新疆阿克苏河流域降水空间变异特征分析  

Research on the Spatial Variability of Rainfall in Akesu River Basin

作　　者：蒋艳[1] 周成虎[1] 程维明[1] 

机构地区：[1]中国科学院地理科学与资源研究所资源与环境信息系统国家重点实验室,北京100101

出　　处：《地球信息科学》2006年第1期131-138,共8页Geo-information Science

基　　金：国家自然科学杰出青年基金"空间数据挖掘与知识发现"(40225004);世行合作项目"新疆塔里木河流域水量调度管理决策支持系统"(THSD-7)

摘　　要：根据阿克苏河流域降水空间观测数据,其降水稀疏且分布不均匀的特点,选取不同模型对降水空间变化规律进行研究,其结果精确性差异很大。通常应用地统计理论研究降水空间变异性,一般只涉及单个变量,传统的多元回归分析虽然涉及多个变量的影响,但缺乏区域化的空间结构特征。揭示具有协同区域化特征的降水空间变异现象及建立其空间分布模型,既要考虑多元信息的空间位置关系,即同一变量在不同地理位置上的相关性,又要考虑多元信息由于空间重复性引起的协同关系,即同一地理位置上不同变量的相关性。本文用阿克苏河流域范围内的降水观测数据建立析取-协克立格模型,考虑高程变量对降水量空间分布的影响,定量地揭示降水区域化变量的空间变异规律,并将其结果用于降水量的空间最优插值。The spatial variability of rainfall is a major problem in its description and prediction. Akesu river basin lies in the semi-arid and arid region of northwestern China. The local climate and topographic factors affect the magnitude and distribution of rainfall. Rainfall has very low total magnitude in the southeast region, but in the northwest region comparatively more precipitation events occur. Fifteen representative stations were selected for years to reflect the regional rainfall patterns throughout the region. The study area lies between 40°00′-42°00′N and 76°00′-82°00′E. The network of rainfall gauging stations in the southwest is sparse and the available data are insufficient to characterize the highly variable spatial distribution of rainfall in this mountain area. Therefore, in areas where data are not available, it is necessary to develop methods to estimate rainfall using data from the surrounding measuring stations. The major goal of this study is to characterize the spatial variability of annual rainfall by suitable variogram models, which are then used in the kriging process for assigning values to ungauged locations for compiling mean rainfall isoline map. The multiple-regression model for predicting rainfall results in many prediction errors because this kind of model considers precipitation to be independent without spatial distribution pattern and mutual interdependence. Geostatistics, which is based on the theory of regionalized variables, is increasingly preferred in hydrology and meteorology because it allows one to capitalize on the spatial correlation between neighboring observations to predict attribute values at unsampled locations. More and more cases have shown that the geostatistical prediction technique provides better estimates of rainfall than conventional methods. Kriging is a geostatistical estimation technique for regionalized variables that exhibit an autocorrelation structure. Kriging algorithm, based on unbiased and minimum-variance estimates, involves a linear sy

关 键 词：阿克苏河流域 地统计模型 空间变异 半变异函数 

分 类 号：P333[天文地球—水文科学] TV87[水利工程—水文学及水资源]