Comparison of two atmospheric correction methods for Landsat TM thermal band

The high resolution of Landsat Thematic Mapper (TM) thermal infrared data can be interesting in applications such as energy balance and crop water stress studies. However, surface temperature measurements are strongly affected by atmospheric effects, particularly by water vapour absorption, therefore a correction method is absolutely necessary. The impossibility of applying a split-window equation and the difficulty of having temporarily coincident radiosondes produce a slight use of Landsat TM data to obtain surface temperature. Vidal et al. show that in a zone with homogeneous climatic conditions, atmospheric correction can be evaluated by using energy balance equation and standard meteorological parameters. Using data of the EFEDA project, a comparison between two different atmospheric correction methods has been made in this paper. The first one was proposed by Vidal el al. and the second one was proposed by Coil et al. in 1994, where atmospheric parameters are calculated from a radiative transfer model, using atmospheric profiles obtained from local temporarily coincident radiosondes. Comparing blackbody temperatures calculated with both methods, a root mean square error of 1.6 degrees C has been obtained. A complete discussion of the sensitivity of both methods is also included in order to explain the differences of blackbody temperatures obtained using each method.