Nonlinear Threshold Effects of Built Environment on Metro Ridership: Implications for Sustainable Urban Mobility and Parking Planning

With rapid urbanization and increasing motorization, understanding the impact of parking facilities on urban metro ridership is crucial for alleviating traffic congestion and promoting public transport priority strategies. However, few studies have systematically examined the influence of built environment characteristics, especially parking facilities, on metro ridership. To address this research gap, this study utilizes metro ridership data and parking facility data, and employs a gradient boosting regression tree (GBRT) model to analyze the relationship between built environment factors (including parking) and metro ridership. Additionally, accumulated local effects (ALE) plots are used to reveal nonlinear effects and interaction patterns. The empirical results demonstrate that parking space density has a significant nonlinear influence on metro ridership. Furthermore, a clear threshold effect is observed in the joint impact of parking space density and distance to the central business district (CBD), providing valuable theoretical and practical insights for optimizing park and ride (P+R) facility planning around metro stations. These findings contribute to a deeper understanding of the interplay between built environment factors and metro ridership, offering evidence-based guidance for sustainable urban transport planning.