Abstract
Abstract. Dynamic downscaling is the best way to get high resolution quantitative precipitation forecasts. The Weather Research and Forecasting Model (WRF) was utilized as a dynamic downscaling tool in this study. The influence of the horizontal resolutions and the model domains on precipitation forecasts has been analysed to establish an optimized dynamic downscaling scheme. Three precipitation events over Xijiang basin, China, were simulated with different horizontal resolutions and domains. The results indicate that both the horizontal resolution and model domain have an influence on the precipitation forecast. However, the correlation between high precipitation forecast accuracy and the high horizontal resolution or the large model domain were not very strong. Comprehensive consideration of the results shows that the accuracy of forecast is best when the horizontal resolution is 20-km. Although the model domain size has no significant influence on the precipitation forecast, a larger domain may improve the stability of forecasting.
Highlights
China is located in the eastern part of Asia and has complicated topography, and its climate varies greatly from region to region
Further increase of the horizontal resolution leads to the occurrence of false rainstorm centres
A similar conclusion was found by Shi et al (2012) utilizing Weather Research and Forecasting Model (WRF) to analysis the influence of horizontal resolution (27 km, 9 km) on precipitation and temperature forecasts; the results were better when the horizontal resolution was 27 km
Summary
China is located in the eastern part of Asia and has complicated topography, and its climate varies greatly from region to region. The South of China is greatly affected by the monsoon climate and tropical cyclones, and flood disasters take place frequently. Flood forecasting, as one of the nonstructural flood protection measures, is capable of reducing the risk of flood disaster effectively. Precipitation is the most important atmospheric input to hydrological models, and is always acquired by gauge-based precipitation measurements. Utilizing observed precipitation data as input to hydrological models would limit the flood lead time. Compared with traditional methods (Lu et al, 2006), using quantitative precipitation forecasts with a certain accuracy as input can be an effective way to provide longer flood lead times
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
