An error updating system for real-time flood forecasting based on robust procedure

An updating technique is a tool to update the forecasts of mathematical flood forecasting model based on observed real-time data, and is an important element in a flood forecasting. The quality of the final updating results depends primarily on the quality of observed data. However, the input flow to reservoirs, or so called “observed input flow”, is usually calculated by the observed water stage, outflow, and stage-storage relation curve. In this process, coarse errors which have great negative influence on updating are often created. This paper introduces a robust theory in real-time forecasting updating, and proposes an error updating system with robust procedure to limit the influence of coarse errors. Cases of ten reservoirs are studied to test the proposed system. Results indicate that the proposed robust procedure deflates the influence of outliers. The results of real-time updating with robust procedure in flood forecasting are better than those without robust procedure.