Using the multiple regression model to predict the land surface temperature in Al Najaf province, Iraq


  • Rana Kamel Albonwas, Ebtihal T AL-Khakani


LST, NDVI, BSI, Multiple regression, Landsat


Land surface temperature (LST) has a significant impact on local climate and ecosystem. Landsat data provided many potentials for understanding the land processes through remote sensing. In this study, LST was identified using the thermal bands of satellite imageries of Landsat 5 TM of 2000 and 2010, and Landsat 8 OLI of 2021 for comparing the data. The image was also employed to create the normalized difference index (NDVI) and bare soil index ( BSI). The correlations between LST, NDVI, and BSI were expressed using linear regression analysis.

The research found that the vegetation area (NDVI) has a strong negative relationship with LST, whereas Bare soil Land (LST) has a positive strong relationship with LST. The study adopted NDVI and BSI indicators to apply the multiple regression equation for LST prediction. The results of multiple regression for the three years showed a significant match between actual and predicted temperature.