Potential of incorporating satellite soil moisture observations in flood modeling in Kelani river basin, Sri Lanka

Abstract

Floods are frequent and devastating natural disasters, causing significant social and economic losses worldwide. Flood modelling plays a crucial role in predicting and mitigating the impacts of floods. In recent years, satellite remote sensing technology has emerged as a valuable tool for flood modelling, especially in estimating soil moisture, a critical parameter for flood forecasting. The potential of incorporating satellite soil moisture observations in flood modelling within the highly flood-prone Kelani River Basin up to Glencourse in Sri Lanka was investigated in this study. The study utilizes ESA CCI Soil moisture data and employs the Pearson correlation coefficient to evaluate the linear correlation between the modified Antecedent Precipitation Index (API) and the European Space Agency's Climate Change Initiative for Soil Moisture (ESA CCI SM). The Hydrologic Engineering Center's Hydrologic Modelling System (HEC-HMS) is taken as the preferred hydrological model, with the Green-Ampt loss method, Clark's unit hydrograph method, and recession constant baseflow method chosen for the lumped daily resolution event-based model. Twelve flood events are carefully selected for model calibration and validation. Events EI to E6 are used for calibration, while events E7 to E12 are utilized for model validation. Nash-Sutcliffe efficiency (NSE), percent bias (PBIAS), and the ratio of root mean square error to the standard deviation of measured data (RSR) is employed to assess the predictive accuracy of the model in comparison to observed streamflow. The findings indicate a robust association between the modified API and the soil moisture data from ESA CCI SM. The calibrated HEC-HMS model exhibits an NSE of 0.76, an RSR value of 0.45, and a PBIAS of 5% during calibration, while validation yields an average NSE of 0.726, RSR of 0.51, and a PBIAS of 6%. The incorporation of ESA CCI SM data leads to marginal improvements in model performance for some events while showing negative impacts for others. The study reveals that incorporating ESA CCI SM observations results in minor enhancements in the model's predictive accuracy. Furthermore, the integration of ESA CCI SM data has the potential to improve the accuracy of total stream flow predictions and peak flow predictions. These findings contribute to the advancement of flood modelling techniques, providing valuable insights for flood mitigation efforts in the Kelani River Basin and similar flood-prone regions globally.