Revealing the driving factors of urban wetland park cooling effects using Random Forest regression and SHAP algorithm

•Utilized RF-SHAP to explore the driving factors of park cooling effects.•Percentage of water bodies inside parks is critical in cooling intensity.•Park area is the most important driving factor on cooling efficiency.•The interactions between driving factors are important for cooling effects. Urban...

Full description

Saved in:
Bibliographic Details
Published inSustainable cities and society Vol. 120; p. 106151
Main Authors Deng, Yue, Jiang, Weiguo, Ling, Ziyan, Liu, Lulu, Sun, Shujuan
Format Journal Article
LanguageEnglish
Published Elsevier Ltd 15.02.2025
Subjects
Online AccessGet full text
ISSN2210-6707
DOI10.1016/j.scs.2025.106151

Cover

More Information
Summary:•Utilized RF-SHAP to explore the driving factors of park cooling effects.•Percentage of water bodies inside parks is critical in cooling intensity.•Park area is the most important driving factor on cooling efficiency.•The interactions between driving factors are important for cooling effects. Urban wetland parks (UWP) can effectively mitigate urban extreme heat events in an environmentally friendly manner. Most studies ignored the interactions among the driving factors on the cooling effects of UWP. In this study, we integrated Landsat 8 imagery, Random Forest (RF) regression and SHapley Additive exPlanations (SHAP) algorithm to comprehensively investigate the cooling effects of 477 UWP across China from cooling intensity and efficiency perspectives, and to reveal the interactions of driving factors. Interaction between vegetation outside UWP and temperature had the strongest effect on UWPCI, while park area and park perimeter had the strongest effect on UWPCE. A low percentage of water body inside UWP (< 20%) could increase cooling intensity when percentage of impervious land outside UWP is greater than 50%. On average, the interactions between driving factors account for 92.15% and 129.01% of the direct influence on UWPCI and UWPCE, respectively. From the perspective of cooling efficiency, the optimal park area is 1.25 km2. This findings contribute to a deeper understanding of the driving factors behind the cooling effects of UWP and better serve UWP planning to mitigate urban extreme heat events.
ISSN:2210-6707
DOI:10.1016/j.scs.2025.106151