中国降水序列均一性研究及对比分析

文章采用全国2415个站1951-2009年逐日降水量资料，采用标准正态均一性检验（SNHT）方法对年序列进行均一性检测，得到存在不连续的断点的站点数为114个，占所有检测站点数的4．72％;对其中70个与对比站相关系数在0．7以上断点的逐年、逐月和逐日降水序列进行了订正：分析发现，订正对冬季降水序列的影响比对夏季序列要大。将订正前后资料按不同的条件分为山区、台站稀疏地区等进行了分析比较，发现订正前后降水序列的变差系数各站差别不大，说明订正的结果是可信的；降水距平对比分析结果表明，多数台站订正前后值差异很小，台站稀疏地区亦是如此，山区站降水量订正前后值差异略大；对均一化前后降水序列趋势分析结果表明，序列趋势增减并不大，多数增减值占原值的比例在10％以下，年及绝大多数月份降水序列订正前后趋势的增减符号是一致的，台站稀疏地区亦是如此，山区站趋势增减幅度略大，即山区迁站对降水量趋势影响较大，但总体上并没有影响到趋势的显著性。

Homogeneity Study and Comparison Analysis on Precipitation Series over China

Based on the precipitation data during 1951-2009 for all the 2415 stations in China, an SNHT (Standard Normal Homogeneity Test) method is used to test the homogeneity of annual data series. The result shows that about 114 stations contain data discontinuities, which covers about 4.72 percent of the total. In the following step, the inhomogeneous series at the 70 stations where the candidate series have good correlation with the reference stations with correlation coefficients above 0.7, have been adjusted in annual, monthly and daily time scales. Further comparison and analysis show that, the adjustment effects on winter precipitation series are much more than those on summer precipitation series. Finally the raw and adjusted data over plain, mountainous and station denser or sparser areas have been analyzed, and the result shows that the adjustment is reasonable. The variation coefficients, precipitation anomalies and linear trends have not changed greatly for the most of the adjusted series, with an exception in mountainous areas, the changes are slightly larger.