Towards the Operational Spatialization of the Single Band Thermal Atmospheric Correction. Application to Landsat 7 ETM

This work aims to improve the accuracy in Land Surface Temperature (LST) from single-channel thermal sensors by providing spatialized maps of transmittance, upwelling and downwelling atmospheric radiances required in the radiative transfer equation. Two different techniques are introduced for the estimation of pixel-by-pixel atmospheric parameters, focusing on the correction of Landsat Thermal Infrared (TIR) data. First technique is based on the linearization of the atmospheric parameters with the total column water vapor (W), extracted from the MOD05 product, whereas a second technique uses the Single Band Atmospheric Correction (SBAC) tool. Ground-measured values of LST in an agricultural area in central Spain, covering a variety of surface conditions, were used for a local assessment. Very similar estimation errors <2 K were obtained using both, linearization with W and SBAC techniques applied to a set of 12 Landsat 7 images. W values ranged 0.6-4 cm in this work. Wetter conditions might deteriorate the performance of the linearization technique. A preliminary analysis of the operational spatialization of this pixel-by-pixel atmospheric correction is also included.