Spatial patterns of urban population growth-shrinkage differentiation in China and their associations with multidimensional flow networks

Urban population growth and shrinkage are intertwined with cities' positions in wider systems of intercity relations. This study examines 330 prefecture-level and above cities in China by combining population change during 2020–2023 with three intercity networks observed in 2023: transportation mobility, online-search attention, and patent collaboration. Spatial analysis and social network analysis are used to characterize population and network patterns, and statistical associations between population-change types and multidimensional network positions. Results show that: (1) Population decline was widespread but generally modest, with 219 cities (66.4%) recording decline and slight shrinkage accounting for 57.3% of the sample. Moderate shrinkage was more spatially concentrated, particularly in Northeast China, whereas growth was geographically heterogeneous. (2) The three networks display distinct structural characteristics. Transportation and innovation networks exhibit clearer hierarchical differentiation, while the online-search network has broad binary connectivity but greater inequality in attention intensity and intermediary positions. Population-change types also differ selectively across network measures: moderate-shrinkage cities generally occupy lower transportation and innovation positions, whereas slight-growth cities show higher mobility intensity and stronger innovation-network centrality. (3) Three multidimensional node types are identified: comprehensive hub–innovation-leading, regional hub–functional intermediary, and peripheral–weakly connected cities. Growth accounts for 70.4% of the first type, whereas moderate shrinkage is disproportionately concentrated in the third type. However, the coexistence of growth and shrinkage within each node type indicates a probabilistic rather than deterministic correspondence. These findings show that recent urban population differentiation is associated with multidimensional relational position in network-specific ways, it provided a structural basis for place-sensitive urban governance.