Mapping environmental influences on walking behavior across urban morphologies: A local climate zone-based study

Walking plays a critical role in promoting healthier and more sustainable cities, yet its environmental determinants are often examined without considering three-dimensional urban morphology. This study introduces a morphology-sensitive framework based on the Local Climate Zone (LCZ) classification to evaluate how built environment, climate, and air pollution factors jointly influence walking across distinct urban forms. Urban morphology is operationalized as a combination of building height and density, defined by the LCZ framework. Using large-scale walking trajectory data from Shenzhen, China, we applied Extreme Gradient Boosting (XGBoost) with SHAPley Additive exPlanations (SHAP) to identify key environmental factors at the citywide scale and across seven LCZ-defined morphological types.

Results reveal that environmental influences on walking vary significantly by urban morphology. In low-rise zones, walking is supported by transit access and walkability features, while natural environments such as NDVI and proximity to water play a minor role; air pollution increasingly constrains walking in high-density settings. Mid-rise medium-density zones present a transitional condition, where transit access, walkability features, and natural environments contribute positively, yet moderate suppression from air pollution and heat occurs. In contrast, high-rise areas experience strong compounded suppression from both air pollution and heat, while the benefits of green infrastructure become increasingly marginal.

These findings challenge the assumption of spatial uniformity in environment–walking relationships and highlight the moderating role of urban morphology. By integrating LCZ classification with interpretable machine learning, this study offers a novel approach to identify form-specific environmental constraints on walking, providing actionable insights for climate-adaptive urban design.