基于空间分析方法的中国区域差异研究

Evaluation of Regional Disparity in China Based on Spatial Analysis

利用中国1952～2000年各省区的人口、土地利用和人均GDP等数据,采用空间分析方法对中国区域社会经济发展差异问题进行了实证研究.对中国大陆的几何中心、可利用土地、人口和GDP总量的空间中心计算表明,空间中心统计是一种简单有效刻画要素空间分布的方法.通过对中国各省区人均GDP的空间自相关系数Moran's I的计算,表明中国社会经济发展主要指标存在强烈的空间自相关,空间分析方法不失为一种深入理解区域经济格局及其变化的重要和有益的方法.对不同年度的Moran's I和Local Moran I的分析也揭示了中国区域经济格局的时空变化规律.

Based on the social-economic data from 1952 to 2000 in China, the question of regional disparity in China was reconsidered by using spatial analysis methods. Spatial analysis needed in social and economic science for the observed spatial values are not independent each other, i. e. , （or） they don＇t follow the same distribution, and （or） there is a trend along different directions. In this paper, the centrographic statistic was used to estimate basic parameters about the spatial distribution. The geometric center of Chinese Mainland with the centers of arable land area, population and GDP based on provincial level data were compared. The results show that the center of population and GDP had a significant offset with the center of geometry and land use. These are the key to understand the spatial disparity in China. The centers of population have an obviously trend of shift from the east to the west of China Mainland. This may be caused by the family planning and the other polices. But the shifts of the population center will help to improve the development of the west of China. As comparison with population, the shifts of the center of GDP had a different direction. It moved to south since the 1978 while the opening-up policy was adopt in china. In other words, the south of China had more rapid increase than north since 1978. This is not consistent with the opinion that the disparity of regional economic is great from east to west in China. The different shifts of centers of population and GDP also indicates that the economy of west did not increase with the increasing of population. The shifts of population and GDP centers indicate the change of the social and economic pattern in China. The difference of them also indicates that the imbalance of development in China. The Exploratory Spatial Data Analysis （ESDA）, which based on the computing spatial autocorelation and spatial heterogeneity, was also used to detect the geographical dynamics of Chinese regional disparity patterns. There are significant positive spatial autocorrelation （Moran＇s I ）of per capita GDP in China. That is, the relatively high （low） developed province tends to be located nearby other high （low） developed provinces more often than expected due to random chance, and then each province should not be viewed as an independent observation. The econometric estimations based on geographical data （ i. e. localized data） have to take into account the fact that economic phenomena do not be randomly spatially distributed. We also compared the temporal change of the spatial autocorrelation in China, and found that there is an obviously temporal increase of Moran＇s I since 1952 to 1995. This means that the disparity is increased in the same periods. But this trend does not keep on since 1990s because we found that the Moran＇s I soothed with a little fluctuating. Moran＇s I Scatterplots and LISA （Local Indicators of Spatial Association, LISA） cluster maps were used to test the local pattern of the Chinese economic development. The results of local statistic show that the two types of clusters （ High-High and Low-Low） are increasing which means that the heterogeneous is increasing too. And this is the other indicator of the regional disparity in China.