Feed

Numerous studies demonstrate that urban green spaces enhance residents' health. However, limited clarity in green space classification and the complex interplay between green space attributes and other variables have constrained our comprehension of the nuanced relationships between specific green space attributes and health benefits. This study employs a case study in Beijing's central area, utilizing machine learning techniques (random forest modeling) combined with SHapley Additive exPlanations (SHAP) analysis to scrutinize the heterogeneous health impacts of various green space types. We examined 36 neighborhoods using multi-source data, including green space metrics, socioeconomic indicators, and health survey outcomes from 1116 residents. The results indicate that different types of green spaces contribute variably to health benefits. Park green spaces (PG) exhibited the most significant positive effect on residents' Physical type Summary (PCS) and Mental type Summary (MCS), surpassing the remaining green space types. Notably, residential green space with sampled communities (RG-I) exhibited a positive effect on residents' PCS, whereas residential green space with non-sampled communities (RG-O) was negatively associated with both the PCS and MCS. Street green space (SG) also showed a positive effect on PCS. Specifically, among all the green space metrics, Area_PG was the most critical determinant of PCS, followed by Proximity_PG and NDVI_PG. For MCS, Area_PG and Proximity_PG exhibited the most importance. Moreover, we delved into the interactions among predictors for health. In high-density neighborhoods, increasing the area of park green spaces significantly enhanced physical health benefits. While in neighborhoods with large area of park green spaces, positive social relationships more effectively boosted mental health benefits. These findings highlighted the distinct effects of various green space attributes in affecting health, especially the characteristics of land functions and utilization. Our study identifies the specific green space metrics in the neighbourhood that significantly impact residents' health, offering insights for more precise and effective urban green space regulation strategies aimed at improving human health.