汉江流域1960~2014年降雨极值时空变化特征

深入研究降雨极值的时空变化规律和特点，有助于提高应对极端灾害的能力。通过构建PDS/GP模型，并引入降雨极值变化指标，结合统计检验分析汉江流域15个气象站点1960~2014年春、夏、秋季以及全年月降雨超定量系列的年际变化特征；利用复杂网络理论的聚类系数和节点度，结合各站点在典型年份春、夏和秋季的降雨极值变化指标，分析月降雨极值的空间分布特征。分析结果表明：从时间上看，汉江流域月降雨极值年际变化的季节性差异较大，近55 a没有明显的一致性变化趋势。受季风气候影响，夏季的月降雨极值年际变化与春、秋季相反，而秋季的变异指标年际波动最大。从空间上看，受季风气候、下垫面条件和人类活动等多种因素的综合作用，中下游月降雨极值变化的差异性要高于上游。随着相关阈值的增大，流域站网总体关联度有所降低。相同阈值下的结果表明空间上相邻的站点其关联性差异较大，部分相距较远的站点具有更大的关联性，分析结果可为降雨极值的空间插值提供参考。

TEMPORAL AND SPATIAL VARIATION OF RAINFALL EXTREMES IN HAN RIVER BASIN FROM 1960 TO 2014

A deeper understanding of the spatiotemporal variation of rainfall extremes will help improve the ability in response to extreme disasters. The PDS/GP model for Han River basin was constructed and the Rainfall Extremes Variation Index (REVI) was proposed to analyze the annual variation of monthly rainfall extremes for the different seasons and full year from 1960 to 2014 by combing the statistical test. The study was based on a 55-year partial duration series of rainfall data from 15 gauging stations in Han River basin. Meanwhile the clustering coefficient and node degree based on the network theory was employed to study the spatial distribution on the basis of REVI in typical years for different seasons. The results show that the seasonal difference of monthly rainfall extremes was quite remarkable in Han River basin, and there were no uniform trends in recently 55 years. Among the three seasons, the situation in summer is quite opposite to that in spring and autumn, while the highest annual fluctuation was present in autumn because of monsoon climate. Spatially, by the comprehensive effects of monsoon climate, underlying surface conditions and human activities, the most dramatic change can be found in the upper reach of Han River. With increasing correlation threshold, the connectivity between stations in the whole basin decreased, which was larger in the upstream than in the middle and lower reaches of Han River. For the same correlation threshold, the results also indicated that even nearest stations had significantly different connectivity, and even some distant stations had very similar connectivity properties. The results can provide a reference for spatial interpolation of rainfall extremes.