Donghui Bai, Chunping Miao, Yafei Xi, Hongyi Liu, Chao He
Building and green space morphology are considered key factors in urban heat regulation. Investigating their interactive effects on the urban heat environment provides both a decision support basis for urban planning and a new theoretical perspective for understanding the relationship between urban morphology and heat environments. This study analyzed the interaction between building morphology and green space on land surface temperature (LST) by integrating XGBoost and Geodetector and explored the nonlinear regression relationships between these factors and LST in Xi’an, China. The results revealed strengthened interactive effects between the average building height, average building volume, building surface area, and floor area ratio with green space on LST. In contrast, the interaction between the building coverage ratio and green space on LST was nonlinearly weakened. LST was the lowest when the green space’s area fractal dimension was 0.75–1.25. A green coverage ratio of 0.18–0.7 realized relatively strong cooling effects. Shapley Additive Explanation (SHAP) values increased with the building coverage ratio but were negative when the building coverage ratio was below 0.6, where its interaction with smaller green space area, lower total edge length, and higher area fractal dimension produced lower SHAP values. The findings offer scientific support for mitigating the urban heat island effect, thus contributing to the ecological construction and sustainable development of cities.