泾河流域气温时序变异特征识别

Detecting Temporal Variations of Temperature Characteristics in Jinghe Watershed

摘要: 气温变异不仅呈现时序上总体升高的趋势性，而且呈现出不同时段内升高趋势及变化幅度的差异。本文基于施瓦兹信息准则的突变检验法，分析泾河流域4个基本气象站最高气温、最低气温、平均气温均值、方差以及趋势项线性变化的变异特征。研究结果表明：各气象站年最高、最低、平均气温均呈上升趋势，突变发生时间大都在20世纪90年代中期，但在变点前后各站气温变异类型存在差异。与Mann-Kendall方法相比，基于施瓦兹信息准则的突变检验不仅可以检测出突变发生的时间，而且可以定量描述变点前后气温变化特征。

Abstract: Temporal variation of temperature not only presents overall rising trend, but also presents differences both the rising trend and range in different period. Variations characteristics of mean, variance and linear trend term changes in maximum, minimum and mean temperatures of four basic meteorological stations in the Jinghe Watershed were detected using change-point test method based on Schwarz information criterion. The research results indicate that the annual maximum, minimum and average temperatures in every meteorological station show increasing trends, the mutation of temperature occurred mainly in the mid-1990s and the mutation type present some differences before and after the change. Compared with Mann-Kendall method, the proposed method that based on Schwarz information criterion not only can test the mutation of the time, but also can quantitatively describe the temperature variation characteristics.