Mapping in-situ eddy covariance measurements of terrestrial land-atmosphere fluxes to the globe is a key method for diagnosing the Earth system from a data-driven perspective. We describe the first global products (called X-BASE) from a newly implemented up-scaling framework, FLUXCOM-X. The X-BASE products comprise of estimates of CO2 net ecosystem exchange (NEE), gross primary productivity (GPP) as well as evapotranspiration (ET) and, for the first time, a novel fully data-driven global transpiration product (ETT), at high spatial (0.05°) and temporal (hourly) resolution. X-BASE estimates the global NEE at -5.75 ± 0.33 Pg C ⋅ yr-1 for the period 2001–2020, showing a much higher consistency with independent atmospheric carbon cycle constraints compared to the previous versions of FLUXCOM. The improvement of global NEE was likely only possible thanks to the international effort to increase the precision and consistency of eddy covariance collection and processing pipelines, as well as to the extension of the measurements to more site-years resulting in a wider coverage of bio-climatic conditions. However, X-BASE global net ecosystem exchange shows a very low inter-annual variability, which is common to state-of-the-art data-driven flux products and remains a scientific challenge. With 125 ± 2.1 Pg C ⋅ yr-1 for the same period, X-BASE GPP is slightly higher than previous FLUXCOM estimates, mostly in temperate and boreal areas. X-BASE evapotranspiration amounts to 74.7x10³ ± 0.9x10³ km3 globally for the years 2001–2020, but exceeds precipitation in many dry areas likely indicating overestimation in these regions. On average 57 % of evapotranspiration are estimated to be transpiration, in good agreement with isotope-based approaches, but higher than estimates from many land surface models. Despite considerable improvements to the previous up-scaling products, many further opportunities for development exist. Pathways of exploration include methodological choices in the selection and processing of eddy-covariance and satellite observations, their ingestion into the framework, and the configuration of machine learning methods. For this, the new FLUXCOM-X framework was specifically designed to have the necessary flexibility to experiment, diagnose, and converge to more accurate global flux estimates.

Mapping in situ eddy covariance measurements of terrestrial land-atmosphere fluxes to the globe is a key method for diagnosing the Earth system from a data-driven perspective. We describe the first global products (called X-BASE) from a newly implemented upscaling framework, FLUXCOM-X, representing an advancement from the previous generation of FLUXCOM products in terms of flexibility and technical capabilities. The X-BASE products are comprised of estimates of CO2 net ecosystem exchange (NEE), gross primary productivity (GPP), evapotranspiration (ET), and for the first time a novel, fully data-driven global transpiration product (ETT), at high spatial (0.05 degrees) and temporal (hourly) resolution. X-BASE estimates the global NEE at -5.75 +/- 0.33 Pg C yr(-1) for the period 2001-2020, showing a much higher consistency with independent atmospheric carbon cycle constraints compared to the previous versions of FLUXCOM. The improvement of global NEE was likely only possible thanks to the international effort to increase the precision and consistency of eddy covariance collection and processing pipelines, as well as to the extension of the measurements to more site years resulting in a wider coverage of bioclimatic conditions. However, X-BASE global net ecosystem exchange shows a very low interannual variability, which is common to state-of-the-art data-driven flux products and remains a scientific challenge. With 125 +/- 2.1 Pg C yr(-1) for the same period, X-BASE GPP is slightly higher than previous FLUXCOM estimates, mostly in temperate and boreal areas. X-BASE evapotranspiration amounts to 74.7x10(3) +/- 0.9x10(3) km(3) globally for the years 2001-2020 but exceeds precipitation in many dry areas, likely indicating overestimation in these regions. On average 57 % of evapotranspiration is estimated to be transpiration, in good agreement with isotope-based approaches, but higher than estimates from many land surface models. Despite considerable improvements to the previous upscaling products, many further opportunities for development exist. Pathways of exploration include methodological choices in the selection and processing of eddy covariance and satellite observations, their ingestion into the framework, and the configuration of machine learning methods. For this, the new FLUXCOM-X framework was specifically designed to have the necessary flexibility to experiment, diagnose, and converge to more accurate global flux estimates.

Separating the components of ecosystem-scale carbon exchange is crucial in order to develop better models and future predictions of the terrestrial carbon cycle. However, there are several uncertainties and unknowns related to current photosynthesis estimates. In this study, we evaluate four different methods for estimating photosynthesis at a boreal forest at the ecosystem scale, of which two are based on carbon dioxide (CO2) flux measurements and two on carbonyl sulfide (COS) flux measurements. The CO2-based methods use traditional flux partitioning and artificial neural networks to separate the net CO2 flux into respiration and photosynthesis. The COS-based methods make use of a unique 5-year COS flux data set and involve two different approaches to determine the leaf-scale relative uptake ratio of COS and CO2 (LRU), of which one (LRUCAP) was developed in this study. LRUCAP was based on a previously tested stomatal optimization theory (CAP), while LRUPAR was based on an empirical relation to measured radiation. For the measurement period 2013-2017, the artificial neural network method gave a GPP estimate very close to that of traditional flux partitioning at all timescales. On average, the COS-based methods gave higher GPP estimates than the CO2-based estimates on daily (23% and 7% higher, using LRUPAR and LRUCAP, respectively) and monthly scales (20% and 3% higher), as well as a higher cumulative sum over 3 months in all years (on average 25% and 3% higher). LRUCAP was higher than LRU estimated from chamber measurements at high radiation, leading to underestimation of midday GPP relative to other GPP methods. In general, however, use of LRUCAP gave closer agreement with CO2-based estimates of GPP than use of LRUPAR. When extended to other sites, LRUCAP may be more robust than LRUPAR because it is based on a physiological model whose parameters can be estimated from simple measurements or obtained from the literature. In contrast, the empirical radiation relation in LRUPAR may be more site-specific. However, this requires further testing at other measurement sites.

Monitoring ecosystem functions in forests is a priority in a climate change scenario, as climate-induced events may initially alter the functions more than slow-changing attributes, such as biomass. The ecosystem functional properties (EFPs) are quantities that characterize key ecosystem processes. They can be derived by point observations of gas and energy exchanges between the ecosystems and the atmosphere that are collected globally at FLUXNET flux tower sites and upscaled at ecosystem level. The properties here considered describe the ability of ecosystems to optimize the use of resources for carbon uptake. They represent functional forest information, are dependent on environmental drivers, linked to leaf traits and forest structure, and influenced by climate change effects. The ability of vegetation optical depth (VOD) to provide forest functional information is investigated using 2011-2014 satellite data collected by the Soil Moisture and Ocean Salinity mission and using the EFPs as reference dataset. Tropical forests in Africa and South America were analyzed, also according to ecological homogeneous units. VOD jointly with water deficit information explained 93% and 87% of the yearly variability in both flux upscaled maximum gross primary productivity and light use efficiency functional properties, in Africa and South America forests respectively. Maps of the retrieved properties evidenced changes in forest functional responses linked to anomalous climate-induced events during the study period. The findings indicate that VOD can support the flux upscaling process in the tropical range, affected by high uncertainty, and the detection of forest anomalous functional responses. Preliminary temporal analysis of VOD and EFP signals showed fine-grained variability in periodicity, in signal dephasing, and in the strength of the relationships. In selected drier forest types, these satellite data could also support the monitoring of functional dynamics.

FLUXNET comprises globally distributed eddy-covariance-based estimates of carbon fluxes between the biosphere and the atmosphere. Since eddy covariance flux towers have a relatively small footprint and are distributed unevenly across the world, upscaling the observations is necessary to obtain global-scale estimates of biosphere–atmosphere exchange. Based on cross-consistency checks with atmospheric inversions, sun-induced fluorescence (SIF) and dynamic global vegetation models (DGVMs), here we provide a systematic assessment of the latest upscaling efforts for gross primary production (GPP) and net ecosystem exchange (NEE) of the FLUXCOM initiative, where different machine learning methods, forcing data sets and sets of predictor variables were employed. Spatial patterns of mean GPP are consistent across FLUXCOM and DGVM ensembles (R2>0.94 at 1∘ spatial resolution) while the majority of DGVMs show, for 70 % of the land surface, values outside the FLUXCOM range. Global mean GPP magnitudes for 2008–2010 from FLUXCOM members vary within 106 and 130 PgC yr−1 with the largest uncertainty in the tropics. Seasonal variations in independent SIF estimates agree better with FLUXCOM GPP (mean global pixel-wise R2∼0.75) than with GPP from DGVMs (mean global pixel-wise R2∼0.6). Seasonal variations in FLUXCOM NEE show good consistency with atmospheric inversion-based net land carbon fluxes, particularly for temperate and boreal regions (R2>0.92). Interannual variability of global NEE in FLUXCOM is underestimated compared to inversions and DGVMs. The FLUXCOM version which also uses meteorological inputs shows a strong co-variation in interannual patterns with inversions (R2=0.87 for 2001–2010). Mean regional NEE from FLUXCOM shows larger uptake than inversion and DGVM-based estimates, particularly in the tropics with discrepancies of up to several hundred grammes of carbon per square metre per year. These discrepancies can only partly be reconciled by carbon loss pathways that are implicit in inversions but not captured by the flux tower measurements such as carbon emissions from fires and water bodies. We hypothesize that a combination of systematic biases in the underlying eddy covariance data, in particular in tall tropical forests, and a lack of site history effects on NEE in FLUXCOM are likely responsible for the too strong tropical carbon sink estimated by FLUXCOM. Furthermore, as FLUXCOM does not account for CO2 fertilization effects, carbon flux trends are not realistic. Overall, current FLUXCOM estimates of mean annual and seasonal cycles of GPP as well as seasonal NEE variations provide useful constraints of global carbon cycling, while interannual variability patterns from FLUXCOM are valuable but require cautious interpretation. Exploring the diversity of Earth observation data and of machine learning concepts along with improved quality and quantity of flux tower measurements will facilitate further improvements of the FLUXCOM approach overall.

The eddy covariance (EC) technique is used to measure the net ecosystem exchange (NEE) of CO2 between ecosystems and the atmosphere, offering a unique opportunity to study ecosystem responses to climate change. NEE is the difference between the total CO2 release due to all respiration processes (RECO), and the gross carbon uptake by photosynthesis (GPP). These two gross CO2 fluxes are derived from EC measurements by applying partitioning methods that rely on physiologically based functional relationships with a limited number of environmental drivers. However, the partitioning methods applied in the global FLUXNET network of EC observations do not account for the multiple co-acting factors that modulate GPP and RECO flux dynamics. To overcome this limitation, we developed a hybrid data-driven approach based on combined neural networks (NNC-part). NNIC-part part incorporates process knowledge by introducing a photosynthetic response based on the light-use efficiency (LUE) concept, and uses a comprehensive dataset of soil and micrometeorological variables as fluxes drivers. We applied the method to 36 sites from the FLUXNET2015 dataset and found a high consistency in the results with those derived from other standard partitioning methods for both GPP (R-2 > .94) and RECO (R-2 > .8). High consistency was also found for (a) the diurnal and seasonal patterns of fluxes and (b) the ecosystem functional responses. NNC-part performed more realistic than the traditional methods for predicting additional patterns of gross CO2 fluxes, such as: (a) the GPP response to VPD, (b) direct effects of air temperature on GPP dynamics, (c) hysteresis in the diel cycle of gross CO2 fluxes, (d) the sensitivity of LUE to the diffuse to direct radiation ratio, and (e) the post rain respiration pulse after a long dry period. In conclusion, NNC-part is a valid data-driven approach to provide GPP and RECO estimates and complementary to the existing partitioning methods.

Global estimation of Gross Primary Production (GPP) - the uptake of atmospheric carbon dioxide by plants through photosynthesis - is commonly based on optical satellite remote sensing data. This presents a source driven approach since it uses the amount of absorbed light, the main driver of photosynthesis, as a proxy for GPP. Vegetation Optical Depth (VOD) estimates obtained from microwave sensors provide an alternative and independent data source to estimate GPP on a global scale, which may complement existing GPP products. Recent studies have shown that VOD is related to aboveground biomass, and that both VOD and temporal changes in VOD relate to GPP. In this study, we build upon this concept and propose a model for estimating GPP from VOD. Since the model is driven by vegetation biomass, as observed through VOD, it presents a carbon sink driven approach to quantify GPP and, therefore, is conceptually different from common source-driven approaches. The model developed in this study uses single frequencies from active or passive microwave VOD retrievals from C-, X- and Ku-band (Advanced Scatterometer (ASCAT) and Advanced Microwave Scanning Radiometer for Earth Observation (AMSR-E)) to estimate GPP at the global scale. We assessed the ability for temporal and spatial extrapolation of the model using global GPP from FLUXCOM and in situ GPP from FLUXNET. We further performed upscaling of in situ GPP based on different VOD data sets and compared these estimates with the FLUXCOM and MODerate-resolution Imaging Spectroradiometer (MODIS) GPP products. Our results show that the model developed for individual grid cells using VOD and change in VOD as input performs well in predicting temporal patterns in GPP for all VOD data sets. For spatial extrapolation of the model, however, additional input variables are needed to represent the spatial variability of the VOD-GPP relationship due to differences in vegetation type. As additional input variable, we included the grid cell median VOD (as a proxy for vegetation cover), which increased the model performance during cross validation. Mean annual GPP obtained for AMSR-E X-band data tends to overestimate mean annual GPP for FLUXCOM and MODIS but shows comparable latitudinal patterns. Overall, our findings demonstrate the potential of VOD for estimating GPP. The sink-driven approach provides additional information about GPP independent of optical data, which may contribute to our knowledge about the carbon source-sink balance in different ecosystems.

Although a key driver of Earth’s climate system, global land-atmosphere energy fluxes are poorly constrained. Here we use machine learning to merge energy flux measurements from FLUXNET eddy covariance towers with remote sensing and meteorological data to estimate global gridded net radiation, latent and sensible heat and their uncertainties. The resulting FLUXCOM database comprises 147 products in two setups: (1) 0.0833° resolution using MODIS remote sensing data (RS) and (2) 0.5° resolution using remote sensing and meteorological data (RS + METEO). Within each setup we use a full factorial design across machine learning methods, forcing datasets and energy balance closure corrections. For RS and RS + METEO setups respectively, we estimate 2001–2013 global (±1 s.d.) net radiation as 75.49 ± 1.39 W m −2 and 77.52 ± 2.43 W m −2 , sensible heat as 32.39 ± 4.17 W m −2 and 35.58 ± 4.75 W m −2 , and latent heat flux as 39.14 ± 6.60 W m −2 and 39.49 ± 4.51 W m −2 (as evapotranspiration, 75.6 ± 9.8 × 10 3 km 3 yr −1 and 76 ± 6.8 × 10 3 km 3 yr −1 ). FLUXCOM products are suitable to quantify global land-atmosphere interactions and benchmark land surface model simulations.

The relationship between vegetation optical depth (VOD) retrieved by L-band SMOS radiometer and forest above ground biomass (AGB) was investigated in tropical areas of Africa and South America. VOD was retrieved from the latest version of level 2 SMOS algorithm, while reference AGB was obtained from a pantropical database, encompassing a large number of ground plot data derived from field surveys conducted on both continents. In Africa and South-America, VOD increased with AGB, reaching saturation at about 350 Mg ha(-1). The strength of the relation was improved selecting VOD data in appropriate seasons, characterized by a higher dynamic range of values. The capability of VOD data to estimate AGB was further investigated using Random Forest decision trees, adding to VOD selected climate variables from the Climatic Research Unit (temperature, potential evapotranspiration, and precipitation) and water deficit data, and validating regression tests with ground data from the reference AGB database. The results for the five analyzed years indicate that the best estimates of AGB are obtained by the joined use of VOD and potential evapotranspiration input data, but all climate variables brought an improvement in AGB estimates. AGB estimates were relatively stable for the considered period, with limited variations possibly due to changes in biomass and to data quality of VOD and of climate variables. The VOD signal and estimated AGB were also analyzed according to ecological homogeneous units (ecoregions), evidencing data clusters, partially overlapped to each other, in the VOD - AGB plane.

At the global scale, the uptake of atmospheric carbon dioxide by terrestrial ecosystems through photosynthesis is commonly estimated through vegetation indices or biophysical properties derived from optical remote sensing data. Microwave observations of vegetated areas are sensitive to different components of the vegetation layer than observations in the optical domain and may therefore provide complementary information on the vegetation state, which may be used in the estimation of Gross Primary Production (GPP). However, the relation between GPP and Vegetation Optical Depth (VOD), a biophysical quantity derived from microwave observations, is not yet known. This study aims to explore the relationship between VOD and GPP. VOD data were taken from different frequencies (L-, C-, and X-band) and from both active and passive microwave sensors, including the Advanced Scatterometer (ASCAT), the Soil Moisture Ocean Salinity (SMOS) mission, the Advanced Microwave Scanning Radiometer for Earth Observation System (AMSR-E) and a merged VOD data set from various passive microwave sensors. VOD data were compared against FLUXCOM GPP and Solar-Induced chlorophyll Fluorescence (SIF) from the Global Ozone Monitoring Experiment-2 (GOME-2). FLUXCOM GPP estimates are based on the upscaling of flux tower GPP observations using optical satellite data, while SIF observations present a measure of photosynthetic activity and are often used as a proxy for GPP. For relating VOD to GPP, three variables were analyzed: original VOD time series, temporal changes in VOD (ΔVOD), and positive changes in VOD (ΔVOD≥0). Results show widespread positive correlations between VOD and GPP with some negative correlations mainly occurring in dry and wet regions for active and passive VOD, respectively. Correlations between VOD and GPP were similar or higher than between VOD and SIF. When comparing the three variables for relating VOD to GPP, correlations with GPP were higher for the original VOD time series than for ΔVOD or ΔVOD≥0 in case of sparsely to moderately vegetated areas and evergreen forests, while the opposite was true for deciduous forests. Results suggest that original VOD time series should be used jointly with changes in VOD for the estimation of GPP across biomes, which may further benefit from combining active and passive VOD data.

Data from the Soil Moisture and Ocean Salinity (SMOS) mission were exploited to explore its potential in providing ecosystem information. Here the strength of an innovative relationship, between SMOS V620 algorithm data and Ecosystem Functional Properties (EFPs) derived from flux towers data, was preliminary investigated for Africa and South America forests. Correlation was also explored for above ground forest biomass. High correlation values were found between SMOS vegetation optical depth (VOD) and a AGB reference map, and with EFPs at year level (2014). Different VOD-EFPs trends were found for different latitudinal belts, characterizing wet and dry forests. The results suggest that SMOS VOD data represent a tool able to provide repeated information on forest biomass and forest processes in tropical and subtropical ranges.

Large interannual variations in the measured growth rate of atmospheric carbon dioxide (CO2) originate primarily from fluctuations in carbon uptake by land ecosystems. It remains uncertain, however, to what extent temperature and water availability control the carbon balance of land ecosystems across spatial and temporal scales. Here we use empirical models based on eddy covariance data and process-based models to investigate the effect of changes in temperature and water availability on gross primary productivity (GPP), terrestrial ecosystem respiration (TER) and net ecosystem exchange (NEE) at local and global scales. We find that water availability is the dominant driver of the local interannual variability in GPP and TER. To a lesser extent this is true also for NEE at the local scale, but when integrated globally, temporal NEE variability is mostly driven by temperature fluctuations. We suggest that this apparent paradox can be explained by two compensatory water effects. Temporal water-driven GPP and TER variations compensate locally, dampening water-driven NEE variability. Spatial water availability anomalies also compensate, leaving a dominant temperature signal in the year-to-year fluctuations of the land carbon sink. These findings help to reconcile seemingly contradictory reports regarding the importance of temperature and water in controlling the interannual variability of the terrestrial carbon balance. Our study indicates that spatial climate covariation drives the global carbon cycle response.

Groundwater is an integral component of the water cycle, and it also influences the carbon cycle by supplying moisture to ecosystems. However, the extent and determinants of groundwater-vegetation interactions are poorly understood at the global scale. Using several high-resolution data products, we show that the spatial patterns of ecosystem gross primary productivity and groundwater table depth are correlated during at least one season in more than two thirds of the global vegetated area. Positive relationships, i.e., larger productivity under shallower groundwater table, predominate in moisture-limited dry to mesic conditions with herbaceous and shrub vegetation. Negative relationships, i.e., larger productivity under deeper groundwater, predominate in humid climates with forests, possibly indicating a drawdown of groundwater table due to substantial ecosystem water use. Interestingly, these opposite groundwater-vegetation interactions are primarily associated with differences in vegetation than with climate and surface characteristics. These findings put forth the first evidence, and a need for better representation, of extensive and non-negligible groundwater-vegetation interactions at the global scale.

Understanding the global carbon (C) cycle is of crucial importance to map current and future climate dynamics relative to global environmental change. A full characterization of C cycling requires detailed information on spatiotemporal patterns of surface–atmosphere fluxes. However, relevant C cycle observations are highly variable in their coverage and reporting standards. Especially problematic is the lack of integration of the carbon dioxide (CO2) exchange of the ocean, inland freshwaters and the land surface with the atmosphere. Here we adopt a data-driven approach to synthesize a wide range of observation-based spatially explicit surface–atmosphere CO2 fluxes from 2001 to 2010, to identify the state of today's observational opportunities and data limitations. The considered fluxes include net exchange of open oceans, continental shelves, estuaries, rivers, and lakes, as well as CO2 fluxes related to net ecosystem productivity, fire emissions, loss of tropical aboveground C, harvested wood and crops, as well as fossil fuel and cement emissions. Spatially explicit CO2 fluxes are obtained through geostatistical and/or remote-sensing-based upscaling, thereby minimizing biophysical or biogeochemical assumptions encoded in process-based models. We estimate a bottom-up net C exchange (NCE) between the surface (land, ocean, and coastal areas) and the atmosphere. Though we provide also global estimates, the primary goal of this study is to identify key uncertainties and observational shortcomings that need to be prioritized in the expansion of in situ observatories. Uncertainties for NCE and its components are derived using resampling. In many regions, our NCE estimates agree well with independent estimates from other sources such as process-based models and atmospheric inversions. This holds for Europe (mean ± 1 SD: 0.8 ± 0.1 PgC yr−1, positive numbers are sources to the atmosphere), Russia (0.1 ± 0.4 PgC yr−1), East Asia (1.6 ± 0.3 PgC yr−1), South Asia (0.3 ± 0.1 PgC yr−1), Australia (0.2 ± 0.3 PgC yr−1), and most of the Ocean regions. Our NCE estimates give a likely too large CO2 sink in tropical areas such as the Amazon, Congo, and Indonesia. Overall, and because of the overestimated CO2 uptake in tropical lands, our global bottom-up NCE amounts to a net sink of −5.4 ± 2.0 PgC yr−1. By contrast, the accurately measured mean atmospheric growth rate of CO2 over 2001–2010 indicates that the true value of NCE is a net CO2 source of 4.3 ± 0.1 PgC yr−1. This mismatch of nearly 10 PgC yr−1 highlights observational gaps and limitations of data-driven models in tropical lands, but also in North America. Our uncertainty assessment provides the basis for setting priority regions where to increase carbon observations in the future. High on the priority list are tropical land regions, which suffer from a lack of in situ observations. Second, extensive pCO2 data are missing in the Southern Ocean. Third, we lack observations that could enable seasonal estimates of shelf, estuary, and inland water–atmosphere C exchange. Our consistent derivation of data uncertainties could serve as prior knowledge in multicriteria optimization such as the Carbon Cycle Data Assimilation System (CCDAS) and atmospheric inversions, without over- or under-stating bottom-up data credibility. In the future, NCE estimates of carbon sinks could be aggregated at national scale to compare with the official national inventories of CO2 fluxes in the land use, land use change, and forestry sector, upon which future emission reductions are proposed.