Abstract Submission: Accurate flood forecasting has long been the dream of the world hydrological community, and the physically based, distributed hydrological model(PBDHM) has been regarded as the new generation hydrological model for flood forecasting. Unfortunately, several challenges, such as the data acquisition for model structure building, model parameter recognition, computing power, and others, limited its application in real-time flood forecasting, thus constraining its application in scientific studies mainly. In this study, several progresses have been presented which makes the PBDHMs more appropriate for real-time flood forecasting. First, model parameter optimization concept and theory for PBDHMs have been proposed, and the Particle Swarm Optimization (PSO) algorithm is developed for Liuxihe model which was proposed for watershed flood forecasting. Dozens of watersheds have been studied, and it has been found that parameter optimization is not only necessary but also critical for PBDHMs to improve their performances. It also has been found that hydrological data from only one flood event is enough for this purpose, not as currently doing in lumped hydrological model parameter calibration, which requires hydrological data from dozens of flood events. This finding has made PBDHMs parameter optimization feasible for most of the world watersheds. To improve the computation efficiency, parallel computing algorithm has been presented and developed for Liuxihe model, which not only makes the parameter optimization feasible, but also makes it applicable for real-time flood forecasting as it provides rapid results. The successful implementation of Liuxihe model for real-time watershed flood forecasting has proven that PBDHMs has entered a new era for the real world application.
