摘要
元胞自动机模型是土地利用变化模拟最有效和常用的方法之一,其模拟精度受元胞大小、时间间隔和邻域结构及大小等参数的影响。以玛纳斯河流域绿洲区耕地为研究对象,基于1975年、1995年、2005年、2010年及2015年土地利用数据构建耕地扩张CA模型,定量分析CA模型的参数敏感性。结果表明:(1)CA模型模拟精度具有时空尺度依赖性,时间间隔越短,精度越高,其中时间间隔为5 a的数量精度、位置精度及Kappa精度分别比时间间隔为20 a的高13.89%、2.63%和8.49%;空间尺度上,元胞大小越小,模拟精度越高,元胞大小为30 m的Kappa系数值(0.8671)高于120 m的模拟结果(0.8259)。(2)邻域结构及大小模拟结果表明,摩尔邻域模拟精度优于冯诺依曼邻域;同等邻域下,邻域大小为7×7模拟精度比3×3邻域模拟精度高,7×7摩尔邻域是模拟耕地的最优选择。(3)耕地模拟结果表明,2020年耕地面积较2015年相比减少149.29 km2,耕地占比已由2015年的67.40%降至66.18%,表现出因增长饱和而逐渐减缓的趋势。此研究可为全面了解元胞自动机模型敏感性提供一定的理论基础。
The cellular automaton model is one of the most effective and common methods for land-use change simulation.The accuracy of the simulation is affected by parameters such as cell size,time interval,and neighborhood structure and size.Following the land use data of the Oasis of Manas River Basin in 1975,1995,2005,2010,and 2015,this study constructed a cultivated land expansion CA model to analyze the parameter sensitivity of the CA model in the method of quantitative analysis.The results showed that(1)the simulation accuracy of the CA model depends on a spatial and temporal scale.The shorter the time interval,the higher the accuracy.The quantitative accuracy,position accuracy,and Kappa accuracy of a 5-year interval are 13.89%,2.63%,and 8.49%higher than those of a 20-year interval.At the spatial scale,the smaller the cell size,the higher the simulation accuracy will be.The Kappa coefficient value(0.8671)of a cell size of 30 m is higher than the simulation result of 120 m(0.8259).(2)The simulation results of neighborhood structure and size showed that the Moore neighborhood accuracy is better than Von Neumann’s.Under the same neighborhood,the neighborhood accuracy with 7×7 size is higher than that of the 3×3 size.The 7×7 Moore neighborhood is the best choice for simulating cultivated land.(3)The simulation results of cultivated land manifested that the cultivated land decreased 149.29 km^2 in 2020 compared with the area in 2015.The proportion of cultivated land decreased from67.40%in 2015 to 66.18%in 2020,which presented a gradual slow down due to growth saturation.This study can help provide a theoretical basis for a comprehensive understanding of the sensitivity of the cellular automaton model.
作者
朱磊
夏鑫鑫
杨爱民
靳含
ZHU Lei;XIA Xin-xin;YANG Ai-min;JIN Han(College of Grassland and Environment Sciences,Xinjiang Agricultural University,Urumqi 830052,Xinjiang,China;Xinjiang Key Laboratory of Soil and Plant Ecological Proceses,Urumqi 830052,Xinjiang,China)
出处
《干旱区研究》
CSCD
北大核心
2020年第5期1327-1336,共10页
Arid Zone Research
基金
国家自然科学基金项目(41561019)资助
作者简介
朱磊(1982-),男,博士,副教授,硕士生导师,研究方向为土地利用变化、地理信息系统及遥感应用.E-mail:smartzhulei04@163.com
