Exploring Seasonal and Diurnal Variations of the Thermal Environment in Metropolitan Scale Analysis Based on Remote Sensing Data

Urban morphology, including land surface, building heights, vegetation, water bodies, and terrain, exerts a significant influence on the urban thermal environment. The complex and nonlinear pathways through which these factors exert influence present significant challenges in urban climate studies. However, existing studies of statistical approaches to the urban thermal environment have primarily focused on linear relationships, often overlooking the complex and nonlinear effects of these factors. Additionally, previous research on those approaches has not adequately addressed the seasonal and diurnal variations in land surface temperature, nor has it examined the extent to which urban morphology influences these variations. While simulation-based approaches can address these nonlinearities and temporal variations, they require large parameter sets and extensive high-resolution input data, making them computationally demanding. This gap limits the ability to develop targeted and effective urban heat mitigation strategies. Recent advancements in remote sensing technologies have revolutionized our ability to analyze these complexities using medium-resolution data products. In this study, we apply a polynomial regression model with an elastic net to represent the impact of terrain and urban morphological factors on the urban thermal environment, considering its seasonal and diurnal variations, taking the case of the Osaka Metropolitan Area. This approach is unique in terms of capturing the nonlinearity of the impacts based on earth observation data by remote sensing and efficiently captures complex relationships while maintaining interpretability and reducing computational overhead. The study leverages MODIS Terra thermal infrared data from 2018, supplemented by Sentinel-2 and Copernicus Land Cover data. The results reveal significant seasonal and diurnal variations in the thermal environment, indicating that building height influences LST non-monotonically, with daytime cooling effects in dense urban areas (0.12 to 0.19 °C decrease) but nighttime heat retention in suburban zones (0.06 to 0.13 °C increase). Similarly, vegetation coverage reduces nighttime LST more effectively, particularly beyond a critical density threshold (NDVI > 0.4). These findings suggest that by optimizing urban design by considering building height effects, strategic design of vegetation coverage can help mitigate heat/cold stress and improve thermal comfort throughout seasons. These findings contribute to sustainable urban development and heat mitigation efforts by providing data-driven insights into urban morphology’s impact on the thermal environment.