港口连通性及其影响因素的时空差异——以中国环渤海港口为例

为明确港口连通性及其影响因素的时空差异,基于港口供应链视角,从港口面向内陆、内贸和外贸连通能力构建港口连通性模型,并借助空间计量模型对2002-2017年间中国环渤海港口连通性及其影响因素的时空差异进行测度。结果表明：①除天津、青岛和大连港连通性最强外,内陆、内贸和外贸连通性最强分别为日照、唐山和烟台港,连通性最弱分别为威海、丹东和盘锦港,黄骅港增速最快;②连通性分布具有多核心-边缘特征和多门户港口并存格局;③连通性影响因素具有空间相关性和异质性,但均对连通性有正向促进作用。本文旨在使决策者明确港口运输的连通能力及其影响因素,为港口规划布局和提升在港口供应链中地位提供决策支持。

PORT CONNECTIVITY AND SPATIAL-TEMPORAL DIFFERENCE OF ITS INFLUENCING FACTORS: THE CASE OF CHINA BOHAI RIM PORTS

In order to measure port connectivity and spatial-temporal difference of its influencing factors, a port comprehensive connectivity model is constructed from the aspect of port supply chain, taking into account three levels of port hinterland connectivity, domestic trade and foreign trade connectivity capacity, based on improved gravity model and location condition. The port connectivity in China Bohai Rim is research during 2002-2017, and its influencing factors are explored by adopting the spatial econometric model, including spatial error model, spatial lag model and Geographically Weighted Regression (GWR). The results show that: 1) There are great differences in hinterland connectivity, domestic trade, foreign trade and comprehensive connectivity of ports in Bohai Rim. Tianjin and Qingdao ports have the highest hinterland connectivity, Dandong and Weihai ports have the lowest, and Huanghua and Tangshan ports have the fastest growth of connectivity. 2) The pattern of port connectivity in Bohai Rim has the characteristics of "multi- core-edge" distribution, that is, the Beijing-Tianjin-Hebei region with Tianjin port as its core, the Shandong Peninsula with Qingdao port as its core and Liaoning Peninsula with Dalian port as its core. 3) There are some spatial correlation (or dependence), heterogeneity and spillover effects in the port connectivity in Bohai Rim, and the difference of connectivity among ports is gradually narrowing. Spatial econometric regression results show that all the factors have positive effect on connectivity, the logistics trade scale of has the most significant impact, and its connectivity distribution conforms to the Matthew effect.