Future Satellite Gravimetry for Global Total Drainable Water Storage Estimation: Mission Performance and Basin Controls 

Abstract

Total Drainable Water Storage (TDWS) represents the portion of basin water storage that can drain and sustain river discharge, providing a basin-scale indicator of freshwater availability. Reliable global estimation of TDWS could therefore improve assessments of freshwater availability and help identify basins with limited capacity to sustain discharge. However, estimating TDWS globally from space-based Terrestrial Water Storage Anomalies (TWSA) remains constrained by both satellite-observation errors and differences in basin hydrological behavior. We use closed-loop satellite gravimetry simulations incorporating realistic instrumental and aliasing errors to estimate TDWS for mission configurations representative of the Gravity Recovery and Climate Experiment Continuity (GRACE-C), the Next Generation Gravity Mission (NGGM), and the Mass-change and Geosciences International Constellation (MAGIC). Assuming a linear runoff–storage (R–S) relationship, TDWS is estimated for more than 200 globally distributed river basins. Two independent Earth system model realizations, ESA ESM2.0 and ESM3.0, are used as reference TWSA datasets. Mission-based TDWS estimates are evaluated against their corresponding references, and retrieval errors are examined in relation to basin area, directional shape, storage and discharge variability, terrain relief, land cover, climate, and R–S linearity. NGGM and MAGIC consistently improve TDWS estimation relative to GRACE-C. Median relative errors decrease from 7.9% for GRACE-C to 4.2% and 4.3% for NGGM and MAGIC, respectively, under ESM2.0, and to 4.5% and 4.1% under ESM3.0. The largest size-dependent improvements occur for basins between 200,000 and 500,000 km², with error reductions of about five percentage points relative to GRACE-C. Directional basin geometry also affects retrieval performance, with improvements increasing toward strongly north–south-oriented basins under ESM3.0. Environmental controls remain evident: rugged basins generally show larger TDWS errors than flat basins, while under ESM3.0 MAGIC reduces the median error in polar climates from 10.0% to 4.4%. The reference R–S relationship achieves a Nash–Sutcliffe Efficiency (NSE) of at least 0.7 in 86% of basins under ESM2.0 and 79% under ESM3.0. NGGM and MAGIC reproduce the reference NSE more closely than GRACE-C, demonstrating the benefit of improved gravity-field recovery. Nevertheless, improved satellite observations cannot compensate for a weak or unsuitable R–S relationship. These simulations therefore provide more than a comparison of future mission performance: they identify the observational, geometric, and hydrological conditions that control basin-scale TDWS retrieval. Moving toward robust global TDWS estimation will require both next-generation satellite gravimetry and basin-specific assessments of hydrological retrieval suitability.