Inflow Forecast Using Downscaled Rainfall From Global NWP For Real-Time Flood Control

Abstract

Many reservoirs are operational with single objective like hydropower reservoirs. Its objective function is to maximize water availability for power generation. While objective functions related to vulnerability, for example flood control, are not prioritized or at very limited capability. In many cases, improper reservoir operation may produce extra flood threats downstream because of dam safety issues. This paper presents inflow forecast for real-time flood control based on combination of downscaled precipitation from the global NWP model using local scaling method and a distributed rainfall runoff model. Altitudinal dependence of rainfall is utilized to determine the basin average scaling factor. Results show that inflow prediction based on precipitation forecast using the local scaling factor exhibits significant improvement. It has demonstrated an added value in reservoir inflow prediction for flood control practices, especially in unobserved, remote catchments, that aims to reduce downstream flood.