Evaluating certain approaches for computing land surface temperature over alluvial plains of Northern India

Dual thermal band capability of Landsat 8 allows for the retrieval of Land Surface Temperature (LST), which is essential in agricultural and environmental studies. The accuracy of LST retrieval depends on precisely capturing land surface emissivity (ε). The scientific community has developed various approaches to determine LST using ε as a key input. This research evaluates the Twumasi, SEBAL, and Stathopoulou approaches for determining LST over the alluvial plains of Naini, Prayagraj (India). LST was estimated using these approaches with Landsat 8 OLI/TIR data from 08-12-2021 to 12-11-2023 and validated against ground-based air temperature observations recorded via Eddy Covariance measurements—a highly sophisticated and precise technique for turbulent flux measurement. The study also attempts to modify the Stathopoulou approach to estimate LST using a single thermal band (Band 10) instead of both Bands 10 and 11. Results revealed that the Twumasi and SEBAL approaches yielded more accurate LST estimates than the original Stathopoulou approach. However, improved error statistics—such as mean error, mean bias error, mean absolute error, mean relative error, and root mean square error along with high values of Willmott’s Index of Agreement, Nash–Sutcliffe Efficiency, and Pearson correlation coefficient were observed between observed air temperature and LST derived using the modified Stathopoulou approach. Therefore, the use of the modified Stathopoulou approach is recommended to enhance the quality and reliability of decision-making processes.