Extents of Predictors for Land Surface Temperature Using Multiple Regression Model

Land surface temperature (LST) is a key factor in numerous areas such as climate change, land use/land cover in the urban areas, and heat balance and is also a significant participant in the creation of climate models. Landsat data has given numerous possibilities to understand the land processes by means of remote sensing. The present study has been performed to identify the LST of the study region using Landsat 8 OLI/TIRS satellite images for two time periods in order to compare the data. The study also attempted to identify and predict the role and importance of NDVI, NDBI, and the slope of the region on LST. The study concludes that the maximum and minimum temperatures of 40.44 C and 20.78 C were recorded during the November month whereas the maximum and minimum LST for month March has increased to 42.44 C and 24.57 C respectively. The result indicates that LST is inversely proportional to NDVI (−6.369) and slope (−0.077) whereas LST is directly proportional to NDBI (+14.74). Multiple linear regression model has been applied to calculate the extents of NDVI, NDBI, and slope on the LST. It concludes that the increase in vegetation and slope would result in slight decrease in temperature whereas the increase in built-up will result in a huge increase in temperature.