摘要:
长江流域是中国重要的水能基地，水患灾害频繁，研究该流域的极端降水事件的分布及变化特征，对预测及应对旱涝灾害影响有重要意义。本文利用长江流域1961~2010年最近50年逐日降水量资料，分析了4个极端降水量指标和2个极端降水频率指标的时空变化及趋势变化特征，特别地利用Logistic回归模型，分析了2个极端降水频率指标的非线性趋势特征。结果表明：整个长江流域各种极端降水量指标的空间差异明显，大值区分布在长江流域上游四川盆地向青藏高原的过渡区和中下游地区，极端最大值分布在流域中几个典型的高山测站；最大1日降水量和超90%阈值降水总量随时间有显著的线性增大现象，但是区域平均年降水量和极端过程降水随时间没有明显的线性趋势；长江上游和中游一带极端频率指标中存在显著非线性增大趋势，四川盆地和下游地区极端频率的非线性趋势不明显。极端降水量大的区域，极端指标出现频率也很高，对年降水量贡献很大，是产生洪涝灾害的重要原因。
Abstract:
The Yangtze River basin is the most important hydropower base in China; however, drought and flood occurred frequently in this region recently. It is important to investigate extreme events for predicting and coping with the drought and flood disasters. On the basis of daily precipitation (1961-2010, 138 meteorological sites) for the Yangtze River basin, the temporal and spatial varia-tion of 6 indices of extreme rainfall is analyzed. What’s more, the nonlinear characteristics of 2 extreme rainfall frequency indices are analyzed by Logistic Regression Model. The results show that the spatial differences between the 6 indexes are obvious. The extreme values for the indices distribute in the transition zone from the Sichuan basin to the Qinghai-Tibet plateau and the middle to the lower reach areas of the Yangtze River basin, and the extreme maximum values distribute in several typical mountain stations. There is no obvious linear trend for the mean annual rainfall and extreme duration rainfall; however, the maximum 1-day rainfall and the total rainfall exceeding 90% threshold increase significantly. Nonlinear increasing trend of extreme frequency indices exists in the upstream and middle reaches of the Yangtze River, while it is not significant in the Sichuan basin and the downstream reaches. In a word, there are more extreme indices in districts with more extreme rainfall, contributing a lot to annual rainfall. This is the main reason of flood disaster.