[Exploring the Characteristics of Influencing Factors and Mitigation Strategies for Surface Urban Heat Island...

With the acceleration of global urbanization, the urban heat island （UHI） effect has become increasingly pronounced, posing severe health threats to urban populations during extreme high-temperature events. However, due to the limitations of the urban-rural dichotomy, comparative studies on the UHI effect across multiple cities are fraught with significant uncertainties. There is an urgent need to conduct in-depth analyses to identify underlying patterns, which would facilitate the development of effective and targeted heat mitigation strategies applicable to multiple cities. We enhanced conventional Local Climate Zone （LCZ） classification methods by incorporating three additional indicators-Building Volume Density （BVD）, Height-to-Width ratio （H/W）, and Average Nighttime Light Intensity （ANLI）-and employed multi-source data to generate high-resolution LCZ maps for Hangzhou, Nanjing, Hefei, and Shanghai. We then applied various machine learning algorithms to analyze and compare the factors influencing daytime SUHII across the four major cities of the Yangtze River Delta （YRD）. The results indicated that： ① The average overall accuracy of LCZ classification across the four cities exceeded 80%. ② Among these cities, Hangzhou exhibited the highest summer SUHII （TDI = 1.37）, while Nanjing recorded the highest winter SUHII （TDI = 1.12）. ③ Seasonal variations in SUHII across different LCZs were evident, with LCZ1-LCZ6 contributing to cooling effects during summer, whereas LCZ8 and LCZ10 displayed significant heating effects. ④ Both the Extreme Gradient Boosting （XGBoost） and Adaptive Boosting （AdaBoost） algorithms achieved high accuracy levels across the four cities and performed well in modeling the LCZ-SUHII relationship. Furthermore, the contribution of SUHII influencing factors varied significantly among cities, with the Pervious Surface Fraction （PSF）, Normalized Difference Vegetation Index （NDVI）, and Modified Normalized Difference Water Index （MNDWI） exhibiting distinct cooling effects in different cities. For example, MNDWI affected the surface urban heat island intensity in all four cities, but the importance of the water factor in Hangzhou was several times stronger than in Nanjing, Hefei, and Shanghai. In contrast, in Shanghai, one of the most highly urbanized cities in China, PSF was the most prominent influencing factor. This may be because the main urban area of Shanghai lacks large areas of natural surfaces. Instead, numerous fragmented and scattered natural or artificial water bodies forming pervious surfaces have become the most significant factor influencing the SUHII in Shanghai. These findings provide precise, effective, and cost-effective UHI mitigation strategies not only for China but also for other major cities worldwide. The addition of the three new indicators-BVD, H/W, and ANLI-has enhanced the accuracy and reliability of the LCZ classification, allowing for a more nuanced understanding of the urban thermal environment. By integrating multi-source data and advanced machine learning techniques, we offer a robust framework for assessing and mitigating the UHI effect in urban areas. The insights gained from this research can inform urban planning and policymaking, helping to create more resilient and sustainable cities that are better equipped to cope with the challenges of climate change.