杨佳佳, 吕骏超, 金珊合, 刘洋, 李文博. 全球黑土区耕地动态变化及其驱动力Logistic回归分析[J]. 地质与资源, 2023, 32(5): 584-591. DOI: 10.13686/j.cnki.dzyzy.2023.05.008
    引用本文: 杨佳佳, 吕骏超, 金珊合, 刘洋, 李文博. 全球黑土区耕地动态变化及其驱动力Logistic回归分析[J]. 地质与资源, 2023, 32(5): 584-591. DOI: 10.13686/j.cnki.dzyzy.2023.05.008
    YANG Jia-jia, LYU Jun-chao, JIN Shan-he, LIU Yang, LI Wen-bo. FARMLAND DYNAMIC CHANGE AND ITS DRIVING FORCE ANALYSIS IN GLOBAL BLACK SOIL REGIONS BASED ON LOGISTIC REGRESSION MODEL[J]. Geology and Resources, 2023, 32(5): 584-591. DOI: 10.13686/j.cnki.dzyzy.2023.05.008
    Citation: YANG Jia-jia, LYU Jun-chao, JIN Shan-he, LIU Yang, LI Wen-bo. FARMLAND DYNAMIC CHANGE AND ITS DRIVING FORCE ANALYSIS IN GLOBAL BLACK SOIL REGIONS BASED ON LOGISTIC REGRESSION MODEL[J]. Geology and Resources, 2023, 32(5): 584-591. DOI: 10.13686/j.cnki.dzyzy.2023.05.008

    全球黑土区耕地动态变化及其驱动力Logistic回归分析

    FARMLAND DYNAMIC CHANGE AND ITS DRIVING FORCE ANALYSIS IN GLOBAL BLACK SOIL REGIONS BASED ON LOGISTIC REGRESSION MODEL

    • 摘要: 按照美国土壤分类系统, 全球广义黑土分为四大片.为探究全球黑土区耕地利用变化规律和驱动机制, 基于2005、2010、2015和2019年4期全球黑土区土地利用遥感监测数据, 借助GIS技术及数理统计软件, 分析全球黑土区耕地在14年间的动态变化特征.同时, 选取自然地理因素、社会因素为自变量, 以土地变化二分类(增加、减少)为因变量, 采用Logistic回归模型, 探讨研究区耕地变化的驱动因子.结果表明: 研究区内耕地变化显著, 3个阶段(2005-2010、2010-2015、2015-2019)先后出现小幅增加-大幅减少-小幅增加的趋势; 14年间耕地共减少58.77×104 km2, 其中亚洲黑土区耕地减少幅度最大, 占比31.70%, 减少的耕地有76.04%转变为未利用地.建立的耕地变化Logistic回归分析模型有效, 结果表明: 耕地变化在第1阶段(2005-2010年)的主要驱动因素是到最近道路距离、降水和高程(DEM); 第2阶段(2010-2015年)的驱动因素是到最近河流距离和高程; 第3阶段(2015-2019年)的驱动因素是到最近道路距离和温度.

       

      Abstract: According to the Soil Taxonomy, the black soil in the world is generally divided into four large tracts. To explore the change rule and driving mechanism of farmland use in the black soil regions, the dynamic change characteristics of farmland in global black soil regions during the past 14 years are analyzed based on the remote sensing monitoring data of land use during three stages of 2005-2010, 2010-2015 and 2015-2019 with GIS and mathematical statistics software. Besides, physical geographical and social factors are selected as independent variables and land change dichotomy of increase and decrease as dependent variables to explore the driving factors of farmland change in the study area. The results show that the farmland change is significant with a trend of slight increase-sharp decrease-slight increase during the above three stages. A total of 58.77×104 km2 of farmland has decreased in 14 years, of which the black soil region of Asia has the largest reduction, accounting for 31.70%, with 76.04% of the reduced farmland converted to unused land. The established Logistic regression analysis model of farmland change proves to be effective and shows that the main driving factors are distance to the nearest road, precipitation and digital elevation model (DEM) in the first stage (2005-2010), distance to the nearest river and DEM in the second stage (2010-2015) and distance to the nearest road and temperature in the third stage (2015-2019).

       

    /

    返回文章
    返回