A simple and fast atmospheric correction for spacebome remote sensing of surface temperature

Accurate surface temperature retrieval using thermal infrared observations from satellites is important for surface energy balance modeling; however it is difficult to achieve without proper correction for atmospheric effects. Typically the atmospheric correction is obtained from radiosonde profiles and a radiative transfer model (RTM). But rigorous RTM processing is impractical for routine continental scale modeling because of long computational times. An alternative, simpler, and faster approach for correcting observations in the 10-12.5 mum band is developed from a previously published water vapor continuum absorption function. Using the RTM program MODTRAN as a reference, the function is calibrated against 159 radiosondes, and then validated against the TIGR radiosonde (1761 profiles) data set. Implementation of the calibrated absorption function usually produced larger temperature corrections than without calibration, an effect due to water vapor band type absorption and to non-water vapor constituents. The resulting surface temperature estimates, within 0.8 degreesC of MODTRAN estimates, were achieved at 15 x less processing time than MODTRAN. (C) 2003 Elsevier Inc. All rights reserved.