基于多元线性回归的东北地区春玉米单产空间化模拟  

Spatialization of Spring Maize Yield Area in Northeast China Based on Multiple Linear Regression

作　　者：赵雪晴 金涛[2] 董雯怡 刘美霞 刘勤[1] 刘恩科[1,2,3] ZHAO Xue-qing;JIN Tao;DONG Wen-yi;LIU Mei-xia;LIU Qin;LIU En-ke(Institute of Environment and Sustainable Development in Agriculture,Chinese Academy of Agricultural Sciences,Beijing 100081,China;Tibet Academy of Agricultural and Animal Husbandry Sciences,Lhasa 850030;National Dryland Agriculture Science and Technology Innovation Alliance,Beijing 100081)

机构地区：[1]中国农业科学院农业环境与可持续发展研究所,北京100081 [2]西藏自治区农牧科学院,拉萨850030 [3]国家旱地农业科技创新联盟,北京100081

出　　处：《中国农业气象》2023年第11期1022-1031,共10页Chinese Journal of Agrometeorology

基　　金：国家重点研发计划“政府间国际科技创新合作重点专项”项目(2021YFE0101300)。

摘　　要：春玉米是东北地区主要粮食作物,研究分析不同气候和土壤模式下春玉米产量的空间差异对于指导农业生产和保障粮食安全具有重要意义。本研究选取土壤、地形和气候因素三个方面共13个关键影响因素,采用多元线性回归分析法构建春玉米单产和13个影响因素的多元线性回归模型,并利用Arc GIS软件将春玉米单产进行栅格化,对东北地区春玉米单产空间分布情况进行分析。结果表明:(1)经多元线性回归模型计算的空间化数据(空间分辨率为1km)与春玉米单产统计数据基本一致,东北地区春玉米单产在2482.49~10147.10kg·hm^(-2)。(2)春玉米单产空间分布图客观反映了春玉米单产的空间分布趋势,大体呈现由中部地区向四周不断减少的格局。本研究准确取得栅格尺度的东北地区春玉米单产空间化的模拟结果,平均相对误差1.45%,可为东北地区农业生产布局优化和决策制定等提供方法参考。Spring maize is the primary food crop in Northeast China.Researching and analyzing the spatial difference in yield under diverse climate and soil models holds immense significance in guiding agricultural production and ensuring food security.In this study,a total of 13 key influencing factors were selected from the three aspects of soil,topography and climate,and a multivariate stepwise linear regression model was constructed for spring maize per unit yield and 13 influencing factors by using multivariate stepwise regression analysis.Then the spatial distribution of spring maize per unit area yield in Northeast China was analyzed by using ArcGIS software to rasterize the spring maize yield per unit area.The results showed that:(1)the spatialized data(with a spatial resolution of 1km)calculated by the multiple linear regression model are basically consistent with the statistical data of spring maize yield per unit area.The spring maize yield in Northeast China is between 2482.49-10147.10kg·ha^(-1).(2)The spatial distribution map of spring maize yield objectively reflects that the spatial distribution trend of spring maize yield,which generally shows a pattern of decreasing from the central to the surrounding areas.This study accurately obtained the grid-scale simulation results of spring maize yield per unit area spatialization in Northeast China (the average relative error is 1.45%), which provides a method reference for the optimization of agricultural production layout and decision-making in Northeast China.

关 键 词：玉米单产 空间化 多元线性回归 东北地区 

分 类 号：S513[农业科学—作物学] S162.5