (all cumulative numbers are rounded to the nearest 5 GtC). These estimates are broadly consistent with indirect constraints from vegetation biomass observations, giving a cumulative source of 155 ± 50 GtC over the 1901–2012 period (Li et al., 2017). However, given the large spread, a best estimate is difficult to ascertain.

In contrast to growing fossil emissions, CO2 emissions from land use, land-use change, and forestry have remained relatively constant, at around 1.3 ± 0.7 GtC yr-1 over the 1970–1999 period, and even show a slight decrease over the last 20 years, reaching 1.1 ± 0.7 GtC yr-1 for the 2011–2020 period (Table 6, Fig. 4), but with large spread across estimates (Table 5, Fig. 7). Emissions have been relatively constant in the DGVMs ensemble of models since the 1970s, with similar mean values until the 1990s as the bookkeeping mean and large model spread (Table 5, Fig. 7). The DGVMs average grows larger than the bookkeeping average in the recent decades and shows no sign of decreasing emissions, which is, however, expected as DGVM-based estimates include the loss of additional sink capacity, which grows with time, while the bookkeeping estimates do not (Appendix D4).

ELUC is a net term of various gross fluxes, which comprise emissions and removals. Gross emissions are on average 2–4 times larger than the net ELUC emissions, and remained largely constant over the last 60 years, with a moderate increase from an average of 3.4 ± 0.9 GtC yr-1 for the decade of the 1960s to an average of 3.8 ± 0.6 GtC yr-1 during 2011–2020 (Fig. 7, Table 5), showing the relevance of land management such as harvesting or rotational agriculture. Increases in gross removals, from 1.9 ± 0.4 GtC yr-1 for the 1960s to 2.7 ± 0.4 GtC yr-1 for 2011–2020, were larger than the increase in gross emissions. Since the processes behind gross removals, foremost forest regrowth and soil recovery, are all slow, while gross emissions include a large instantaneous component, short-term changes in land-use dynamics, such as a temporary decrease in deforestation, influence gross emissions dynamics more than gross removal dynamics. It is these relative changes to each other that explain the decrease in net ELUC emissions over the last two decades and the last few years. Gross fluxes differ more across the three bookkeeping estimates than net fluxes, which is expected due to different process representation; in particular, treatment of shifting cultivation, which increases both gross emissions and removals, differs across models.

There is a decrease in net CO2 emissions from land-use change over the last decade (Fig. 7, Table 6), in contrast to earlier estimates of no clear trend across ELUC estimates (Friedlingstein et al., 2020; Hong et al., 2021). The trend in the last decade is now about -4 % yr-1, compared to the +1.8 % yr-1 reported by Friedlingstein et al. (2020). This decrease is principally attributable to changes in ELUC estimates from BLUE and OSCAR, which relate to changes in the underlying land-use forcing, LUH2 (Chini et al., 2021; Hurtt et al., 2020), based on HYDE3.3 (Klein Goldewijk et al., 2017a, b): HYDE3.3 now incorporates updated estimates of agricultural areas by the FAO and uses multi-annual land-cover maps from satellite remote sensing (ESA CCI Land Cover) to constrain contemporary land-cover patterns (see Appendix C2.2 for details). These changes lead to lower global ELUC estimates in the last two decades compared to earlier versions of the global carbon budget due most notably to lower emissions from cropland expansion, particularly in the tropical regions. Rosan et al. (2021) showed that for Brazil, the new HYDE3.3 version is closer to independent, regional estimates of land-use and land-cover change (MapBiomas, 2021) with respect to spatial patterns, but it shows less land-use and land-cover changes than these independent estimates, while HYDE3.2-based estimates had shown higher changes and lower emissions. The update in land-use forcing leads to a decrease in estimated emissions in Brazil across several models after the documented deforestation peak of 2003–2004 that preceded policies and monitoring systems decreasing deforestation rates (Rosan et al., 2021). However, estimated emissions based on the new land-use forcing do not reflect the rise in Brazilian deforestation in the last few years (Silva Junior et al., 2021), and associated increasing emissions from deforestation would have been missed here. The update in FAO agricultural areas in Brazil also implied that substantial interannual variability reported to earlier FAO assessment and captured by the HYDE3.2 version since 2000 was removed. Due to the asymmetry of (fast) decay (like clearing by fire) and (slower) regrowth, such reduced variability is expected to decrease annual emissions. Also, the approach by Houghton and Nassikas (2017) smooths land-use area changes before calculating carbon fluxes by a 5-year running mean, hence the three emission estimates are in better agreement than in previous GCB estimates. However, differences still exist, which highlight the need for accurate knowledge of land-use transitions and their spatial and temporal variability. A further caveat is that global land-use change data for model input does not capture forest degradation, which often occurs on small scales or without forest cover changes easily detectable from remote sensing and poses a growing threat to forest area and carbon stocks that may surpass deforestation effects (e.g. Matricardi et al., 2020; Qin et al., 2021).

Overall, therefore, we assign low confidence to the change towards a decreasing trend of land-use emissions over the last two decades as seen compared to the estimate of the global carbon budget 2020 (Friedlingstein et al., 2020). Our approach aims at using the most up-to-date data and methods, such as accounting for revisions of living databases of country-level agricultural statistics from FAO or including satellite remote-sensing information for spatial allocation. While we start from a well-documented methodology to provide gridded land-use data (Chini et al., 2021), not all changes in individual components are always documented, complicating the explanation of changes from one GCB to the next. The rising number of pan-tropical or global estimates of carbon stock changes based on satellite remote sensing of carbon densities and forest cover changes (Fan et al., 2019; Qin et al., 2021; Xu et al., 2022; Feng et al., 2022) may seem a promising path for independent evaluation of the land-use emissions term. However, comparison of satellite-derived fluxes to global model estimates is hampered for several reasons discussed by Pongratz et al. (2021). Most importantly, satellite-based estimates usually do not distinguish between anthropogenic drivers and natural forest cover losses (e.g. from drought or natural wildfires), which have also increased over time in some regions, including the tropics; ancillary information would be needed to attribute the observed signal of vegetation or carbon stock change to different drivers. Further, satellite-based estimates often only provide sub-component fluxes of ELUC, excluding soil or product pool changes. Since forest cover loss is better detectable from space than regrowth, satellite-based products often limit their estimates to emissions from forest loss, neglecting carbon uptake from regrowth of forests, as may occur following wood harvesting, abandonment, or natural disturbances; such products thus provide a subset of the gross emissions term (Fig. 7b) and cannot be compared to net emissions. Lastly, satellite-based fluxes typically quantify committed instead of actual emissions, i.e. legacy CO2 fluxes from potentially slow processes such as slash, soil carbon or product decay, or forest regrowth are not captured at the time they actually occur but are attributed to the time of the land-use change event (Pongratz et al., 2021). Using data on drivers of forest cover loss to isolate fluxes from agricultural expansion, and looking into gross emissions instead of the net land-use change flux, Feng et al. (2022) suggest a stronger increase in global gross emissions (though generally a smaller flux) than the bookkeeping models do (see gross fluxes in Fig. 7b). This is in line with Rosan et al. (2021) suggesting that the trend of net emissions in Brazil may be underestimated by the updated land-use data (though patterns have improved). Further studies are needed to robustly estimate the trend of global net land-use emissions. Progress is also needed on accurate quantifications of land-use dynamics, including less well observable management types such as shifting cultivation and wood harvesting, and their distinction from natural disturbances (Pongratz et al., 2021).

The highest land-use emissions occur in the tropical regions of all three continents, including the Arc of Deforestation in the Amazon basin (Fig. 6b). This is related to massive expansion of cropland, particularly in the last few decades in Latin America, Southeast Asia, and sub-Saharan Africa (Hong et al., 2021), to a substantial extent for export (Pendrill et al., 2019). Emission intensity is high in many tropical countries, particularly of Southeast Asia, due to high rates of land conversion in regions of carbon-dense and often still pristine, undegraded natural forests (Hong et al., 2021). Emissions are further increased by peat fires in equatorial Asia (GFED4s, van der Werf et al., 2017). Uptake due to land-use change occurs, particularly in Europe, partly related to expanding forest area as a consequence of the forest transition in the 19th and 20th century and subsequent regrowth of forest (Fig. 6b) (Mather, 2001; McGrath et al., 2015).

While the mentioned patterns are robust and supported by independent literature, we acknowledge that model spread is substantially larger on regional than on global level, as has been shown for bookkeeping models (Bastos et al., 2021) as well as DGVMs (Obermeier et al., 2021). A detailed analysis of country-level or regional uncertainties globally is beyond the scope of this study. Assessments for individual regions will be performed as part of REgional Carbon Cycle Assessment and Processes (RECCAP2; Ciais et al., 2022) or already exist for selected regions (e.g. for Europe Petrescu et al., 2020, for Brazil Rosan et al., 2021).

National GHG inventory data (NGHGI) under the LULUCF sector or data submitted by countries to FAOSTAT differ from the global models' definition of ELUC we adopt here in that in the NGHGI reporting, the natural fluxes (SLAND) are counted towards ELUC when they occur on managed land (Grassi et al., 2018). In order to compare our results to the NGHGI approach, we perform a re-mapping of our ELUC estimate by including the SLAND over managed forest from the DGVM simulations (following Grassi et al., 2021) to the bookkeeping ELUC estimate (see Appendix C2.3). For the 2011–2020 period, we estimate that 1.5 GtC yr-1 of SLAND occurred on managed forests and is then reallocated to ELUC here, as done in the NGHGI method. Doing so, our mean estimate of ELUC is reduced from a source of 1.1 GtC to a sink of -0.4 GtC, very similar to the NGHGI estimate of -0.6 GtC (Table A8).

Though estimates between GHGI, FAOSTAT, individual process-based models, and the mapped budget estimates still differ in value and need further analysis, the approach taken here provides a possibility to relate the global models' and NGHGI approach to each other routinely and thus link the anthropogenic carbon budget estimates of land CO2 fluxes directly to the Global Stocktake, as part of the UNFCCC Paris Agreement.

The global CO2 emissions from land-use change are estimated as 0.9 ± 0.7 GtC in 2020, 0.2 GtC lower than 2019, which had featured particularly large peat and tropical deforestation and degradation fires. The surge in deforestation fires in the Amazon, causing about 30 % higher emissions from deforestation and degradation fires in 2019 over the previous decade, continued into 2020 (GFED4.1s, van der Werf et al., 2017). However, the unusually dry conditions for a non-El Niño year that occurred in Indonesia in 2019 and led to fire emissions from peat burning, deforestation, and degradation in equatorial Asia to be about twice as large as the average over the previous decade (GFED4.1s, van der Werf et al., 2017) ceased in 2020. However, confidence in the annual change remains low. While the mentioned fires are clearly attributable to land-use activity, foremost deforestation and peat burning, and may have been reinforced by dry weather conditions, as was the case in Indonesia in 2019, wildfires also occur naturally. In particular, the extreme fire events in recent years in Australia, Siberia, and California were unrelated to land-use change and are thus not attributed to ELUC, but to the natural land sink, and are discussed in Sect. 3.6.2.

Land-use change and related emissions may have been affected by the COVID-19 pandemic (e.g. Poulter et al., 2021). Although emissions from tropical deforestation and degradation fires have been decreasing from 2019 to 2020 on the global scale, they increased in Latin America (GFED4s; van der Werf et al., 2017). During the period of the pandemic, environmental protection policies and their implementation may have been weakened in Brazil (Vale et al., 2021). In other countries, too, monitoring capacities and legal enforcement of measures to reduce tropical deforestation have been reduced due to budget restrictions of environmental agencies or impairments to ground-based monitoring that prevents land grabs and tenure conflicts (Brancalion et al., 2020; Amador-Jiménez et al., 2020). Effects of the pandemic on trends in fire activity or forest cover changes are hard to separate from those of general political developments and environmental changes, and the long-term consequences of disruptions in agricultural and forestry economic activities (e.g. Gruère and Brooks, 2020; Golar et al., 2020; Beckman and Countryman, 2021) remain to be seen.

With wet conditions in Indonesia and a below-average fire season in South America our preliminary estimate of ELUC for 2021 is substantially lower than the 2011–2020 average. By the end of September 2021 emissions from tropical deforestation and degradation fires were estimated to be 222 TgC, down from 347 TgC in 2019 and 288 in 2020 (315 TgC 1997–2020 average). Peat fire emissions in equatorial Asia were estimated to be 1 TgC, down from 117 TgC in 2019 and 2 TgC in 2020 (74 TgC 1997–2020 average) (GFED4.1s, van der Werf et al., 2017). Based on the fire emissions until the end of September, we expect ELUC emissions of around 0.8 GtC in 2021. Note that although our extrapolation is based on tropical deforestation and degradation fires, degradation attributable to selective logging, edge effects, or fragmentation will not be captured.

Atmospheric CO2 concentration was approximately 277 parts per million (ppm) in 1750 (Joos and Spahni, 2008), reaching 300 ppm in the 1910s, 350 ppm in the late 1980s, and 412.44 ± 0.1 ppm in 2020 (Dlugokencky and Tans, 2022; Fig. 1). The mass of carbon in the atmosphere increased by 48 % from 590 GtC in 1750 to 876 GtC in 2020. Current CO2 concentrations in the atmosphere are unprecedented in the last 2 million years, and the current rate of atmospheric CO2 increase is at least 10 times faster than at any other time during the last 800 000 years (Canadell et al., 2022).

The growth rate in atmospheric CO2 level increased from 1.7 ± 0.07 GtC yr-1 in the 1960s to 5.1 ± 0.02 GtC yr-1 during 2011–2020 with important decadal variations (Table 6, Figs. 3 and 4).

During the last decade (2011–2020), the growth rate in atmospheric CO2 concentration continued to increase, albeit with large interannual variability (Fig. 4).

The airborne fraction (AF), defined as the ratio of atmospheric CO2 growth rate to total anthropogenic emissions, 2AF=GATM/(EFOS+ELUC) provides a diagnostic of the relative strength of the land and ocean carbon sinks in removing part of the anthropogenic CO2 perturbation. The evolution of AF over the last 60 years shows no significant trend, remaining at around 45 %, albeit showing a large interannual variability driven by the year-to-year variability in GATM (Fig. 8). The observed stability of the airborne fraction over the 1960–2020 period indicates that the ocean and land CO2 sinks have been removing on average about 55 % of the anthropogenic emissions (see Sect. 3.5 and 3.6).

The growth rate in atmospheric CO2 concentration was 5.0 ± 0.2 GtC (2.37 ± 0.08 ppm) in 2020 (Fig. 4; Dlugokencky and Tans, 2022), very close to the 2011–2020 average. The 2020 decrease in EFOS and ELUC of about 0.7 GtC propagated to an atmospheric CO2 growth rate reduction of 0.38 GtC (0.18 ppm), given the significant interannual variability of the land carbon sink.

The 2021 growth in atmospheric CO2 concentration (GATM) is projected to be about 5.3 GtC (2.49 ppm) based on GLO observations until the end of December 2021, bringing the atmospheric CO2 concentration to an expected level of 414.67 ppm averaged over the year, 50 % over the pre-industrial level.

Cumulated since 1850, the ocean sink adds up to 170 ± 35 GtC, with two-thirds of this amount being taken up by the global ocean since 1960. Over the historical period, the ocean sink increased in pace with the anthropogenic emissions exponential increase (Fig. 3b). Since 1850, the ocean has removed 26 % of total anthropogenic emissions.

The ocean CO2 sink increased from 1.1 ± 0.4 GtC yr-1 in the 1960s to 2.8 ± 0.4 GtC yr-1 during 2011–2020 (Table 6), with interannual variations of the order of a few tenths of GtC yr-1 (Fig. 9). The ocean-borne fraction (SOCEAN/(EFOS+ELUC) has been remarkably constant, around 25 % on average (Fig. 8). Variations around this mean illustrate decadal variability of the ocean carbon sink. So far, there is no indication of a decrease in the ocean-borne fraction from 1960 to 2020. The increase in the ocean sink is primarily driven by the increased atmospheric CO2 concentration, with the strongest CO2-induced signal in the North Atlantic and the Southern Ocean (Fig. 10a). The effect of climate change is much weaker, reducing the ocean sink globally by 0.12 ± 0.07 GtC yr-1 or 5 % (2011–2020, range -0.8 % to -7.4 %), and does not show clear spatial patterns across the GOBMs ensemble (Fig. 10b). This is the combined effect of change and variability in all atmospheric forcing fields, previously attributed to wind and temperature changes in one model (Le Quéré et al., 2010).

The global net air–sea CO2 flux is a residual of large natural and anthropogenic CO2 fluxes into and out of the ocean with distinct regional and seasonal variations (Figs. 6 and B1). Natural fluxes dominate on regional scales but largely cancel out when integrated globally (Gruber et al., 2009). Mid-latitudes in all basins and the high-latitude North Atlantic dominate the ocean CO2 uptake where low temperatures and high wind speeds facilitate CO2 uptake at the surface (Takahashi et al., 2009). In these regions, mode, intermediate, and deep-water masses are formed that transport anthropogenic carbon into the ocean interior, thus allowing for continued CO2 uptake at the surface. Outgassing of natural CO2 occurs mostly in the tropics, especially in the equatorial upwelling region, and to a lesser extent in the North Pacific and polar Southern Ocean, mirroring a well-established understanding of regional patterns of air–sea CO2 exchange (e.g. Takahashi et al., 2009; Gruber et al., 2009). These patterns are also noticeable in the Surface Ocean CO2 Atlas (SOCAT) dataset, where an ocean fCO2 value above the atmospheric level indicates outgassing (Fig. B1). This map further illustrates the data sparsity in the Indian Ocean and the Southern Hemisphere in general.

Interannual variability of the ocean carbon sink is driven by climate variability with a first-order effect from a stronger ocean sink during large El Niño events (e.g. 1997–1998) (Fig. 9; Rödenbeck et al., 2014; Hauck et al., 2020). The GOBMs show the same patterns of decadal variability as the mean of the fCO2-based data products, with a stagnation of the ocean sink in the 1990s and a strengthening since the early 2000s (Fig. 9, Le Quéré et al., 2007; Landschützer et al., 2015, 2016; DeVries et al., 2017; Hauck et al., 2020; McKinley et al., 2020). Different explanations have been proposed for this decadal variability, ranging from the ocean's response to changes in atmospheric wind and pressure systems (e.g. Le Quéré et al., 2007; Keppler and Landschützer, 2019), including variations in upper ocean overturning circulation (DeVries et al., 2017), to the eruption of Mount Pinatubo and its effects on sea surface temperature and slowed atmospheric CO2 growth rate in the 1990s (McKinley et al., 2020). The main origin of the decadal variability is a matter of debate with a number of studies initially pointing to the Southern Ocean (see review in Canadell et al., 2022), but also contributions from the North Atlantic and North Pacific (Landschützer et al., 2016; DeVries et al., 2019) or a global signal (McKinley et al., 2020) were proposed.

Although all individual GOBMs and data products fall within the observational constraint, the ensemble means of GOBMs and data products adjusted for the riverine flux diverge over time with a mean offset increasing from 0.24 GtC yr-1 in the 1990s to 0.66 GtC yr-1 in the decade 2011–2020 and reaching 1.1 GtC yr-1 in 2020. The SOCEAN trend diverges with a factor-of-2 difference since 2002 (GOBMs: 0.3 ± 0.1 GtC yr-1 per decade, data products: 0.7 ± 0.2 GtC yr-1 per decade, best estimate: 0.5 GtC yr-1 per decade) and with a factor-of-3 since 2010 (GOBMs: 0.3 ± 0.1 GtC yr-1 per decade, data products: 0.9 ± 0.3 GtC yr-1 per decade, best estimate: 0.6 GtC yr-1 per decade). The GOBM estimate is lower than in the previous global carbon budget (Friedlingstein et al., 2020), because one high-sink model was not available. The effect of two models (CNRM, MOM6-COBALT) revising their estimates downwards was largely balanced by two models revising their estimate upwards (FESOM-REcoM, PlankTOM).

The discrepancy between the two types of estimates stems mostly from a larger Southern Ocean sink in the data products prior to 2001, and from a larger SOCEAN trend in the northern and southern extra-tropics since then (Fig. 12). Possible explanations for the discrepancy in the Southern Ocean could be missing winter observations and data sparsity in general (Bushinsky et al., 2019; Gloege et al., 2021), model biases (as indicated by the large model spread in the south, Fig. 12, and the larger model–data mismatch, Fig. B2), or uncertainties in the regional river flux adjustment (Hauck et al., 2020; Lacroix et al., 2020).

During 2010–2016, the ocean CO2 sink appears to have intensified in line with the expected increase from atmospheric CO2 (McKinley et al., 2020). This effect is stronger in the fCO2-based data products (Fig. 9, GOBMs: +0.43 GtC yr-1, data products: +0.56 GtC yr-1). The reduction of -0.09 GtC yr-1 (range: -0.30 to +0.12 GtC yr-1) in the ocean CO2 sink in 2017 is consistent with the return to normal conditions after the El Niño in 2015–2016, which caused an enhanced sink in previous years. After 2017, the GOBMs ensemble mean suggests the ocean sink levels off at about 2.5 GtC yr-1, whereas the data products' estimate increases by 0.3 GtC yr-1 over the same period.

The estimated ocean CO2 sink was 3.0 ± 0.4 GtC in 2020. This is the average of GOBMs and data products, and is a small increase of 0.02 GtC compared to 2019, in line with the competing effects from an expected sink strengthening from atmospheric CO2 growth and expected sink weakening from La Niña conditions. There is, however, a substantial difference between GOBMs and fCO2-based data products in their mean 2020 SOCEAN estimate (GOBMs: 2.5 GtC, data products: 3.5 GtC). While the GOBMs simulate a stagnation of the sink from 2019 to 2020 (-0.02 ± 0.11 GtC), the data products suggest an increase by 0.06 GtC, although not significant at the 1σ level (±0.13 GtC). Four models and four data products show an increase in SOCEAN (GOBMs up to +0.18 GtC, data product up to +0.21 GtC), while four models and three data products show no change or a decrease in SOCEAN (GOBMs down to -0.12 GtC, data products down to -0.13 GtC; Fig. 9). The data products have a larger uncertainty at the tails of the reconstructed time series (e.g. Watson et al., 2020). Specifically, the data products' estimate of the last year is regularly adjusted in the following release owing to the tail effect and an incrementally increasing data availability with 1–5 years lag (Fig. 9 bottom).

Using a feed-forward neural network method (see Sect. 2.4) we project an ocean sink of 2.9 GtC for 2021. This is a reduction of the sink by 0.1 GtC relative to the 2020 value, which we attribute to La Niña conditions in January to May 2021 and projections of a re-emergence of La Niña later in the year.

The evaluation of the ocean estimates (Fig. B2) shows an RMSE from annually detrended data of 1.3 to 2.8 µatm for the seven fCO2-based data products over the globe, relative to the fCO2 observations from the SOCAT v2021 dataset for the period 1990–2020. The GOBMs RMSEs are larger and range from 3.3 to 5.9 µatm. The RMSEs are generally larger at high latitudes compared to the tropics, for both the data products and the GOBMs. The data products have RMSEs of 1.3 to 3.6 µatm in the tropics, 1.3 to 2.7 µatm in the north, and 2.2 to 6.1 µatm in the south. Note that the data products are based on the SOCAT v2021 database; hence the latter are not an independent dataset for the evaluation of the data products. The GOBM RMSEs are more spread across regions, ranging from 2.7 to 4.3 µatm in the tropics, 2.9 to 6.9 µatm in the north, and 6.4 to 9.8 µatm in the south. The higher RMSEs occur in regions with stronger climate variability, such as the northern and southern high latitudes (poleward of the subtropical gyres). The upper ranges of the model RMSEs have decreased somewhat relative to Friedlingstein et al. (2020), owing to one model with upper-end RMSE not being represented this year, and the reduction of RMSE in one model (MPIOM-HAMOCC6), presumably related to the inclusion of riverine carbon fluxes.

The additional simulation C allows the steady-state anthropogenic carbon component (sim C – sim B) to be separated and the model flux and DIC inventory change to be compared directly to the interior ocean estimate of Gruber et al. (2019) without further assumptions. The GOBMs ensemble average of steady-state anthropogenic carbon inventory change 1994–2007 amounts to 2.1 GtC yr-1 and is significantly lower than the 2.6 ± 0.3 GtC yr-1 estimated by Gruber et al. (2019). Only the three models with the highest sink estimate fall within the range reported by Gruber et al. (2019). This suggests that most of the models underestimate anthropogenic carbon uptake by the ocean likely due to biases in ocean carbon transport and mixing from the surface mixed layer to the ocean interior.

The reported SOCEAN estimate from GOBMs and data products is 2.1 ± 0.4 GtC yr-1 over the period 1994 to 2007, which is in agreement with the ocean interior estimate of 2.2 ± 0.4 GtC yr-1 when accounting for the climate effect on the natural CO2 flux of -0.4 ± 0.24 GtC yr-1 (Gruber et al., 2019) to match the definition of SOCEAN used here (Hauck et al., 2020). This comparison depends critically on the estimate of the climate effect on the natural CO2 flux, which is smaller from the GOBMs (Sect. 3.5.2) than in Gruber et al. (2019).

Cumulated since 1850, the terrestrial CO2 sink amounts to 195 ± 45 GtC, 30 % of total anthropogenic emissions. Over the historical period, the sink increased in pace with the anthropogenic emissions exponential increase (Fig. 3b).

The terrestrial CO2 sink increased from 1.2 ± 0.5 GtC yr-1 in the 1960s to 3.1 ± 0.6 GtC yr-1 during 2010–2019, with important interannual variations of up to 2 GtC yr-1 generally showing a decreased land sink during El Niño events (Fig. 7), responsible for the corresponding enhanced growth rate in atmospheric CO2 concentration. The larger land CO2 sink during 2010–2019 compared to the 1960s is reproduced by all the DGVMs in response to the combined atmospheric CO2 increase and the changes in climate, and consistent with constraints from the other budget terms (Table 5).

Over the period 1960 to present the increase in the global terrestrial CO2 sink is largely attributed to the CO2 fertilization effect in the models (Prentice et al., 2001; Piao et al., 2009), directly stimulating plant photosynthesis and increased plant water use in water-limited systems, with a small negative contribution of climate change (Fig. 10). There is a range of evidence to support a positive terrestrial carbon sink in response to increasing atmospheric CO2, albeit with uncertain magnitude (Walker et al., 2021). As expected from theory the greatest CO2 effect is simulated in the tropical forest regions, associated with warm temperatures and long growing seasons (Hickler et al., 2008) (Fig. 10a). However, evidence from tropical intact forest plots indicate an overall decline in the land sink across Amazonia (1985–2011), attributed to enhanced mortality offsetting productivity gains (Brienen et al., 2005; Hubau et al., 2020). During 2011–2020 the land sink is positive in all regions (Fig. 6) with the exception of central and eastern Brazil, southwest USA and northern Mexico, southeast Europe and central Asia, South Africa, and eastern Australia, where the negative effects of climate variability and change (i.e. reduced rainfall) counterbalance CO2 effects. This is clearly visible in Fig. 10 where the effects of CO2 (Fig. 10a) and climate (Fig. 10b) as simulated by the DGVMs are isolated. The negative effect of climate is the strongest in most of South America, Central America, southwest USA, and central Europe (Fig. 10b). Globally, climate change reduces the land sink by 0.45 ± 0.39 GtC yr-1 or 15 % (2011–2020).

In the past years several regions experienced record-setting fire events. While global burned area has declined over the past decades mostly due to declining fire activity in savannas (Andela et al., 2017), forest fire emissions are rising and have the potential to counter the negative fire trend in savannas (Zheng et al., 2021). Noteworthy events include the 2019–2020 Black Summer event in Australia (emissions of roughly 0.2 GtC; van der Velde et al., 2021) and Siberia in 2021, where emissions approached 0.4 GtC or 3 times the 1997–2020 average according to GFED4s. While other regions, including western USA and Mediterranean Europe, also experienced intense fire seasons in 2021 their emissions are substantially lower.

Despite these regional negative effects of climate change on SLAND, the efficiency of land to remove anthropogenic CO2 emissions has remained broadly constant over the last six decades, with a land-borne fraction (SLAND/(EFOS+ELUC) of ∼ 30 % (Fig. 8).

The terrestrial CO2 sink from the DGVM ensemble was 2.9 ± 1.0 GtC in 2020, slightly below the decadal average of 3.1 GtC yr-1 (Fig. 4, Table 6). We note that the DGVM estimate for 2020 is significantly larger than the 2.1 ± 0.9 GtC yr-1 estimate from the residual sink from the global budget (EFOS+ELUC-GATM-SOCEAN) (Table 5).

Using a feed-forward neural network method (see Sect. 2.5) we project a land sink of 3.3 GtC for 2021. This is an increase in the land sink by 0.3 GtC relative to the 2020 value which we attribute to La Niña conditions in 2021.

The evaluation of the DGVMs (Fig. B3) shows generally high skill scores across models for runoff, and to a lesser extent for vegetation biomass, GPP, and ecosystem respiration (Fig. B3a). Skill score was lowest for leaf area index and net ecosystem exchange, with the widest disparity among models for soil carbon. Further analysis of the results will be provided separately, focusing on the strengths and weaknesses in the DGVMs ensemble and its validity for use in the global carbon budget.

In the period 2011–2020, the bottom-up view of total global carbon sinks provided by the GCB (SOCEAN+SLAND-ELUC) agrees closely with the top-down budget delivered by the atmospheric inversions. Figure 11 shows both total sink estimates of the last decade split by land and ocean, which match the difference between GATM and EFOS to within 0.06–0.17 GtC yr-1 for inverse models and to 0.3 GtC yr-1 for the GCB mean. The latter represents the BIM discussed in Sect. 3.8, which by design is minimal for the inverse models.

The distributions based on the individual models and data products reveal substantial spread but converge near the decadal means quoted in Tables 5 and 6. Sink estimates for SOCEAN and from inverse models are mostly non-Gaussian, while the ensemble of DGVMs appears more normally distributed justifying the use of a multi-model mean and standard deviation for their errors in the budget. Noteworthy is that the tails of the distributions provided by the land and ocean bottom-up estimates would not agree with the global constraint provided by the fossil fuel emissions and the observed atmospheric CO2 growth rate (EFOS-GATM). This illustrates the power of the atmospheric joint constraint from GATM and the global CO2 observation network it derives from.

The total atmosphere-to-land fluxes (SLAND-ELUC), calculated here as the difference between SLAND from the DGVMs and ELUC from the bookkeeping models, amounts to a 1.9 ± 0.9 GtC yr-1 sink during 2011–2020 (Table 5). Estimates of total atmosphere-to-land fluxes (SLAND-ELUC) from the DGVMs alone (1.6 ± 0.6 GtC yr-1) are consistent with this estimate and also with the global carbon budget constraint (EFOS-GATM-SOCEAN, 1.7 ± 0.8 GtC yr-1 Table 5). Consistent with the bookkeeping model estimates, the DGVM-based ELUC is substantially lower than in Friedlingstein et al. (2020) due to the improved land-cover forcing (see Sect. 3.2.2), increasing their total atmosphere-to-land fluxes and hence the consistency with the budget constraint. For the last decade (2011–2020), the inversions estimate the net atmosphere-to-land uptake to lie within a range of 1.3 to 2.0 GtC yr-1, consistent with the GCB and DGVM estimates of SLAND-ELUC (Fig. 11, Fig. 2, top row).