中文摘要：

利用TRMM卫星上携带的闪电探测仪（LIS）所获取的9年闪电资料（1998-2006年）对华南地区闪电活动的时空分布特征进行了分析。该地区闪电次数的年差异较大，最多年份是最少年份的2倍多，闪电活动季节性特征非常明显，闪电主要集中发生在春末仲夏，呈现双峰值特征，4—8月是闪电高发期（约占全年总闪电活动的81．91％）。闪电活动的日变化表明，8月份闪电活动绝大多数发生在午后至傍晚时分，这也与对流活动相对应，5月份闪电活动除在午后有一个峰值区外，在凌晨也有一个不小的峰值区。华南地区的闪电密度高值区主要有：清远-广州一带、廉江市附近、海南岛中部，闪电密度低值区主要位于南海水面上。分析表明：华南地区闪电时空分布除与大尺度的天气系统背景有关外，还与太阳辐射、地形抬升、下垫面的性质等有关。太阳辐射的季节变化和日变化等是造成闪电时间分布的重要原因；地形的抬升作用和下垫面的性质及其差异是造成气候意义上中小尺度闪电空间分布差异的重要原因。

英文摘要：

The temporal and spatial distributions of lightning activities in South of China are studied by using nine years （from 1998 to 2006） of observation data from the Lightning Imaging Sensor （LIS） on board the Tropical Rainfall Measurement Mission （TRMM） satellite. The result indicates that the number of lightning flashes captured by LIS varies year by year, and the maximum is about twice as many as the minimum. The lightning activities clearly show seasonal variation in South of China. Late spring through midsummer is the main lightning activity season. The seasonal variety of the flash number is a pair of peak value, about 81.19% of the lightning flashes of a year occurs frequently between April and August. It is shown that lightning activities in August primarily occur between afternoon and early evening, and lightning activities in May mostly occur after midday, followed by another substantial peak around midnight. Lightning flash density centers are found mainly over the Qingyuan and Guangzhou areas, Lianjiang area and central Hainan Island. Fewer lighting flashes are found over the waters of South China Sea. Results show that the temporal and spatial distributions of lightning activities in South of China are closely related to both synoptic climatic background and solar radiation, the topography and underlying surface, the features of surface, conditions of the atmosphere and so on. The temporal change in solar radiation brings about that of lighting activity. The physical features of surface and topography may be major causes for the difference in the spatial distribution of lightning on mesoscale.