作　　者： ; (凡宸）; (秦建桥）; (高洋）; (吴海华）;

机构地区： 惠州市大亚湾高新经济技术开发区环保局

出　　处： 《中国人口·资源与环境》 2014年第S1期226-230,共5页

摘　　要： 人口调查统计以行政区划为基本单元,数据精度不能满足栅格尺度上的空间结构分析,也难以与生态环境综合研究中的自然地理数据相匹配,而人口数据降尺度空间模拟是解决这一问题的有效途径。本文基于统计学方法和GIS技术,对惠东县乡镇人口数据与空间因子进行相关性分析,并筛选出建设用地指数、高程、到居民点距离等因子用于回归分析,分别采用一元和两种多元回归方法建立人口密度数据空间化模型,最终在GIS平台中实现人口密度的降尺度模拟,获取200m×200m栅格尺度的人口密度数据。一元回归分析中,建设用地指数因子的模拟结果最优,R2为0.734,可作为快速、粗略模拟惠东县人口密度的模型;多元回归分析中,逐步和向后回归模型的R2分别达到0.775和0.886,模拟结果均较为理想。通过对模型的分析可知:①多元模型明显优于一元模型;②向后回归模型优于逐步回归模型。研究结果表明此人口密度模拟方法具有较强的操作性,可为县域尺度人口密度空间化的应用研究提供借鉴。 Census statistics used administrative region as abasic unit. Data accuracy not only cannot meet space structure analysis on grid scale,but also cannot match natural geographic data on eco-environmental comprehensive research. Space downscaling-simulation of demographic data can be an effective method to solve this problem. This article carried out research based on statistical method and GIS technique. Firstly,we use Pearson correlation to discuss the relevency between demographic data and space factors on town's administrative zone of Huidong County. Secondly,we screen out the following factors : construction land index,elevation,distance-to-settlement,etc. in regression analysis. Thirdly,we established some spatial models of population density with simple linear regression, backward regression model and stepwise regression model. Finally,we realized downscaling simulation of population density on GIS platform,and obtained population density data on raster scale( 200 m by 200 m). On the analysis of simple linear regression,the simulation result of construction land index was the best,R2 reached 0. 734; therefore,we use this index as a quick and rough model to simulate population density of Huidong County. On the analysis of multiple regression,R2 of backward regression model reached 0. 775,R2 of stepwise regression model reached 0. 886,and simulation result was reasonable. On the analysis of model result,we find out that the result of multiple regression had the advantage of simple linear regression,and backward regression model was superior to stepwise regression model. These research results demonstrated that this population density simulation method had strong operability,and could provide the same research with use for reference.

关 键 词： 人口密度 统计数据 空间分析 惠东县

领　　域： [社会学] [天文地球] [天文地球]