Spatial regression analysis of land use impact on land surface temperature in four East Asian metropolises

The escalating impact of global climate change on the urban heat island effect presents a significant challenge for sustainable urban planning and the preservation of thermal comfort in metropolitan areas. The impact of land use composition on land surface temperature varies across different locations and climate zones. This study employed spatial regression, including Spatial Lag Model (SLM) and Spatial Error Model (SEM), to investigate the correlation between land use types and temperature in four East Asian metropolises: Beijing, Seoul, Tokyo, and Taipei. Moderate Resolution Imaging Spectroradiometer (MODIS) datasets are applied to generate the summer land surface temperature. At the same time, Sentinel-2 imageries and Visible Infrared Imaging Radiometer Suite (VIIRS) are integrated to derive the land use and cover, effectively distinguishing high- and low-development areas. The results indicate that Tokyo experienced the most severe heat among the four metropolises during the summer of 2020. Both spatial lag and spatial error models can effectively improve the OLS model, making its coefficients more representative. Additionally, we observe that low-density building areas have a higher impact on temperature than high-density building areas. The outcome of this study can be utilized to develop effective large-scale strategies for mitigating the urban heat island effect.