华北地区夏季降雨量与南海海温长期变化的关系

比较了华北地区7个站与17个站1951-1997年夏季（6，7，8月）降雨量与气候随时间的变化特征，并对其成因作了探讨。结果表明，用北京、天津、邢台、烟台、郑州、太原和济南等7个站可代表该地区夏季降雨量与气候的多尺度变化特征，过去47a该地区依次经历了湿凉、湿热、湿凉、干热、湿热几个时期，降雨量的长期变化与南海前冬（1-2月）海温成负相关。前冬南海海温偏高，意味着初夏南海地区大气对流低频振动偏弱，南海夏季风爆发较晚，西南季风较弱，夏季西太平洋副高位置偏南，华北地区大气低层北风加强，华北地区夏季少雨，前冬南海海温偏低时情况则相反，考虑冬季（1-2月）南海南温和7-8月西太平洋副高脊线位置（纬度）的影响用均生函数建模，试验结果与用子波变换重构方法考虑华北地区夏季降雨量的变化趋势比较，二者相吻合，预测试验结果与过去3a的实况基本一致。

RELATIONSHIP BETWEEN LONG-TERM CHANGES OF SUMMER RAINFALL IN NORTH CHINA AND SEA SURFACE TEMPERATURE OVER SOUTH CHINA SEA

The changes of summer (June, July and August) rainfall and climate with lime at 7 stations were compared with those at 17 stations in North China for the period from 1951 to 1997. The causes of the changes were also explored. It was found that the changes of summer rainfall and climate at 7 stations (Beijing, Tianjin, Xingtai, Yantai, Zhengzhou, Taiyuan and Jinan) can represent the multi-scale characteristics of those in North China, which has experienced wet cool, wet hot, wet cool, dry hot and wet hot periods successively during the last 47 years. The long term variation of summer rainfall is negatively related to the sea surface temperature (SST) over the South China Sea (SCS) during early winter (January and February). The higher SST than normal over SCS during early winter implies a weak low-frequency oscillation of the atmosphere over SCS in early summer, late onset time of SCS summer monsoon, weak southwest monsoon, southward location of western Pacific subtropical high in summer, strong north wind in the lower atmosphere layer and less rainfall in North China. The situation is on the contrary when SST over SCS is lower than normal during early winter. The test results, considering the affections of SST over SCS during winter (January and February) and location (latitude) of western Pacific subtropical high in July and August and by u-sing mean generating function to develop the mode, are consistent with the long-term trend of summer rainfall in North China using reconstruction method of wavelet transforms. The forecast results based on the former are basically consistent with the observation results in the last three years.