基于人口流的中国省域城镇化空间关联特征及形成机制

空间关联是区域相互作用和影响的纽带,以人口流动为载体的城镇化空间关联影响并不断调整着中国区域经济社会发展格局。根据省际人口流动数据,构建城镇化空间关联模型,通过分析1990-2015年中国省域城镇化空间关联强度与关联网络结构变化,揭示城镇化空间关联特征及形成机制。结果发现：① 中国省域城镇化空间关联强度演变过程可以解析出7种类型,其中主导类型与全国的吻合程度最高,并决定着全国城镇化空间关联强度演变过程;非主导类型的城镇化空间关联强度演变特征以2005年为界,之前增长缓慢甚至为负,之后与主导类型共同推动全国城镇化空间关联强度持续增长。② 对城镇化空间关联进行不同尺度的结构分析,全国城镇化空间关联网络结构日趋复杂,关联路径由集中转向分散,不同属性的地域子群间关联密度差异开始缩小,高等级节点由东部向中西部扩散。③ 采用面板向量自回归和面板分位数回归检验城镇化空间关联强度与关联网络中心度的互动机制,并解释了居民收入水平和区域产业结构对城镇化空间关联特征的影响方式。城镇化空间关联不仅深化了城镇化的研究,也为中国人口城镇化发展和产业空间转移调控提供理论依据。

Characteristics and formation mechanism of China's provincial urbanization spatial correlation based on population flow

Spatial correlation is the link of regional interaction and influence; the urbanization spatial correlation is spatial relationship of urbanization based on population flow, which influences and continuously adjusts the regional economic and social development pattern in China. By analyzing the change of the spatial correlation intensity and the spatial correlation network of provincial urbanization in China during 1990-2015, this paper reveals the spatial correlation characteristics, formation process and influencing factors of China's provincial urbanization since China's reform and opening up. The results show that: (1) 7 types of spatial correlation intensity are identified in the evolution of provincial urbanization of China, among which the dominant type has the highest degree of coincidence with the whole country, and determines the evolution process of the spatial correlation intensity of urbanization in China. The non-dominant types of urbanization spatial correlation intensity evolution feature 2005 as the critical year, and their growth was slow or even negative before the year; then together with the dominant type, they have promoted a continuous increase in the urbanization spatial correlation intensity across the country. (2) Urbanization spatial correlation network structure characteristics are analyzed from national, regional and provincial scales. The structure of the spatial correlation network of urbanization is becoming more and more complicated, and the correlation paths have shifted from centralized to scattered. The differences in the correlation density between regional subgroups with different attributes have begun to shrink, and the high-level nodes have spread from the eastern to the central and western regions. (3) Further, the panel-vector auto-regression and panel quantile regression are used to test the interaction mechanism between urbanization spatial correlation intensity and correlation network centrality, which explains the effect of residents' income level and regional industrial structure on urbanization spatial correlation characteristics. The specialization of urbanization spatial correlation not only deepens the research on urbanization theory and methodology, but also provides a theoretical basis for the development of population urbanization and the adjustment of industrial space in China.