A General Split Window Algorithm for Land Surface Temperature Estimation

摘要

A general multichannel algorithm for land surface temperature estimation is described. It has been set up by simulations of NOAA / AVHRR-2 measurements in order to parameterize the temperature correction due to atmospheric water vapor absorption and deviation of surface emissivity from unity at 11 and 12 μm wavelengths. The parameterization is based on model calculations performed with FASCODE, a line - by -line radiative transfer program. A comparison with one particular algorithm selected by ESA to produce global thermal maps of the Earth shows no significant difference in the retrieved surface temperature for water vapor around 2.5 g/cm~2. For greater deviations from this value of water vapor content the difference in the results can be remarkably large. Model calculations have been performed for deviations of each parameter (water vapor and surface emissivity) individually and for both parameters simultaneously. The approach used allows the application of the algorithm to other sensors. In fact, the next step of our activity on this subject envisages its application to the appropriate bands of the MODIS sensor.