基于WRF模式的好溪流域致洪暴雨千米尺度预报研究

Research on Forecast of Flood Rainstorm at Kilometer Scale in Haoxi Basin Based on WRF Model

鉴于好溪流域内多山地和丘陵,梅汛期和台汛期常突发强暴雨,致使其千米尺度下的数值预报难度较大的问题,选取好溪流域2015~2020年的10场致洪暴雨事件,利用中尺度数值天气预报模式WRF和GFS预报数据进行6、30、54 h的回顾性预报,并测试了5种云微物理参数化方案的敏感性。结果表明,在千米尺度下,WRF模式模拟的降雨量偏小,对山区的预报效果优于河谷地区;Lin方案的整体预报效果最佳,而WSM3方案最差;选取的10场极端降水事件受到台风和西南低空急流等异常天气的影响,Lin方案和WSM5方案适用性较好;WRF模式对暴雨的预报效果优于大暴雨。研究结果对于提高山区小流域致洪降雨的模拟和预报精度具有借鉴意义。

Given the presence of mountains and hills in the Haoxi River Basin,as well as the frequent occurrence of intense rainstorms during the plum rain season and typhoon season,there are significant challenges in accurately forecasting rainfall at a kilometer scale.This study selected 10 flood-causing rainstorm events in the Haoxi River Basin from 2015 to 2020 and conducted retrospective forecasts for 6 hours,30 hours,and 54 hours using the mesoscale numerical weather prediction model WRF and GFS forecast data.The sensitivity of five cloud microphysics parameterization schemes was tested.The results show that at the kilometer scale,the rainfall simulated by the WRF model is underestimated,and its forecasting performance is better for mountainous regions than for valley areas.The Lin scheme performs the best overall,while the WSM3 scheme performs the worst.The selected ten extreme precipitation events are influ-enced by abnormal weather conditions such as typhoons and southwest low-level jet streams,with the Lin scheme and WSM5 scheme being more applicable.The WRF model performs better in forecasting heavy rain compared to extreme rain.This study provides valuable insights for improving the simulation and forecasting accuracy of flood-causing rainfall in small watersheds in mountainous areas.