初始场中尺度信息对暴雨预报的影响

THE EFFECT OF INITIAL MESOSCALE MESOSCALE FEATURE ERRORS ON THE PREDICTION OF PRECIPITATION

由于观测资料分辨率与模式分辨率的不同,利用高分辨率模式对暴雨进行预报时,常规观测资料形成的初始场不能直接分辨出中尺度系统,这种中尺度系统特征的缺少可以认为是初始场的一种信息误差。利用中尺度天气分析的尺度分离方法可以提取这种中尺度信息。通过分析初始场中尺度信息的结构、演变特征及其对暴雨预报影响的机理,发现初始场中尺度信息的结构在主要雨带的对流敏感区具有明确的天气学意义,包含了有利暴雨产生的信息;其能量随时间也是增长的,特别是在积分12小时以后,能量迅速增长然后趋于稳定,超过了初始随机扰动的能量增长。利用减弱和增强初始场中中尺度信息的两种初始场作暴雨预报,其结果反映了初始场中尺度信息对暴雨预报的重要性,特别是对雨团位置和强度的预报,这些信息会直接影响暴雨的精细预报。

Because of the existence of resolution difference between observational data and model analysis, mesoscale information cannot be directly indentified in the traditional observation. The lack of mesoscale features can be regarded as information errors of the initial field, which can be extracted through scale decomposition. The structure and evolvement of such errors, along with their effect on the prediction of precipitation, are discussed in this paper. It indicates that the errors possess certain dynamic structures which are responsible for the rainfall in convective regions. The energy growth of meso-scale feature errors is faster than that of random errors after 12-h model integration and saturates at larger amplitude. The simulations of different initial fields, which are initialized by adding and omitting meso-scale feature errors respectively, reveal the significant effect of initial meso-scale features on the prediction of precipitation. Such effect is very important to precise prediction, especially to the forecast of rain-cell position and rain intensity.