A Semi-Empirical Model for Computing Land Surface Emissivity in the Microwave Region

In an effort to better simulate land surface microwave emissivity, a semi-empirical technique is developed and tested over the U.S. Southern Great Plains (SGP) area. A physical model is used to calculate emissivity at the 10-GHz frequency, combining contributions from the underlying soil and vegetation layers, including the dielectric and roughness effects of each medium. Adjustments are added for post-precipitation surface water effects on emissivity of the soil and water-coated vegetation emissivity. A five-year data set of retrieved emissivities from the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) during clear-sky conditions is employed for calculation of a robust set of channel covariances. These covariances, combined with the modeled 10-GHz emissivities, provide emissivity values for each AMSR-E channel, which are then used to compute top of the atmosphere brightness temperatures Tbs. Results comparing these calculated Tbs to observed AMSR-E values show correlations of 0.85-0.93 and biases generally less than 1 K, with the largest bias appearing in the highest AMSR-E frequency. Such a modeling system could be easily implemented for the emissivity calculation required for atmospheric retrievals over similar land surfaces.