中文摘要：

以广东北江流域18测站1965～2007年的月降水数据为基础，统计发现该区域降水前汛期突出，后汛期不明显，年内降水呈明显的“单峰型”分布。Mann-Kendall趋势检验表明，4月、5月降水主要呈下降趋势，6月降水主要呈上升趋势。考虑到前期海温及遥相关指数对降水具有预测性的指示意义，本文将流域前汛期各月份降水经验正交函数（EOF）分析的第一模态时间系数与同期及前期1～12月全球海温及10种遥相关指数进行时滞相关。结果发现影响前汛期各月降水的海温异常关键区并不局限于广东近海，各月的海温异常关键区、遥相关指数及其影响时段存在较大差异。海温异常关键区面积各月均表现为负相关海域大于正相关海域；4月、5月正相关海域主要分布在南半球，6月正相关海域主要分布在北太平洋海域；有些海区在滞后0～12月均与前汛期降水相关显著。研究有助于减少广东北江流域前汛期降水预报的不确定性，为指导工农业生产及防灾减害提供一定的科学依据，也可为研究区域气候变化规律及其影响因素提供一种思路。

英文摘要：

Statistics show that there was more precipitation in the first rainy season （April, May, June）, but the second rainy seasonal （July, August, September） precipitation is usually not obvious, and the yearly variation of precipitation is obvious unimodal distribution, base on the monthly precipitation data of 18 stations in Guangdong Beijiang river basin from 1965 to 2007. The precipitation in April and May were mainly down- ward trend, but mainly upward trend in June from 1965 to 2007 by Mann-Kendall test. Based on the signifi- cance of early sea surface temperature （SST） and teleconnection indices to the prediction of precipitation, the paper analyses the correlation between the PC 1 of EOF in the first rainy seasonal precipitation and the same pe- riod and early one to twelve months globe SST and 10 teleconnection indices. The results showed that SST anomaly key areas are not only distributed in China＇ s neighboring seas; SST anomaly key areas and telecon- nection indices and its key periods are larger differences among months in first rainy season. SST anomaly key area of all months showed that sea areas of negative correlation are bigger than the positive correlation. The positive correlational sea areas of April and May are mainly distributed in the southern hemisphere, while June mainly distributed in the North Pacific Ocean. The negative correlational sea areas of April mainly distributed in northem hemisphere; and the negative correlational sea areas of May mainly distributed in the East Atlan- tic-the Mediterranean Sea, northwest Pacific Ocean, South America＇ s southern waters, Australia＇ s western and southem waters; while June mainly distributed in Australia-centered sea areas. Some sea areas are signifi- cantly related to the first rainy seasonal precipitation at the same time and lag one to twelve months. The statis- tics showed that precipitation in the first rainy season correlated significantly with PNA, AAOI and NAO; and PNA was positive correlation, AAOI was negative correlation.