Data release for Remotely Sensed Surface Water Storage Shows Distinct Patterns from SWAT-Simulated Data (ver. 2.0, February 2026)

Understanding and projecting the downstream benefits of terrestrial surface water storage, i.e., volumetric water stored in lakes and wetlands (SWstorage) requires watershed hydrologic models. Use of external datasets to calibrate and validate modeled SWstorage dynamics remains uncommon, particularly across major river basins. Here, we: (1) develop and assess the utility of a novel remote sensing-based (RS) SWstorage approach for verifying model simulated SWstorage estimates and (2) compare simulated and RS SWstorage volumes across the landscape, representing both average conditions and changing conditions through time. These comparisons used SWstorage from Sentinel-1 and -2 (RS SWstorage) and simulated SWstorage from the Soil and Water Assessment Tool (SWAT; SWAT SWstorage) across the ~440,000 km2 Upper Mississippi River Basin. Average RS SWstorage across the entire basin was 0.86 km3 (5.7%) lower than SWAT SWstorage, and time series of SWstorage were only positively correlated in 38.8% of studied subbasins. Our results show that SWAT SWstorage, initialized using only digital elevation model data, does not well represent landscape water distribution, especially in agricultural and wetland-rich regions. While SWstorage-based model calibration is uncommon, we find that calibration to discharge alone may not adequately simulate changes in SWstorage through time. This new RS SWstorage dataset could be used to improve surface water parameterization and calibration in SWAT or similar process-based hydrological models.

First release: 2025 (ver. 1.0)
Revised: February 2026 (ver. 2.0)