2003年汛期淮河流域降水的集合预测试验研究

Experiment of Ensemble Forecast of Heavy Rainfall in the Huaihe River During Rainy Season of 2003

利用并行化全球中期数值预报模式MPGM,对2003年夏季淮河流域特大降水作了集合预报试验。试验中提出一种繁殖循环初始扰动生成方法,产生9个集合成员的初值场,利用该初值集合对这次降水过程进行了集合模拟,然后对集合预报试验的结果同对照试验进行了对比分析,并对降水期间的物理量场及降水情况的模拟结果进行了分析。结果表明:集合预报试验的结果优于不加扰动的对照试验的结果,说明集合预报技术在一定程度上能够减小甚至消除数值预报中的不确定性,提高数值预报的精度;集合预报试验较好地模拟出了降水期间的物理量场与天气形势,并对此次降水情况作了较好的预报,可以对降水预报提供有益的参考。

An ensemble forecast experiment and a control run of the heavy rainfall event in Huaihe River area during June and July 2003 were performed using MPGM model. A method of generating the breeding cycle initial perturbation has been proposed and 9 ensemble numbers of initial data were produced. Then, the results of these two experience have been compared and analyzed. The simulated result of parameter field and precipitation also has been discussed. The results show that ensemble forecast seems to be better than control run, this could imply the technology of ensemble forecast could reduce the uncertainty of numerical prediction and improve the precision of numerical prediction. Meanwhile, the results also show that ensemble prediction is inspiring, it can provided useful reference to the improvement of heavy rainfall prediction.