A Remote sensing approach to evaluate the trend of surface temperature changes using spatial analysis and regression

The purpose of this study is to evaluate the trend of long-term changes in ground surface temperature and heat islands in Ahvaz city using the split window algorithm (SW) in the period from 2014 to 2022. The Thiel-Sen estimator and the Mann-Kendall model were used to accurately estimate the range of changes in the earth's surface temperature. Spatial autocorrelation of heat islands was evaluated using local Moran's index, and the relationship between LST changes and urban use parameters was evaluated using ordinary least square regression (OLS) and weighted regression (GWR). Til-Sen trend analysis has shown that 61.93% of the area has an increasing trend and 6.39% has a decreasing trend. According to the results of the Mann-Kendall significance test, only 6% of the area has a constant decreasing or increasing trend. It is significant and the other sections have no significant trends. The Pearson correlation results between the calculated air temperature and the ground station temperature are equal to 0.716. The evaluation of the earth's surface temperature using the global Moran's spatial correlation method showed that the temperature has a spatial structure with a cluster pattern and its value varies between 0.63 and 0.68. In the results of the local Moran index, Investigating the effect of urban land use factors on the trend of temperature changes using OLS and GWR methods, showed that all independent parameters considered are significant and GWR model has better results than OLS in the studied area.