概率匹配平均法在山东强降水预报中的应用

Applications of probability matching method in heavy rainfall forecast in Shandong province

基于WRF集合预报系统开发了概率匹配平均降水产品,选取了山东省2014—2016年共13次强降水过程,检验评估了概率匹配平均法在山东省强降水预报中的综合表现。结果表明:对于不同的强降水过程,各预报产品的预报能力差异较大,尤其是对暴雨以上量级降水的预报存在较大偏差;概率匹配平均相对集合平均,对大雨以上量级降水预报有明显改善,较WRF确定性预报产品也有一定提高,对强降水预报具有一定指示意义;该方法的改进主要体现在对不同量级降水的调整上,尤其是强降水的落区,相对集合平均增大了强降水的范围和强度,但对整个区域的总降水量预报没有很好的改进作用。

Based on WRF ensemble forecast system,a product of precipitation forecast is developed using probability matching method. Comprehensive evaluations are carried out for thirteen heavy rainfall processes in Shandong province from 2014 to 2016. The results show that skills of different methods are significantly different in precipitation forecast for intense rainfall processes,especially for rainstorms; The probability matching precipitation forecast have a significant improvement compared to the ensemble average products for heavy rainfalls,and it is also better than that of the WRF deterministic forecasts,indicating certain guiding significance in heavy rainfall forecast. Improvements are made by probability matching method in precipitation forecast of all levels,especially in adjusting the area of heavy rainfall.Probability matching products improve the area and intensity forecast of heavy rainfall,but can barely improve the total precipitation.