The satellite remote sensing retrievals are based on pixel by pixel, but it is quite difficult to obtain some atmospheric parameters such as air temperature and air pressure  using the remote sensing technique and field campaign, while numerical modeling can provide them easily. So, the atmospheric parameters from mesoscale model simulation are corrected using the filed observation and then are applied in the estimation of evapotranspiration using EOS/MODIS data. Validation results show that the estimation error of sensible heat flux decreased 5.8% and error of latent heat flux decreased 5.5%  using the numeric results comparing with  the single observation, all errors are less than 10%, which indicate that the land surface flux could be retrieved successfully  using the SEBS model. Analysis on the distribution of the land surface heat flux displays that the introduction of numericalsimulation improves the description on the characteristics of energy balance in the oasis and the deserts. Latent heat flux is little in the deserts and Gobi, while it is greater in the oasis and water.

Download