Comparison of different land surface temperature (LST) estimation algorithms using remotely sensed images

Abstract

Earth Surface Temperature is a key indicator of energy balance on Earth. In recent years,
several methods for obtaining temperature using satellite images have been provided. This
study aimed to compare the performance of different methods to estimate Land Surface
Temperature (LST).Here, LST over Tehran metropolitan city has been estimated through
three methods including Artis, Mono_window and Stefan_Boltzman algorithms applied to
Landsat (TM, ETM +, OLI) and MODIS image, Emissivity obtained through image classification,
the vegetation index (NDVI) and MODIS emissivity product, Then, using an accurate
thermometer, surface and air temperature at a height of one and a half meters were taken the
linear relationship between surface temperature and corresponding air temperature has been
established. A statistical measure namely mean absolute error and T test for selection of best
method were used. The results show that Stefan_Boltzman method with mean absolute error
of 1.540 °C was the best one. Therefore, it is suggested that the Stefan_Boltzman method be
used For other areas with the same weather conditions and geographical parameters.