Fine-Scale Analysis of the Long-Term Urban Thermal Environment in Shanghai Using Google Earth Engine

Abstract

Exploring the spatiotemporal patterns of urban thermal environments is crucial for mitigating the detrimental effects of urban heat islands (UHI). However, the long-term and fine-grained monitoring of UHI is limited by the temporal and spatial resolutions of various sensors. To address this limitation, this study employed the Google Earth Engine (GEE) platform and a multi-source remote sensing data fusion approach to generate a densely time-resolved Landsat-like Land Surface Temperature (LST) dataset for daytime observations spanning from 2001 to 2020 in Shanghai. A comprehensive analysis of the spatiotemporal patterns of UHI was conducted. The results indicate that over the past 20 years, the highest increase in average LST was observed during spring with a growth coefficient of 0.23, while the lowest increase occurred during autumn (growth coefficient of 0.12). The summer season exhibited the most pronounced UHI effect in the region (average proportion of Strong UHI and General UHI was 28.73%), while the winter season showed the weakest UHI effect (proportion of 22.77%). The Strong UHI areas gradually expanded outward over time, with a noticeable intensification of heat island intensity in the northwest and coastal regions, while other areas did not exhibit significant changes. Impervious surfaces contributed the most to LST, with a contribution of 0.96 °C, while water had the lowest contribution (−0.42 °C). The average correlation coefficients between LST and NDVI, NDWI, and NDBI over 20 years were −0.4236, −0.5128, and 0.5631, respectively.