动力协调的扰动场在风暴尺度集合预报中的效果改进试验

The Skills of Dynamicly Coordinated Initial Perturbation in Storm-Scale Ensemble Forecasing

利用WRF模式,对1977年5月20日发生在美国俄克拉荷马州的超级单体风暴进行理想化的集合预报试验。基于具有明确统计意义的蒙特卡洛扰动法和扰动变量动力协调性的考虑,构造了两个集合预报系统(EPS),一个是由包含多时次扰动误差信息并能够使各个扰动变量与动力模式相协调的初始扰动场生成集合成员;一个是由蒙特卡洛扰动法生成集合成员。对比检验了两个EPS的性能,结果表明:与动力模式相协调的扰动场形成初始扰动的EPS预报效果要好于蒙特卡洛扰动法产生的EPS;Talagrand分布和离散度情况表明前者可以获得更适宜的离散度,模式对初始扰动的平滑作用更小,比后者更为合理。

Ideal storm case is simulated by WRF model on the supercell severe local storm,named the Del City storm,occurred in central Oklahoma on 20 May 1977 to find the better initial perturbation method.This paper designs two ensemble forecast system： one is perturbated with dynamicly coordinated initial perturbation;the other is with Monte Carlo methods.Through comparing the skills between the two EPSs,it is found that the first EPSs can gain higher forecast skills and more reasonable spread than the second one.