摘要: 本文以广东北江流域18测站1965~2007年日降水数据为基础，选取6个极端降水指标，应用REOF方法对极端降水场进行分区，利用M-K方法进行变化趋势分析，应用变差系数对长期变化的稳定程度及相对变化特征进行分析，采用滑动检验法进行变异年份分析，应用Morlet小波分析振荡周期。结果表明：R1d、R5d、R95T、CDD呈上升趋势，CWD、R95p呈下降趋势；极端降水的长期变化具有较大的波动性，山区的波动性较平原地区更大；极端降水指数在20世纪90年代初期发生变异的现象较为普遍，显著周期尺度在1990年前后存在明显差异，这些都可能与该区20世纪90年代之后经济高速发展从而造成人类活动的影响较之前更加强烈有关。
Abstract:
The paper selected 6 extreme precipitation indices base on the daily data of 18 stations in Beijiang River basin from 1965 to 2007. The River basin was divided into three different extreme precipitation spatial-temporal characteristic subareas by REOF method. The non-parametric method of Mann-Kendall trend test is applied to analysis variation trend. The change points of extreme precipitation indices were analyzed by the moving t-test method, the long-term changes in stability and relative variation were analyzed by variation coefficient, and the cycle characteristics were analyzed by the Morlet wavelet transform. Results show that, R1d, R5d, R95T and CDD were on the rise, while CWD and R95p declined; the change of extreme precipitation in long-term has large variability, and mountain area more than plain; the extreme indices were common mutation in the early 1990s, significant change cycle around 1990 scale difference, all of these may due to the high-speed economic development after 1990, and the influence of human activity more intense than before.