The impacts of ingesting and updating soil moisture-based loss coefficients on HEC-HMS-based reservoir inflow prediction

Abstract:

HEC-HMS is a modeling system widely adopted by reservoir operators for producing inflow forecasts. As the model is often configured for event-based simulations, the accuracy of prediction depends closely on the judicious choice of loss factors specified ahead of each event. This work explores the use of real-time, satellite-infused soil moisture products as the basis for deriving and adjusting loss factors in order to improve the predictive accuracy of inflow. In particular, we seek to extend an existing regression approach to include subbasin-based soil moisture anomalies as the predictors of loss factors. To assess the impacts of the extension, we perform control experiments in which hindcast is issued using the Army Corps of Engineers HEC-HMS configurations for six inflow points in the Upper Trinity River basin. In the experiment, three sets of loss factors are used, namely the static default loss factors, and loss factors based on the baseline and extended version of the regression approach. The presentation summarizes the preliminary findings of the experiment and outlines the next steps.