Use of Integrated Observations to Improve 0-36 h Flood Forecasting: Development and Application of a Coupled Atmosphere-Hydrology System in the Nanpan River Basin, China

Abstract

The accuracy of 0-36 h real-time flood forecasting is largely determined by the quantitative precipitation forecasts (QPFs), but convective weather remains a significant challenge for numerical weather prediction systems. Therefore, it is crucial to improve the QPFs’ accuracies to predict and prevent flash flood disasters.A coupled atmosphere-hydrology system with the WRF model (together with a three-dimensional variational data assimilation system, 3DVAR) and a distributed biosphere hydrological model (WEB-DHM) is described. This system was then applied to the flood forecasting of the Nanpan River Basin (Yunnan province, China) for 1 July 2008. Based on the available observations (230 surface meteorological sites, 10 conventional Radiosonde sites, and 8 ground-based GPS stations), a series of experiments were conducted with the WRF-3DVAR to investigate the contributions of different observations to QPF accuracy and flood forecasting. Forced with the observations or the WRF model outputs, WEB-DHM predicts stream-flow at the basin outlet. The overall better performance by the assimilation experiments over the no assimilation case has been clearly demonstrated in the predictions of the 0-36 h heavy rainfall (magnitude and spatial pattern) and flash flood occurrence ( peak value and time). The WRF-3DVAR only assimilating GPS data performs poorly, showing the necessity to improve both the assimilation technique and the spatial resolution for the operational numerical weather forecasts. To our knowledge, this work is the first to utilize comprehensive observations around the Tibetan Plateau with a coupled atmosphere-hydrology system to improve short-term flood predictions.