标题:
面向历史土地利用研究的苏皖地区耕地分布的影响因子分析及宜垦性模型构建Analysis of the Impact Factors of Cropland Distribution and Model Building in Jiangsu and Anhui for Historical Land Use Research
作者:
袁存, 叶瑜, 李蓓蓓, 魏学琼, 方修琦
关键字:
因子分析, 网格化重建, 耕地, 模型构建, 苏皖Factors Analysis, Grid Reconstruction, Cropland, Model Building, Jiangsu and Anhui
期刊名称:
《Geographical Science Research》, Vol.3 No.2, 2014-06-26
摘要:
精确的格点化区域历史土地利用数据是进行历史土地利用/土地覆被变化环境效应模拟的基础。由于缺少充足可信的历史图件等资料，历史网格化数据无法直接获取，通常是根据现代耕地分布格局及其形成的驱动因素，对某一区域的历史耕地总和数据进行降尺度的格点化分配。本文以苏皖地区为例，基于现代遥感数据分析了耕地空间分布与气候、坡度、海拔高度、土壤、河流湖泊、人口等因子的关系，发现影响苏皖地区耕地分布最为关键的因素是河湖及坡度。河流湖泊地区为非耕地的剔除区域，坡度与垦殖率的相关系数高达−0.628，苏皖地区海拔未达到气候影响下的耕地垦殖上限。在此基础上构建了耕地空间分布模型，对现代的耕地格局进行10 km × 10 km的模型重建，并进行了模型评估检验。检验结果显示，无论是从均值还是从相关系数或标准偏差来看，宜垦性模型的空间分配结果与现代遥感数据的空间格局差异均不大，两者的相关系数R为0.798，标准偏差S为22.2，两者差值的均值为0.09。该现代宜垦性模型能够用于耕地格局重建，可为该区历史时期耕地数据网格化处理提供参考。
Accurate historical gridded data of land use are the base for the environmental effect simulation of the land use/land cover change. Because of lacking of sufficient credible historical map data, his-torical grid data couldn’t access directly. It is usually assigned downscaling based on the modern distribution pattern of cropland and its driving factors. This paper analyzed the relationship be-tween spatial distribution of cropland and climate, slope, elevation, soil, river and lake, population based on modern remote sensing data in Jiangsu and Anhui. The most key factors in Jiangsu and Anhui are river and elevation. The area that river flows through doesn’t distribute cropland. The correlation coefficient of slope and reclamation rate is as high as −0.628; the highest elevation in Jiangsu and Anhui is under the upper limit elevation that can be cultivated because of climate. The paper also constructed the cropland spatial distribution model with spatial resolution of 10 km × 10 km, and assessed the model. It is suggested that the reconstruction data and the remote sensing data are not different obviously from the correlation coefficient or standard deviation. The corre-lation coefficient is 0.821 and the standard deviation is 20.6. This model can be used in cropland reconstruction in Jiangsu and Anhui, and provide references for historical cropland gridded data.