A long-term reconstructed TROPOMI solar-induced fluorescence dataset using machine learning algorithms

Photosynthesis is a key process linking carbon and water cycles, and satellite-retrieved solar-induced chlorophyll fluorescence (SIF) can be a valuable proxy for photosynthesis. The TROPOspheric Monitoring Instrument (TROPOMI) on the Copernicus Sentinel-5P mission enables significant improvements in providing high spatial and temporal resolution SIF observations, but the short temporal coverage of the data records has limited its applications in long-term studies. This study uses machine learning to reconstruct TROPOMI SIF (RTSIF) over the 2001-2020 period in clear-sky conditions with high spatio-temporal resolutions (0.05° 8-day). Our machine learning model achieves high accuracies on the training and testing datasets (R2 = 0.907, regression slope = 1.001). The RTSIF dataset is validated against TROPOMI SIF and tower-based SIF, and compared with other satellite-derived SIF (GOME-2 SIF and OCO-2 SIF). Comparing RTSIF with Gross Primary Production (GPP) illustrates the potential of RTSIF for estimating gross carbon fluxes. We anticipate that this new dataset will be valuable in assessing long-term terrestrial photosynthesis and constraining the global carbon budget and associated water fluxes.

Application of a Machine Learning Algorithm in Generating an Evapotranspiration Data Product From Coupled Thermal Infrared and Microwave Satellite Observations

Land surface evapotranspiration (ET) is one of the main energy sources for atmospheric dynamics and a critical component of the local, regional, and global water cycles. Consequently, accurate measurement or estimation of ET is one of the most active topics in hydro-climatology research. With massive and spatially distributed observational data sets of land surface properties and environmental conditions being collected from the ground, airborne or space-borne platforms daily over the past few decades, many research teams have started to use big data science to advance the ET estimation methods. The Geostationary satellite Evapotranspiration and Drought (GET-D) product system was developed at the National Oceanic and Atmospheric Administration (NOAA) in 2016 to generate daily ET and drought maps operationally. The primary inputs of the current GET-D system are the thermal infrared (TIR) observations from NOAA GOES satellite series. Because of the cloud contamination to the TIR observations, the spatial coverage of the daily GET-D ET product has been severely impacted. Based on the most recent advances, we have tested a machine learning algorithm to estimate all-weather land surface temperature (LST) from TIR and microwave (MW) combined satellite observations. With the regression tree machine learning approach, we can combine the high accuracy and high spatial resolution of GOES TIR data with the better spatial coverage of passive microwave observations and LST simulations from a land surface model (LSM). The regression tree model combines the three LST data sources for both clear and cloudy days, which enables the GET-D system to derive an all-weather ET product. This paper reports how the all-weather LST and ET are generated in the upgraded GET-D system and provides an evaluation of these LST and ET estimates with ground measurements. The results demonstrate that the regression tree machine learning method is feasible and effective for generating daily ET under all weather conditions with satisfactory accuracy from the big volume of satellite observations.

Remote Sensing, Geophysics, and Modeling to Support Precision Agriculture—Part 2: Irrigation Management

Food and water security are considered the most critical issues globally due to the projected population growth placing pressure on agricultural systems. Because agricultural activity is known to be the largest consumer of freshwater, the unsustainable irrigation water use required by crops to grow might lead to rapid freshwater depletion. Precision agriculture has emerged as a feasible concept to maintain farm productivity while facing future problems such as climate change, freshwater depletion, and environmental degradation. Agriculture is regarded as a complex system due to the variability of soil, crops, topography, and climate, and its interconnection with water availability and scarcity. Therefore, understanding these variables’ spatial and temporal behavior is essential in order to support precision agriculture by implementing optimum irrigation water use. Nowadays, numerous cost- and time-effective methods have been highlighted and implemented in order to optimize on-farm productivity without threatening the quantity and quality of the environmental resources. Remote sensing can provide lateral distribution information for areas of interest from the regional scale to the farm scale, while geophysics can investigate non-invasively the sub-surface soil (vertically and laterally), mapping large spatial and temporal domains. Likewise, agro-hydrological modelling can overcome the insufficient on-farm physicochemical dataset which is spatially and temporally required for precision agriculture in the context of irrigation water scheduling.

Validation of Four Satellite-Derived Soil Moisture Products Using Ground-Based In Situ Observations over Northern China

Accurately obtaining the spatial distribution of soil moisture and its variability are the basis for the land-atmosphere interaction study. We investigated the fidelity of four satellite-based soil moisture products (AMSR2, CCI, SMAP, and SMOS) using in situ observation during the period 2019–2020. The spatial distribution and variability of different soil moisture products in northern China were analyzed for different seasons and climate zones. The satellite products showed the best performance of summer soil moisture with the bias and uncertainty of the three products (CCI, SMAP, and SMOS) being less than 0.041 and 0.097, whereas soil moisture showed a large bias in winter. For all seasons, AMSR2 and CCI demonstrated a positive bias whereas SMAP and SMOS showed a negative bias. CCI product had little bias in spring, summer, and fall in northern China, while SMAP and SMOS had the smallest bias in winter. For different climate zones, CCI product performed better in describing the temporal variability of soil moisture in arid climate zones with the correlation coefficients > 0.50 for most areas, while AMSR2 product provided a similar spatial distribution. In the eastern monsoon region, the soil moisture from SMAP and SMOS was found to have a large bias, whereas the bias in CCI product was small. Four products failed to reproduce the observed soil moisture characteristics in the transitional zones affected by the summer monsoon, with a positive bias found in AMSR2 and CCI and the largest biases in SMAP and SMOS products. We also suggest several reasons for the bias and error in the satellite soil moisture products. These results have important implications for soil moisture studies over midlatitude regions.

An Improved Cloud Gap-Filling Method for Longwave Infrared Land Surface Temperatures through Introducing Passive Microwave Techniques

Satellite-derived land surface temperature (LST) data are most commonly observed in the longwave infrared (LWIR) spectral region. However, such data suffer frequent gaps in coverage caused by cloud cover. Filling these ‘cloud gaps’ usually relies on statistical re-constructions using proximal clear sky LST pixels, whilst this is often a poor surrogate for shadowed LSTs insulated under cloud. Another solution is to rely on passive microwave (PM) LST data that are largely unimpeded by cloud cover impacts, the quality of which, however, is limited by the very coarse spatial resolution typical of PM signals. Here, we combine aspects of these two approaches to fill cloud gaps in the LWIR-derived LST record, using Kenya (East Africa) as our study area. The proposed “cloud gap-filling” approach increases the coverage of daily Aqua MODIS LST data over Kenya from <50% to >90%. Evaluations were made against the in situ and SEVIRI-derived LST data respectively, revealing root mean square errors (RMSEs) of 2.6 K and 3.6 K for the proposed method by mid-day, compared with RMSEs of 4.3 K and 6.7 K for the conventional proximal-pixel-based statistical re-construction method. We also find that such accuracy improvements become increasingly apparent when the total cloud cover residence time increases in the morning-to-noon time frame. At mid-night, cloud gap-filling performance is also better for the proposed method, though the RMSE improvement is far smaller (<0.3 K) than in the mid-day period. The results indicate that our proposed two-step cloud gap-filling method can improve upon performances achieved by conventional methods for cloud gap-filling and has the potential to be scaled up to provide data at continental or global scales as it does not rely on locality-specific knowledge or datasets.

Estimation of Evapotranspiration and Its Components across China Based on a Modified Priestley–Taylor Algorithm Using Monthly Multi-Layer Soil Moisture Data

Although soil moisture (SM) is an important constraint factor of evapotranspiration (ET), the majority of the satellite-driven ET models do not include SM observations, especially the SM at different depths, since its spatial and temporal distribution is difficult to obtain. Based on monthly three-layer SM data at a 0.25° spatial resolution determined from multi-sources, we updated the original Priestley Taylor–Jet Propulsion Laboratory (PT-JPL) algorithm to the Priestley Taylor–Soil Moisture Evapotranspiration (PT-SM ET) algorithm by incorporating SM control into soil evaporation (Es) and canopy transpiration (T). Both algorithms were evaluated using 17 eddy covariance towers across different biomes of China. The PT-SM ET model shows increased R2, NSE and reduced RMSE, Bias, with more improvements occurring in water-limited regions. SM incorporation into T enhanced ET estimates by increasing R2 and NSE by 4% and 18%, respectively, and RMSE and Bias were respectively reduced by 34% and 7 mm. Moreover, we applied the two ET algorithms to the whole of China and found larger increases in T and Es in the central, northeastern, and southern regions of China when using the PT-SM algorithm compared with the original algorithm. Additionally, the estimated mean annual ET increased from the northwest to the southeast. The SM constraint resulted in higher transpiration estimate and lower evaporation estimate. Es was greatest in the northwest arid region, interception was a large fraction in some rainforests, and T was dominant in most other regions. Further improvements in the estimation of ET components at high spatial and temporal resolution are likely to lead to a better understanding of the water movement through the soil–plant–atmosphere continuum.

Terrestrial Water Storage Change Retrieved by GRACE and Its Implication in the Tibetan Plateau: Estimating Areal Precipitation in Ungauged Region

The Tibetan Plateau (TP) is referred to as the water tower of Asia, where water storage and precipitation have huge impacts on most major Asian rivers. Based on gravity recovery and climate experiment data, this study analyzed the terrestrial water storage (TWS) changes and estimated areal precipitation based on the water balance equation in four different basins, namely, the upper Yellow River (UYE), the upper Yangtze River (UYA), the Yarlung Zangbo River (YZ), and the Qiangtang Plateau (QT). The results show that the TWS change exhibits different patterns in the four basins and varies from −13 to 2 mm/year from 2003 to 2017. The estimated mean annual precipitation was 260 ± 19 mm/year (QT), 697 ± 26 mm/year (UYA), 541 ± 36 mm/year (UYE), and 1160 ± 39 mm/year (YZ) which performed better than other precipitation products in the TP. It indicates a potential method for estimating basin-scale precipitation through integrating basin average precipitation from the water balance equation in the poorly gauged and ungauged regions.

Monitoring Residual Soil Moisture and Its Association to the Long-Term Variability of Rainfall over the Upper Blue Nile Basin in Ethiopia

Monitoring soil moisture and its association with rainfall variability is important to comprehend the hydrological processes and to set proper agricultural water use management to maximize crop growth and productivity. In this study, the European Space Agency’s Climate Change Initiative (ESA CCI) soil moisture product was applied to assess the dynamics of residual soil moisture in autumn (September to November) and its response to the long-term variability of rainfall in the Upper Blue Nile Basin (UBNB) of Ethiopia from 1992 to 2017. The basin was found to have autumn soil moisture (ASM) ranging from 0.09–0.38 m3/m3, with an average of 0.26 m3/m3. The ASM time series resulted in the coefficient of variation (CV) ranging from 2.8%–28% and classified as low-to-medium variability. In general, the monotonic trend analysis for ASM revealed that the UBNB had experienced a wetting trend for the past 26 years (1992–2017) at a rate of 0.00024 m3/m3 per year. A significant wetting trend ranging from 0.001 to 0.006 m3/m3 per year for the autumn season was found. This trend was mainly showed across the northwest region of the basin and covers about 18% of the total basin area. The spatial patterns and variability of rainfall and ASM were also found to be similar, which implies the strong relationship between rainfall and soil moisture in autumn. The spring and autumn season rainfall explained a considerable portion of ASM in the basin. The analyses also signified that the rainfall amount and distribution impacted by the topography and land cover classes of the basin showed a significant influence on the characteristics of the ASM. Further, the result verified that the behavior of ASM could be controlled by the loss of soil moisture through evapotranspiration and the gain from rainfall, although changes in rainfall were found to be the primary driver of ASM variability over the UBNB.

Data Assimilation of High-Resolution Thermal and Radar Remote Sensing Retrievals for Soil Moisture Monitoring in a Drip-Irrigated Vineyard

Efficient water use assessment and irrigation management is critical for the sustainability of irrigated agriculture, especially under changing climate conditions. Due to the impracticality of maintaining ground instrumentation over wide geographic areas, remote sensing and numerical model-based fine-scale mapping of soil water conditions have been applied for water resource applications at a range of spatial scales. Here, we present a prototype framework for integrating high-resolution thermal infrared (TIR) and synthetic aperture radar (SAR) remote sensing data into a soil-vegetation-atmosphere-transfer (SVAT) model with the aim of providing improved estimates of surface- and root-zone soil moisture that can support optimized irrigation management strategies. Specifically, remotely-sensed estimates of water stress (from TIR) and surface soil moisture retrievals (from SAR) are assimilated into a 30-m resolution SVAT model over a vineyard site in the Central Valley of California, U.S. The efficacy of our data assimilation algorithm is investigated via both the synthetic and real data experiments. Results demonstrate that a particle filtering approach is superior to an ensemble Kalman filter for handling the nonlinear relationship between model states and observations. In addition, biophysical conditions such as leaf area index are shown to impact the relationship between observations and states and must therefore be represented accurately in the assimilation model. Overall, both surface and root-zone soil moisture predicted via the SVAT model are enhanced through the assimilation of thermal and radar-based retrievals, suggesting the potential for improving irrigation management at the agricultural sub-field scale using a data assimilation strategy.

Evapotranspiration Data Product from NESDIS GET-D System Upgraded for GOES-16 ABI Observations

Evapotranspiration (ET) is a major component of the global and regional water cycle. An operational Geostationary Operational Environmental Satellite (GOES) ET and Drought (GET-D) product system has been developed by the National Environmental Satellite, Data and Information Service (NESDIS) in the National Oceanic and Atmospheric Administration (NOAA) for numerical weather prediction model validation, data assimilation, and drought monitoring. GET-D system was generating ET and Evaporative Stress Index (ESI) maps at 8 km spatial resolution using thermal observations of the Imagers on GOES-13 and GOES-15 before the primary operational GOES satellites transitioned to GOES-16 and GOES-17 with the Advanced Baseline Imagers (ABI). In this study, the GET-D product system is upgraded to ingest the thermal observations of ABI with the best spatial resolution of 2 km. The core of the GET-D system is the Atmosphere-Land Exchange Inversion (ALEXI) model, which exploits the mid-morning rise in the land surface temperature to deduce the land surface fluxes including ET. Satellite-based land surface temperature and solar insolation retrievals from ABI and meteorological forcing from NOAA NCEP Climate Forecast System (CFS) are the major inputs to the GET-D system. Ancillary data required in GET-D include land cover map, leaf area index, albedo and cloud mask. This paper presents preliminary results of ET from the upgraded GET-D system after a brief introduction of the ALEXI model and the architecture of GET-D system. Comparisons with in situ ET measurements showed that the accuracy of the GOES-16 ABI based ET is similar to the results from the legacy GET-D ET based on GOES-13/15 Imager data. The agreement with the in situ measurements is satisfactory with a correlation of 0.914 averaged from three Mead sites. Further evaluation of the ABI-based ET product, upgrade efforts of the GET-D system for ESI products, and conclusions for the ABI-based GET-D products are discussed.

What Rainfall Does Not Tell Us—Enhancing Financial Instruments with Satellite-Derived Soil Moisture and Evaporative Stress

Advanced parametric financial instruments, like weather index insurance (WII) and risk contingency credit (RCC), support disaster-risk management and reduction in the world’s most disaster-prone regions. Simultaneously, satellite data that are capable of cross-checking rainfall estimates, the “standard dataset” to develop such financial safety nets, are gaining importance as complementary sources of information. This study concentrates on the analysis of satellite-derived multi-sensor soil moisture (ESA CCI, Version v04.2), the evapotranspiration-based Evaporative Stress Index (ESI), and CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data) rainfall estimates in nine East African countries. Based on spatial correlation analysis, we found matching spatial/temporal patterns between all three datasets, with the highest correlation coefficient occurring between October and March. In large parts of Kenya, Ethiopia, and Somalia, we observed a lower (partly negative) correlation coefficient between June and August, which was likely caused by issues related to cloud cover and the volume scattering of microwaves in sandy, hot soils. Based on simple linear and logit regression analysis with annual, national maize yield estimates as the dependent variable, we found that, depending on the chosen period (averages per year, growing or harvesting months), there was added value (higher R-squared) if two or all three variables were combined. The ESI and soil moisture have the potential to close sensitive knowledge gaps between atmospheric moisture supply and the response of the land surface in operational parametric insurance projects. For the development and calibration of WII and RCC, this means that better proxies for historical and potential future drought impact can strengthen “drought narratives”, resulting in a better match between calculated payouts/credit repayment levels and the actual needs of smallholder farmers.

The Potential Utility of Satellite Soil Moisture Retrievals for Detecting Irrigation Patterns in China

Climate change and anthropogenic activities, including agricultural irrigation have significantly altered the global and regional hydrological cycle. However, human-induced modification to the natural environment is not well represented in land surface models (LSMs). In this study, we utilize microwave-based soil moisture products to aid the detection of under-represented irrigation processes throughout China. The satellite retrievals used in this study include passive microwave observations from the Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E) and its successor AMSR2, active microwave observations from the Advanced Scatterometer (ASCAT), and the blended multi-sensor soil moisture product from the European Space Agency (i.e., ESA CCI product). We first conducted validations of the three soil moisture retrievals against in-situ observations (collected from the nationwide agro-meteorological network) in irrigated areas in China. It is found that compared to the conventional Spearman’s rank correlation and Pearson correlation coefficients, entropy-based mutual information is more suitable for evaluating soil moisture anomalies induced by irrigation. In general, around 60% of uncertainties in the anomaly of “ground truth” time series can be resolved by soil moisture retrievals, with ASCAT outperforming the others. Following this, the potential utility of soil moisture retrievals in mapping irrigation patterns in China is investigated by examining the difference in probability distribution functions (detected by two-sample Kolmogorov-Smirnov test) between soil moisture retrievals and benchmarks of the numerical model ERA-Interim without considering the irrigation process. Results show that microwave remote sensing provides a promising alternative to detect the under-represented irrigation process against the reference LSM ERA-Interim. Specifically, the highest performance in detecting irrigation intensity is found when using ASCAT in Huang-Huai-Hai Plain, followed by advanced microwave scanning radiometer (AMSR) and ESA CCI. Compared to ASCAT, the irrigation detection capabilities of AMSR exhibit higher discrepancies between descending and ascending orbits, since the soil moisture retrieval algorithm of AMSR is based on surface temperature and, thus, more affected by irrigation practices. This study provides insights into detecting the irrigation extent using microwave-based soil moisture with aid of LSM simulations, which has great implications for numerical model development and agricultural managements across the country.

Developing a remotely sensed drought monitoring indicator for Morocco

Drought is one of the most serious climatic and natural disasters inflicting serious impacts on the socio-economy of Morocco, which is characterized both by low-average annual rainfall and high irregularity in the spatial distribution and timing of precipitation across the country. This work aims to develop a comprehensive and integrated method for drought monitoring based on remote sensing techniques. The main input parameters are derived monthly from satellite data at the national scale and are then combined to generate a composite drought index presenting different severity classes of drought. The input parameters are: Standardized Precipitation Index calculated from satellite based precipitation data since 1981 (CHIRPS), anomalies in the day-night difference of Land Surface Temperature as a proxy for soil moisture, Normalized Difference Vegetation Index anomalies from MODIS data and Evapotranspiration anomalies from surface energy balance modeling. All of these satellite-based indices are being used to monitor vegetation condition, rainfall and land surface temperature. The weighted combination of these input parameters into one composite indicator takes into account the importance of the rainfall based parameter (SPI). The composite drought index maps were generated during the growing seasons going back to 2003. These maps have been compared to both the historical, in situ precipitation data across Morocco and with the historical yield data across different provinces with information being available since 2000. The maps are disseminated monthly to several main stakeholders groups including the Ministry of Agriculture and Department of Water in Morocco.

An Assessment of the Impact of Land Thermal Infrared Observation on Regional Weather Forecasts Using Two Different Data Assimilation Approaches

Recent studies have shown the unique value of satellite-observed land surface thermal infrared (TIR) information (e.g., skin temperature) and the feasibility of assimilating land surface temperature (LST) into land surface models (LSMs) to improve the simulation of land-atmosphere water and energy exchanges. In this study, two different types of LST assimilation techniques are implemented and the benefits from the techniques are compared. One of the techniques is to directly assimilate LST using ensemble Kalman filter (EnKF) data assimilation (DA) utilities. The other is to use the Atmosphere-Land Exchange Inversion model (ALEXI) as an “observation operator” that converts LST retrievals into the soil moisture (SM) proxy based on the ratio of actual to potential evapotranspiration (fPET), which is then assimilated into an LSM. While most current studies have shown some success in both directly the assimilating LST and assimilating ALEXI SM proxy into offline LSMs, the potential impact of the assimilation of TIR information through coupled numerical weather prediction (NWP) models is unclear. In this study, a semi-coupled Land Information System (LIS) and Weather Research and Forecast (WRF) system is employed to assess the impact of the two different techniques for assimilating the TIR observations from NOAA GOES satellites on WRF model forecasts. The NASA LIS, equipped with a variety of LSMs and advanced data assimilation tools (e.g., the ensemble Kalman Filter (EnKF)), takes atmospheric forcing data from the WRF model run, generates updated initial land surface conditions with the assimilation of either LST- or TIR-based SM and returns them to WRF for initializing the forecasts. The WRF forecasts using the daily updated initializations with the TIR data assimilation are evaluated against ground weather observations and re-analysis products. It is found that WRF forecasts with the LST-based SM assimilation have better agreement with the ground weather observations than those with the direct LST assimilation or without the land TIR data assimilation.

Estimating Morning Change in Land Surface Temperature from MODIS Day/Night Observations: Applications for Surface Energy Balance Modeling

Observations of land surface temperature (LST) are crucial for the monitoring of surface energy fluxes from satellite. Methods that require high temporal resolution LST observations (e.g., from geostationary orbit) can be difficult to apply globally because several geostationary sensors are required to attain near-global coverage (60°N to 60°S). While these LST observations are available from polar-orbiting sensors, providing global coverage at higher spatial resolutions, the temporal sampling (twice daily observations) can pose significant limitations. For example, the Atmosphere Land Exchange Inverse (ALEXI) surface energy balance model, used for monitoring evapotranspiration and drought, requires an observation of the morning change in LST - a quantity not directly observable from polar-orbiting sensors. Therefore, we have developed and evaluated a data-mining approach to estimate the mid-morning rise in LST from a single sensor (2 observations per day) of LST from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor on the Aqua platform. In general, the data-mining approach produced estimates with low relative error (5 to 10%) and statistically significant correlations when compared against geostationary observations. This approach will facilitate global, near real-time applications of ALEXI at higher spatial and temporal coverage from a single sensor than currently achievable with current geostationary datasets.

Water, Energy, and Carbon with Artificial Neural Networks (WECANN): A statistically-based estimate of global surface turbulent fluxes and gross primary productivity using solar-induced fluorescence

A new global estimate of surface turbulent fluxes, latent heat flux (LE) and sensible heat flux (H), and gross primary production (GPP) is developed using a machine learning approach informed by novel remotely sensed Solar-Induced Fluorescence (SIF) and other radiative and meteorological variables. This is the first study to jointly retrieve LE, H and GPP using SIF observations. The approach uses an artificial neural network (ANN) with a target dataset generated from three independent data sources, weighted based on triple collocation (TC) algorithm. The new retrieval, named Water, Energy, and Carbon with Artificial Neural Networks (WECANN), provides estimates of LE, H and GPP from 2007 to 2015 at 1° × 1° spatial resolution and on monthly time resolution. The quality of ANN training is assessed using the target data, and the WECANN retrievals are evaluated using eddy covariance tower estimates from FLUXNET network across various climates and conditions. When compared to eddy covariance estimates, WECANN typically outperforms other products, particularly for sensible and latent heat fluxes. Analysing WECANN retrievals across three extreme drought and heatwave events demonstrates the capability of the retrievals in capturing the extent of these events. Uncertainty estimates of the retrievals are analysed and the inter-annual variability in average global and regional fluxes show the impact of distinct climatic events - such as the 2015 El Niño - on surface turbulent fluxes and GPP.

Water, Energy, and Carbon with Artificial Neural Networks (WECANN): A statistically-based estimate of global surface turbulent fluxes and gross primary productivity using solar-induced fluorescence

A new global estimate of surface turbulent fluxes, latent heat flux (LE) and sensible heat flux (H), and gross primary production (GPP) is developed using a machine learning approach informed by novel remotely sensed Solar-Induced Fluorescence (SIF) and other radiative and meteorological variables. This is the first study to jointly retrieve LE, H and GPP using SIF observations. The approach uses an artificial neural network (ANN) with a target dataset generated from three independent data sources, weighted based on triple collocation (TC) algorithm. The new retrieval, named Water, Energy, and Carbon with Artificial Neural Networks (WECANN), provides estimates of LE, H and GPP from 2007 to 2015 at 1° × 1° spatial resolution and on monthly time resolution. The quality of ANN training is assessed using the target data, and the WECANN retrievals are evaluated using eddy covariance tower estimates from FLUXNET network across various climates and conditions. When compared to eddy covariance estimates, WECANN typically outperforms other products, particularly for sensible and latent heat fluxes. Analysing WECANN retrievals across three extreme drought and heatwave events demonstrates the capability of the retrievals in capturing the extent of these events. Uncertainty estimates of the retrievals are analysed and the inter-annual variability in average global and regional fluxes show the impact of distinct climatic events - such as the 2015 El Niño - on surface turbulent fluxes and GPP.