Although soil moisture is a key factor of hydrologic and climate applications, global continuous high resolution soil moisture datasets are still limited. Here we use physics-informed machine learning to generate a global, long-term, spatially continuous high resolution dataset of surface soil moisture, using International Soil Moisture Network (ISMN), remote sensing and meteorological data, guided with the knowledge of physical processes impacting soil moisture dynamics. Global Surface Soil Moisture (GSSM1 km) provides surface soil moisture (0–5 cm) at 1 km spatial and daily temporal resolution over the period 2000–2020. The performance of the GSSM1 km dataset is evaluated with testing and validation datasets, and via inter-comparisons with existing soil moisture products. The root mean square error of GSSM1 km in testing set is 0.05 cm³/cm³, and correlation coefficient is 0.9. In terms of the feature importance, Antecedent Precipitation Evaporation Index (APEI) is the most important significant predictor among 18 predictors, followed by evaporation and longitude. GSSM1 km product can support the investigation of large-scale climate extremes and long-term trend analysis.

Inland surface water is highly dynamic, seasonally and inter-annually, limiting the representativity of the water coverage information that is usually obtained at any single date. The long-term dynamic water extent products with high spatial and temporal resolution are particularly important to analyze the surface water change but unavailable up to now. In this paper, we construct a global water Normalized Difference Vegetation Index (NDVI) spatio-temporal parameter set based on the Moderate-resolution Imaging Spectroradiometer (MODIS) NDVI. Employing the Google Earth Engine, we construct a new Global Surface Water Extent Dataset (GSWED) with coverage from 2000 to 2018, having an eight-day temporal resolution and a spatial resolution of 250 m. The results show that: (1) the MODIS NDVI-based surface water mapping has better performance compared to other water extraction methods, such as the normalized difference water index, the modified normalized difference water index, and the OTSU (maximal between-cluster variance method). In addition, the water-NDVI spatio-temporal parameter set can be used to update surface water extent datasets after 2018 as soon as the MODIS data are updated. (2) We validated the GSWED using random water samples from the Global Surface Water (GSW) dataset and achieved an overall accuracy of 96% with a kappa coefficient of 0.9. The producer’s accuracy and user’s accuracy were 97% and 90%, respectively. The validated comparisons in four regions (Qinghai Lake, Selin Co Lake, Utah Lake, and Dead Sea) show a good consistency with a correlation value of above 0.9. (3) The maximum global water area reached 2.41 million km² between 2000 and 2018, and the global water showed a decreasing trend with a significance of P = 0.0898. (4) Analysis of different types of water area change regions (Selin Co Lake, Urmia Lake, Aral Sea, Chiquita Lake, and Dongting Lake) showed that the GSWED can not only identify the seasonal changes of the surface water area and abrupt changes of hydrological events but also reflect the long-term trend of the water changes. In addition, GSWED has better performance in wetland areas and shallow areas. The GSWED can be used for regional studies and global studies of hydrology, biogeochemistry, and climate models.

Accurate information on surface soil moisture (SSM) content at a global scale under different climatic conditions is important for hydrological and climatological applications. Machine-learning-based systematic integration of in situ hydrological measurements, complex environmental and climate data, and satellite observation facilitate the generation of reliable data products to monitor and analyse the exchange of water, energy, and carbon in the Earth system at a proper space–time resolution. This study investigates the estimation of daily SSM using 8 optimised machine learning (ML) algorithms and 10 ensemble models (constructed via model bootstrap aggregating techniques and five-fold cross-validation). The algorithmic implementations were trained and tested using International Soil Moisture Network (ISMN) data collected from 1722 stations distributed across the world. The result showed that the K-neighbours Regressor (KNR) had the lowest root-mean-square error (0.0379 cm³ cm⁻³) on the “test_random” set (for testing the performance of randomly split data during training), the Random Forest Regressor (RFR) had the lowest RMSE (0.0599 cm³ cm⁻³) on the “test_temporal” set (for testing the performance on the period that was not used in training), and AdaBoost (AB) had the lowest RMSE (0.0786 cm³ cm⁻³) on the “test_independent-stations” set (for testing the performance on the stations that were not used in training). Independent evaluation on novel stations across different climate zones was conducted. For the optimised ML algorithms, the median RMSE values were below 0.1 cm³ cm⁻³. GradientBoosting (GB), Multi-layer Perceptron Regressor (MLPR), Stochastic Gradient Descent Regressor (SGDR), and RFR achieved a median r score of 0.6 in 12, 11, 9, and 9 climate zones, respectively, out of 15 climate zones. The performance of ensemble models improved significantly, with the median RMSE value below 0.075 cm³ cm⁻³ for all climate zones. All voting regressors achieved r scores of above 0.6 in 13 climate zones; BSh (hot semi-arid climate) and BWh (hot desert climate) were the exceptions because of the sparse distribution of training stations. The metric evaluation showed that ensemble models can improve the performance of single ML algorithms and achieve more stable results. Based on the results computed for three different test sets, the ensemble model with KNR, RFR and Extreme Gradient Boosting (XB) performed the best. Overall, our investigation shows that ensemble machine learning algorithms have a greater capability with respect to predicting SSM compared with the optimised or base ML algorithms; this indicates their huge potential applicability in estimating water cycle budgets, managing irrigation, and predicting crop yields.

Land surface energy, water and carbon fluxes are key for understanding Earth's climate system, yet global continuous high resolution fluxes datasets remain scarce. In this study, we present a global long-term (2000–2020) hourly 9 km dataset of terrestrial water-energy-carbon fluxes, generated by integrating model simulations, in-situ measurements, and machine learning with remote sensing and meteorological data. First, the integrated STEMMUS-SCOPE model was deployed to simulate land surface fluxes over 170 sites with in-situ measurements. The selected model-output variables include net radiation (Rn), latent heat flux (LE), sensible heat flux (H), soil heat flux (G), gross primary productivity (GPP), solar-induced fluorescence at 685 and 740 nm (SIF685, SIF740). Next, optimal interpolation was applied to merge Rn, LE, and H from STEMMUS-SCOPE simulations with eddy covariance observations. The optimal estimate of Rn, LE, H alongside STEMMUS-SCOPE simulated G, GPP, SIF685, SIF740 were then used as training data-pairs to develop the emulator using a multivariate Random Forest (RF) regression algorithm, referred to as Random Forest with Optimal Interpolation (RF_OI) to predict terrestrial water-energy-carbon fluxes. The results demonstrate that RF_OI can estimate land surface fluxes with Pearson Correlation Coefficient score (r-score) values higher than 0.88 except for GPP (Rn 0.99, LE 0.88, H 0.92, G 0.92, GPP 0.8, SIF685 0.99, SIF740 0.99). The testing results on independent stations (which were not included for developing the emulator) show r-score values higher than 0.8. The feature importance indicates that incoming shortwave radiation, surface soil moisture, and leaf area index are top predictor variables that determine the prediction performance. FluxHourly enables analysis of ecosystem responses to climate extremes at unprecedented spatiotemporal scales. FluxHourly is available at https://doi.org/10.11888/Terre.tpdc.302319 (Han et al., 2025a).