基于EOF的陕西省降水变化时空分异研究

Analyzing Spatial and Temporal Distributions of Precipitation in Shannxi Province Using Empirical Orthogonal Function

应用经验正交函数法(empirical orthogonal function,EOF)对陕西省23 a来96个气象站点的数据进行EOF分解,分别计算出年降水场的特征向量分布和时间系数序列。结果表明,应用EOF方法可以很好地揭示降水场的时空分布特征,且对陕西省而言前3个特征向量揭示了3种典型的分布场,其累计贡献率达81.10%。陕西省降水的空间分布有3种典型的模式,即全局型、南北型和东西型。总体上3种场的特征向量分布值表现为南大北小,且高值中心都出现在陕西省的最南端,这说明南部虽然多雨,但年际变化量较大,北方少雨,但年际变化量较小。对特征向量所对应的时间系数分析发现,陕西省的降水场主要表现为6种类型。从23 a来的统计结果可以看出它与贡献率所反映的降水场的典型程度基本一致。

Using the method of empirical orthogonal function（EOF） and the data of 23 years at 96 meteorological stations in Shaanxi Province,we calculated the feature vector distribution of precipitation fields and the time-coefficient series.The results showed that the EOF method was capable of revealing the spatial and temporal distributions of precipitation field for Shaanxi Province,with the first three EOFs reflecting the three typical vector distribution fields and explaining 81.10% of total variation.The spatial distribution of precipitation in Shaanxi Province had three typical modes： the whole region,north-south,and east-west patterns.In general,the values of the three feature vector distribution fields increased from north to south,with the largest value centered on the southern tip of the province,indicating a wetter south with a larger inter-annual precipitation variation and a drier north with a smaller inter-annual precipitation variation.Analysis of the eigenvectors of the corresponding time factor coefficients showed that the precipitation fields in Shaanxi Province mainly exhibited six types.Compared to the statistical results of 23 years,these results were consistent with the typical precipitation fields.