Integrated evaluation of remotely sensed water balance components and in-situ river discharge for basin-scale hydrological assessment

Abstract

Understanding basin-scale water balance dynamics is essential for sustainable water resource management in semi-arid, monsoon-driven river systems. This study evaluates the spatio-temporal variability of water yield in the Bhima River Basin using integrated satellite-derived datasets, including precipitation from CHIRPS, GPM, SSEBop and GRACE. Annual water yield was derived by integrating precipitation, evapotranspiration, and GRACE-based terrestrial water storage changes within a water balance framework. The remotely sensed outflow estimates were further evaluated against observed discharge records to assess model consistency and performance. Results indicate pronounced spatial heterogeneity in water yield across the basin. The upper sub-basin consistently generated higher runoff, attributed to relatively greater rainfall and favorable runoff-producing conditions. In contrast, the middle and lower reaches exhibited persistently low water yield, primarily due to high evapotranspiration losses and limited surface runoff generation. Comparative evaluation of precipitation-driven water balances showed that CHIRPS-based estimates exhibited substantially stronger agreement with observed discharge (r = 0.86) compared to GPM (r = 0.41). The superior performance of CHIRPS was further confirmed by a higher coefficient of determination (R² = 0.74) and a positive Nash–Sutcliffe efficiency (NSE = 0.56), whereas GPM demonstrated weaker predictive capability.