Predicting Land Surface Temperature and Land Cover Changes Based on Multisource Remote Sensing Spatio-Temporal Fusion in Hefei, Eastern China

With the rapid urban infrastructure expansion in China, numerous cities have been suffering from the negative effects of urban warming and ensuing social and environmental challenges. Accurate acquisition of long-term dynamics in urban land surface temperature (LST) helps to the development of sustainable urban management actions. In this article, based on the Landsat and MODIS long-term observations from 2000 to 2020 covering Hefei City, the capital of Anhui Province of China, four spatio-temporal fusion algorithms including the spatial and temporal adaptive reflectance fusion model (STARFM), enhanced spatial and temporal adaptive reflectance fusion model, flexible spatio-temporal data fusion model and spatial and temporal nonlocal filter-based fusion model were implemented and compared, followed by the extraction of the spatial distribution characteristics and evolution trend of urban summer average LST. Next, the relationships between LST grading zones and different land cover types were explored, and the 2030 urban LST was predicted by using the patch-generating land use simulation coupled with Markov chain model. The results showed that STARFM was the best spatio-temporal fusion algorithm here. From 2000 to 2020, the high temperature zone expanded from 638.81 to 886.31 km
²
, and there was still a continuous trend of enhancing high temperature zone and built-up area in Hefei during 2020–2030, indicating that the urban heat island effect will intensify continuously. This study provides a feasible and comprehensive approach to monitoring urban surface temperature dynamics. These findings contribute a scientific insight for local governments to formulate urban land management plans.