Developing a Long-Term Global Dataset of Water Storage Anomalies Using GRACE and Model-Based Estimations

The Gravity Recovery And Climate Experiment (GRACE) satellite mission has significantly advanced the remote sensing of total water storage anomalies (TWSA) from regional to continental scales. Building on this foundation, the GRACE Follow-On (GRACE-FO) mission, launched on 22 May 2018, continues to provide valuable data. However, the combined observational period of GRACE and GRACE-FO is limited to two decades of monthly data, with a one-year gap between the missions. This relatively short record limits the ability to observe global and regional climate trends over the long term, which is essential for studies on drought characterization and long-term climate change patterns.To overcome this limitation, using global hydrological, atmospheric, and reanalysis models, we have developed a new gridded TWSA dataset that expands the temporal coverage of GRACE(-FO) observations back to 1980. We employed various machine learning algorithms to combine these models and reconstruct TWSA, including Multivariate Linear Regression (MLR), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), and Gaussian Process Regression (GPR) together with some ensemble methods. Comparisons were made with GRACE data during the GRACE period (Apr 2002 to Dec 2012) with a high-resolution Satellite Laser Ranging (SLR) TWSA product before the GRACE period (1992-2002).Our findings demonstrate significant improvements compared to the basic approaches, highlighting the necessity for advanced and sophisticated methods to reconstruct TWSA accurately across diverse regions and climates. GPR and MLR exhibited superior performance among the tested methods, while SVM and DT displayed poorer performance in most basins. This research presents a new approach for reconstructing long-term total water storage anomaly fields prior to the GRACE period (i.e., before 2002). The newly developed dataset substantially extends the timeline of TWSA observations, providing valuable insights into long-term variations in Earth's water storage.