Comparison between land surface temperature estimation in single and multi-channel method using LandSat images 8

Land surface temperature (LST) is a key parameter in environmental studies particularly for drought monitoring. Due to the ground limitations to measure the LST on a large scale, thermal remote sensing is a unique method for estimating LST. The aim of this article is comparing between LST estimation in single and multi-channel method using Landsat 8 thermal and reflective bands. Necessary ground data from meteorological stations Farabi (Khuzestan) and Karaj (Alborz) were taken to coincide with the dates and times of Landsat 8 overpasses. In this article Land surface emissivity and atmospheric water vapor content are major inputs for single and multi-channel LST estimation. After correction, processing and calculation of interest, LST were estimated. For result evaluation, statistical indices such as Root-Mean-Square Error (RMSE), Mean Absolute Error (MAE) and coefficient of determination (R2) were used. Results show the high value of R2 in all LST estimation method in comparison with ground measurement. In single channel using band 10 highest accuracy with MAE about 1.04 and 0.98 degrees in Karaj and Farabi station was seen respectively. The lowest and highest value of RMSE is in the single channel method (band 10) and multi-channel method (band 10 and 11) respectively. Study area conditions in terms of temperature; land cover and water vapor content affect the results and appropriate thermal band selection. Take-in consideration, especially using multi-band LST estimation method is suggested.