Evaluating SMAP L4 root-zone soil moisture against hydrologically simulated soil-water states in the upper Kelani River Basin, Sri Lanka

Abstract

Soil moisture regulates the conversion of monsoon rainfall into runoff, making it significant for flood forecasting. Satellite products such as the Soil Moisture Active Passive (SMAP) mission provide soilmoisture estimates but remain poorly evaluated in data-scarce tropical monsoon regions, particularly Sri Lanka. This study evaluates the consistency of SMAP L4 Version 7 root-zone soil moisture with soil-water states simulated by a discharge-calibrated Hydrologic Engineering Center–Hydrologic Modeling System (HEC-HMS) model of the upper Kelani River Basin above Hanwella (1,818 km2). A continuous 11-subbasin HEC-HMS model used a 5 km grid, Gridded Deficit and Constant Loss, and SoilGrids-based Saxton and Rawls parameters. It was calibrated against daily discharge for 2020 and validated for 2021 (Nash–Sutcliffe 0.70, R 0.86, PBIAS −8%). Simulated soil-moisture deficit was converted to volumetric soil moisture and compared with 9 km SMAP using anomaly, day-to-day changes, lag, and flood-event analyses. Direct agreement between the two series was weak (R = 0.31; negative Nash–Sutcliffe), partly reflecting grid-scale mismatch between SMAP and HEC-HMS across steep terrain and canopy attenuation of retrievals under dense Wet Zone vegetation. Correspondence was stronger after removing the mean (R = 0.55) and for dayto- day changes (R = 0.39); the catchment-averaged series correlated at 0.58, increasing to 0.72 at a 17-day lag. Overall, SMAP captured aspects of sub-seasonal variability despite limited agreement in absolute magnitude, suggesting greater suitability for monitoring soil-moisture anomalies and wetness trends than directly updating HEC-HMS states in tropical basins.