Assimilating SWOT Level-4 river discharge into a macroscale hydrologic model
Abstract
Accurate quantification of river discharge remains an important challenge as limited ground observations and hydrological model uncertainties constrain our ability to monitor freshwater resources at global scales. The Surface Water and Ocean Topography (SWOT) mission, launched in December 2022, provides unprecedented satellite-based discharge estimates that offer the potential to improve hydrological modeling, particularly in data-sparse regions where traditional gauge networks are inadequate. Here, we evaluate the effectiveness of assimilating SWOT Level-4 river discharge data into the Variable Infiltration Capacity (VIC) hydrological model using a modified Local Ensemble Transform Kalman Filter (LETKF). Our study area was the Ohio River Basin over a 1-year period from October 2023 to September 2024, encompassing 54 river reaches with valid SWOT observations. The modified LETKF incorporates bias correction and Huberization techniques to address inherent errors in SWOT discharge estimates and demonstrated superior performance compared to standard assimilation approaches, improving the Kling-Gupta Efficiency and Nash-Sutcliffe Efficiency at 90% of gauges with median values changing from 0.22 and 0.21 to 0.26, respectively. An idealized observation experiment revealed that observation frequency limitations can significantly influence assimilation effectiveness beyond measurement errors alone. Overall, SWOT discharge observations appear to demonstrate potential for enhancing hydrological model performance through data assimilation, particularly when implemented with bias-aware filtering techniques.
