Comparing different space-borne sensors and methods for the retrieval of land surface temperature

The importance of land surface temperature (LST) is increasingly recognized, and various methodologies have been proposed for the retrieval of LST using space-borne thermal infrared data. However, the selection of LST retrieval from Thermal Infrared Sensor (TIRS) of Landsat-8 based on different methods and the readily available MODIS LST products is still a challenging topic for local and global environmental studies. In this study, the potential of three different methods for retrieving LST using Landsat-8 TIRS data, including Radiative Transfer Equation (RTE), Single Channel (SC), and Split Window (SW) method in comparison with MODIS MOD11A1 LST product was evaluated. For accuracy assessment, 0 cm ground surface temperature (LST_GST) data was used. Our results almost showed same accuracy for RTE_B10 with RMSE = 0.35 °C, followed by MODIS with RMSE = 0.36 °C, and SC_B10 with RMSE = 0.38 °C. Secondly, SC_mean (Mean of B₁₀ and B₁₁), and RTE_mean (Mean of B₁₀ and B₁₁) generate nearly the same accuracy with RMSE = 0.53 °C, and RMSE = 0.54 °C, respectively. The other methods viz., SC_B11, RTE_B11, and SW method slightly showed lower accuracy with RMSE = 0.87 °C, RMSE = 0.88 °C, and RMSE = 0.91 °C, respectively. We found all the methods highly accurate and can be used successfully by climatologists, environmentalists, hydrologists, and urban planners concerning planning, and monitoring of the ever-increasing LST at local and global scale studies.