Combining Optical Remote Sensing, McFLI Discharge Estimation, Global Hydrologic Modeling, and Data Assimilation to Improve Daily Discharge Estimates Across an Entire Large Watershed

Remote sensing has gained attention as a novel source of primary information for estimating river discharge, and the Mass-conserved Flow Law Inversion (McFLI) approach successfully estimated discharge in ungauged basins solely from optical satellite data. However, McFLI currently suffers two major drawbacks: (1) existing satellites lead to temporally spatially sparse estimates (2) because assumptions required, cannot guarantee downstream flow continuity. Hydrological modeling neither drawback, yet model accuracy is frequently limited by lack observations. We therefore combine models data assimilation framework applicable globally. establish daily “ungauged” baseline 28,998 reaches Missouri basin forced recently published global runoff data, which we do not calibrate. estimate via using ?1 million width measurements made 12,000 Landsat scenes assimilate into before validating at 403 USGS gauges. Results show that assimilated discharges did impair already accurate flows achieved median improvements 28% normalized root mean square error, 0.50 Nash–Sutcliffe efficiency (NSE), 0.23 Kling–Gupta where performance was poor (defined negative NSE, 225/403 reaches). ultimately improved 92% these originally poorly modeled gauges, even though images only provide 1.5% 26% simulated days. Our results suggest combination state-of-the-art hydrology can improve estimations