一次典型梅雨锋暴雨过程的多尺度结构特征

利用2006—2014年4—9月湖南117例强降雨天气过程高空、地面观测资料，依据湖南省暴雨预报经验和方法，将强降雨天气过程分为低涡冷槽型、地面暖倒槽锋生型、副高边缘型、台风型、梅雨锋切变型和华南准静止锋型6种类型。在此基础上，以同期222个暴雨日当天08时20°—35°N、105°—120°E范围内500 hPa高度场和850 hPa径向风为参数，采用K-均值聚类法，通过反复迭代得到6类暴雨日客观天气型。结果表明:第1、2、6类暴雨日天气形势场特征和强降雨落区分别有60%以上与台风型、低涡冷槽型、地面暖倒槽锋生型强降雨过程的特征吻合；第3、4、5类暴雨日分别出现在2种或以上强降雨天气过程期间，其天气形势场和强降雨特点对应强降雨过程的不同阶段；6类聚类结果较为客观地反映了湖南省汛期暴雨的天气形势和降雨特点，分型结果可作为湖南汛期暴雨预报客观分型的参考依据。

Mesoscale convective systems

In this study, both subjective and objective classification techniques are employed to identify features of synoptic weather patterns for 117 severe precipitation events in Hunan Province between April and September from 2006 to 2014. The subjective classification is based on the operational experience accumulated by forecasters in Hunan in the past. With the scheme above, we have classified weather patterns of the 117 severe precipitation events into six types, i.e., low vortex with cold trough pattern, frontogenesis in the warm ground inverted trough pattern, edge of subtropical high pattern, typhoon pattern, Meiyu front with shear line pattern, and South China quasi-stationary front pattern. Based on these subjective patterns, we collect meridional wind at 500 hPa and 850 hPa between 20°N and 25°N, 105°E and 120°E at 08:00 BT for the 222 rainstorm days from the 117 severe precipitation events as parameters, and then apply the K-means clustering method and the iterative repetition to obtain six types of objective synoptic patterns of flood-season severe precipitation in Hunan. It is shown that more than 60% of synoptic situation and the severe rainfall distribution for the first, second and sixth type of severe precipitation days corresponded well with those of typhoon pattern, low vortex with cold trough pattern and frontogenesis in the warm ground inverted trough pattern severe precipitation. The third, fourth and fifth type of severe precipitation days occur during two or more patterns of severe precipitation events, and their synoptic situation and severe precipitation distribution corresponded to the different stages of the severe precipitation events. The 6 types of objective clustering results can well represent the synoptic situation and rainfall feature of rainstorms in flood season in Hunan, and also can be used as an important basis for objective classifications of rainstorm.