Boosted Regression Tree Algorithm for the Reconstruction of GRACE-Based Terrestrial Water Storage Anomalies in the Yangtze River Basin

The terrestrial water storage anomaly (TWSA) from the previous Gravity Recovery and Climate Experiment (GRACE) covers a relatively short period (15 years) with several missing periods. This study explores boosted regression trees (BRT) artificial neural network (ANN) to reconstruct TWSA series between 1982 2014 over Yangtze River basin (YRB). Both algorithms are trained hydro-climatic variables (e.g., precipitation, soil moisture, temperature) climate indices for YRB. results this show that BRT is capable of reconstructing shows Nash–Sutcliffe efficiency (NSE) 0.89 root-mean-square error (RMSE) 18.94 mm during test stage, outperforming ANN in about 2.3% 7.4%, respectively. As step further, reliability technique beyond GRACE era was also evaluated. Hence, closed-loop simulation using 1982–2014 under same scenarios actual data can predict (NSE 0.92 RMSE 6.93 mm). Again, outperformed by approximately 1.1% 5.3%, provides new perspective filling gaps GRACE–TWSA data-scarce regions, which desired hydrological drought characterization environmental studies. offers such an opportunity Follow-On mission 11 months relying on limited number predictive variables, hence being adjudged be more economical than ANN.