基于Co-kriging的呼伦贝尔市2018年降水量空间分布特征模拟及插值精度分析  

Simulation of Spatial Distribution Characteristics of Precipitation in Hulunbuir City in 2018 and an Analysis of Interpolation Accuracy Based on Collaborative Kriging Model

作　　者：任宾宾[1,2] 王平[1] 邱绍柱 杜卫红[2] 白絮飞[2] REN Binbin;WANG Ping;QIU Shaozhu;DU Weihong;BAI Xufei(Faculty of geography,Yunnan Normal University,Kunming 650500,China;College of Tourism and Geographical Sciences,Hulunbuir University,Hailaer Inner Mongolia 021008,China;Bureau of Nature and Resources of Manzhouli city,Manzhouli Inner Mongolia 021400,China)

机构地区：[1]云南师范大学地理学部,昆明650500 [2]呼伦贝尔学院旅游与地理学院,内蒙古海拉尔021008 [3]满洲里自然资源局,内蒙古满洲里021400

出　　处：《西南大学学报（自然科学版）》2021年第11期162-171,共10页Journal of Southwest University(Natural Science Edition)

基　　金：内蒙古自治区教育厅项目(NJSY11229);呼伦贝尔学院校级项目(2020ZKYB16).

摘　　要：因降水发生随机性强、连续性差,获得高精度、高时空分辨率的降水空间数据一直是地球科学等相关领域研究的难点和热点.内蒙古呼伦贝尔市地处温带半湿润、半干旱气候区,降水是该区域关键生态因子和稀缺资源,降水空间模拟具有重要意义.本研究根据该市16个国家气象站点2018年逐日降水数据,采用协同克里金(Co-kriging)法模拟了年、季降水量空间差异;根据125个检测站点降水量相对误差分析了空间插值精度及其影响因素.结果表明:①降水量空间差异有统计学意义,自东向西,年降水量由750 mm左右逐渐降低至220 mm左右,等降水量线延伸方向与大兴安岭走向基本一致,夏、秋季降水量空间分布与之相似;②降水量空间插值精度年大于季,夏、秋季大于冬、春季,年降水量预测值与实测值之比近于1(p<0.001),平均相对误差MRE较小(11.4%);③年降水量空间插值精度与降水频率呈正相关(p=0.023),与经度、海拔分别呈极显著正、负相关关系(p<0.001),与之相似,越向东,夏、秋季降水量空间插值精度越高,海拔越高,插值精度越低;降水量较少的季节(冬、春、秋季),降水空间插值精度受降水频率影响较大,且由Co-kriging模型导致的“平滑效应”明显;④通过Co-kriging插值法得到的年、夏季降水空间数据可靠,可供站点稀少的类似区域降水空间插值研究借鉴.Precipitation data with high precision and spatio-temporal resolution are an important parameter in geoscience and related fields.Because of the strong randomness and poor continuity of rainfall,the spatial interpolation of precipitation has always been a focus and difficulty in the study of the spatialization of climatic factors.Hulunbuir city enjoys a vast area and is located in a semi-humid and semi-arid climate region,where rainfall is a key ecological factor and a scarce resource.Affected by the East Asia monsoon and topography,there are large spatio-temporal differences of precipitation in the region.Due to the scarcity of meteorological stations in the area,the spatial simulation of precipitation is of great significance.In the present study,based on the daily precipitation data,starting from January 1 to December 31,2018,of 16 meteorological stations in Hulunbuir city,the Collaborative Kriging(Co-Kriging)Model is adopted to simulate the spatial distribution characteristics of annual and seasonal precipitations.On the basis of the relative errors(RE),which are calculated from the measured and predictive values,in precipitation at the test stations,the spatial interpolation accuracy of annual and seasonal precipitations is analyzed.Because of the randomness of rainfall and the influences of macro-geographical factors,such as longitude,latitude and altitude,the impacts on spatial interpolation accuracy of precipitation caused by precipitation frequency are analyzed.The results show that Co-Kriging method can describe the spatial differences of annual precipitation in Hulunbuir.From the east to the west of the region,the annual precipitation gradually decreases,approximately,from 750 mm to 220 mm.The annual precipitation isoline is roughly parallel to the northeast-southwest range of the Daxinganling Mountains.The annual precipitation occurs mostly in summer and autumn.The spatial distribution of precipitation in summer and autumn is similar to that of annual precipitation.When time resolution gets shorter,th

关 键 词：呼伦贝尔市 降水量 协同克里金 插值精度 降水频率 宏观地理因子 

分 类 号：P332.1[天文地球—水文科学]