The formation and proximity mechanism of population flow networks under multiple traffic in China

Population flow within a specific spatial range can promote social and economic reaggregation and diffusion. However, few empirical studies have paid close attention to population flow networks (PFNs) under different traffic. Tencent migration data were used to analyse the spatial patterns and aims to identify how proximity affects PFNs among 338 Chinese cities in 2018. The results indicated that the weighted average distance of population flow by highways and railways are 252.39 km and 566.16 km, showed restriction by distance decay in space. However, the weighted average distance of population flow by aviation was 1282.71 km, was less affected by geospatial constraints. PFNs involving highways, railways, and aviation display core-periphery, corridor, and hub-spoke structures, respectively. China's PFNs forms a diamond structure under multiple traffic, with Beijing, Shanghai, Shenzhen, Guangzhou, Chongqing, and Chengdu as the vertices. Negative binomial regression models were used to determine how proximities affect PFNs under multiple traffic. The estimated coefficients for geographical proximity, institutional proximity, and cultural proximity were significantly positive at the 1 % level. The influence of climate proximity on the PFNs may be complex, with a positive impact on PFNs on highways and railways but a negative impact on PFNs involving aviation.