Development and validation of algorithms for LST measurement from NOAA-11/AVHRR satellite data

Abstract

The purpose of this project was to develop the algorithms and characterise measurement errors of land surface temperature (LST). An initial investigation of the accuracy of retrieval of LST has employed synthetic radiances, produced using Lowtran7, for a series of land and atmospheric situations representative of the Australian continental climate. The synthetic data have been modelled to incorporate variation of surfaces emissivity, atmospheric temperature, water vapour profile, and the surface or boundary temperature. The study indicated that the success of the scheme depended on the validity of the modelled radiances, particularly with respect to the formulation of the radiative role of the water vapour constituent of the atmosphere. Whirls, the use of regression in deriving the coefficients for LST equation minimised the bias errors in the radiative transfer calculation that may arise from air mass dependence. However, our algorithms derived using statistical regression and synthetic atmospheres indicated good accuracy for LST measurement with and rms error of order 1°C when validated using data for an instrumented field site.