Port System Evolution in the Pearl River Basin: Evidence from Inter-Port-Area Cargo-Flow Networks

As ports become increasingly embedded in broader transport and logistics systems, inter-port linkages have become important for understanding regional port-system evolution. Although network analysis has been used to examine port systems, the characteristics of system evolution reflected in temporal changes in inter-port-area linkages within regional port systems remain insufficiently understood. Accordingly, this study constructs annual directed cargo-flow networks for 42 port areas in the Pearl River Basin from 2020 to 2024. Network analysis examines short-term evolution, while random forest, Shapley additive explanations (SHAP), and accumulated local effects (ALE) analysis investigate factors associated with inter-port-area cargo-flow linkage strength. The results show that inter-port-area linkage coverage expanded during 2020–2024, whereas average linkage strength did not exhibit a corresponding sustained increase. Relational organization remained broadly stable, while cargo-flow linkages extended inland from the Pearl River Delta along the Xijiang corridor and high-strength linkages remained concentrated in established core areas. Minimum port-area scale level, waterway distance, and minimum navigability condition were the most important associated factors, while administrative boundary crossings, differences in hinterland economic scale, and road-network density showed more complex association patterns. Overall, these findings provide empirical evidence for understanding regional port-system evolution from a relational perspective and offer practical insights for coordinated and differentiated port-system planning.