初始扰动对一次华南暴雨预报的影响的研究

本文选取了2006年华南前汛期的一次暴雨过程, 采用AREMv2.3中尺度数值模式进行数值模拟, 分别在模式初始场的物理量场 (温度场、 风场、 湿度场) 上加扰动, 分析不同物理量场上的扰动对降水预报的影响, 以及物理量预报误差和扰动能量的增长情况。同时, 通过本个例讨论误差增长与湿对流的关系, 扰动振幅对误差增长的影响和华南区域的中尺度降水的可预报性问题。数值试验结果表明: 初始时刻不同物理量场加实际振幅的正态分布的随机扰动时, 对降水的影响是不同的。对于24小时降水预报, 温度场对降水的影响最大。误差的增长与湿对流不稳定有着密切的关系。小尺度小振幅误差增长很快, 而且是非线性增长。这意味着短期的较小尺度降水的可预报性很小。与大振幅扰动相比, 小振幅扰动造成的误差较小。但是小振幅扰动的迅速发展, 很快就会对降水预报造成较大的影响。因此, 只能有限地提高预报质量, 而且由于扰动非线性增长很快, 在预报时间的提前上, 不会有太大的改善。

Impacts of Initial Perturbations on Prediction of a Heavy Rain in South China

A rainstorm process in South China during the pre-rainy season of 2006 has been selected in this article. AREMv2.3 mesoscale numerical model is used to simulate the process through adding perturbations to the initial fields of physical variables (temperature, humidity, and wind). The authors analyze the impacts of perturbations of different physical variable fields on the precipitation forecast, the error growth of physical variables, and the perturbation energy growth. And the authors also discuss the relationship between the error growth and moist convection, and the impact of perturbations of different amplitudes on the error growth, and the predictability of the mesoscale precipitation in South China. The numerical simulation results show that: the impacts on precipitation are different when adding the perturbations of the actual amplitude normal distribution at initial time to different physical variable fields. For 24-hour precipitation forecast, the temperature perturbation has the greatest impact on precipitation. Error growth and moist convective instability are closely related. The small-scale and small-amplitude initial errors increase rapidly, furthermore in the form of nonlinear growth. This means that there is less predictability in short-term smaller-scale precipitation forecast. Compared with the larger-amplitude perturbations, smaller-amplitude perturbations can cause smaller error. But smaller-amplitude perturbations develop rapidly and will soon make a great impact on the precipitation forecast. Therefore, it can only improve the prediction quality limitedly, and because the perturbations are in quick nonlinear growth, there will not be much improvement that can be done to bring forward the prediction time.