Abstract: Groundwater circulation is a crucial component of the global terrestrial water cycle, and accurately describing basin  groundwater runoff is essential for understanding the groundwater circulation process within a basin. Using hydrometeorological remote sensing data, this study estimates the groundwater runoff released by precipitation in a basin by constructing memory curve models for evaporation and surface runoff, integrating them with the water balance equation, and conducting parameter uncertainty analysis. Taking the Jequitinhonha River basin in Brazil as the study area, the results indicate that the memory curve model can effectively simulate the processes of evaporation, surface runoff, and terrestrial water storage change in the study area. The simulated evaporation values fit well with GLEAM-based evaporation observations, with R² values of 0.74 and 0.81 for the calibration and validation periods, respectively. The simulated surface runoff values had R² values of 0.69 for the calibration period and 0.63 for the validation period when compared to observations. The simulated total water storage change (TWSC) values had R² values of 0.73 and 0.6 for the calibration and validation periods, respectively, when compared to the GRACE TWSC series. Furthermore, based on the groundwater runoff memory curve model, the estimated monthly cumulative groundwater runoff released by precipitation in the study area fluctuated between 50 and 350 m³/s, showing a clear intra-annual variation pattern. This study provides a new approach for estimating basin groundwater runoff in large-scale areas lacking groundwater observation data.

Key words: memory curve, MCMC method, water balance equation, groundwater runoff volume