基于多源数据的河南省GDP空间化

GDP spatialization in Henan Province based on multi-source data

随着社会经济的快速发展,以县区为单位统计的GDP数据不能客观反映县级内部地区的经济差异,对GDP统计数据进行空间化是解决该问题的手段之一.本文以河南省为例,在GIS平台下采用分产业建模方式,结合土地利用数据、人口数据、DMSP/OLS数据、GDP统计数据,利用相关分析和回归分析的方法,实现GDP的空间化.结果表明,第一产业与土地利用类型有密切的线性关系;第二、三产业之和与人口数据、DMSP/OLS数据都有很好的相关性,将人口数据与DMSP/OLS数据相结合构建的综合因子与GDP_(2,3)之间的相关性更好,相关系数为0.949,R2为0.901.利用综合因子与第二、三产业GDP数据回归建模,可以提高第二、三产业空间化的精度.验证结果显示乡镇尺度模拟值与统计值之间的平均相对误差为10.34%,本研究的模型具有较高的精度.空间化后的GDP的密度图能够反映地区内部的经济情况,对研究该地区的经济空间差异有一定的价值.

With rapid development of social economy,provincial counties as units of GDP data cannot reflect economic differences in different regions of the same county.Therefore,spatializations have become an important issue in GDP research.In the present work we have studied GDP spatialization in Henan Province from land use,population,DMSP/OLS and GDP statistics.It was found that primary industry GDP（GDP1）had good linear correlation with land use types.Secondary and tertiary industry GDP（GDP（2,3））had good correlation with population data and DMSP/OLS data.A composite index（combination of population data and DMSP/OLS data）was found to have significant correlation with GDP（2,3）（R=0.949,R2=0.901）.Regression model using composite index as dependent variable and GDP（2,3） as independent variables were found to improve accuracy of GDP（2,3） spatialization.Accuracy verification found that average relative error between simulated value and statistical value was 10.34%.GDP density map could reflect local economy in Henan Province,and this will likely contribute to the study of economic development in the area.