Vegetative and Atmospheric Corrections for the Soil Moisture Retrieval from Passive Microwave Remote Sensing Data: Results from the Southern Great Plains Hydrology Experiment 1997

Abstract

A radiative transfer model and data from the Southern Great Plains 1997 Hydrology Experiment were used to analyze the dependency of surface emissivity retrieval at 19 GHz on atmospheric and vegetative effects. Volumetric soil moisture obtained from ground measurements in the Central Facility area that show a dynamic range of 25% was highly correlated with the corresponding L-band electronically steered thinned array radiometer (ESTAR) 1.4-GHz and Special Sensor Microwave Imager 19-GHz brightness temperatures. For the Little Washita area, only the ESTAR measurements were well correlated with volumetric soil moisture. Atmospheric corrections, which were calculated from collocated radiosonde measurements, did not improve the soil moisture retrieval significantly. However, a sensitivity study at 19 GHz using a larger dataset of 241 radiosonde ascents indicates that the variability in integrated atmospheric water vapor introduces variations of 0.023 in surface emissivity. This value is ∼36% of the variability caused by changes in soil moisture. Therefore, atmospheric corrections should generally improve the soil moisture retrieval at 19 GHz. Different water vapor absorption schemes and absorption by nonraining clouds do not affect this result. Even for sparse vegetation (vegetation water content of 0.33 kg m−2), the effect on soil emissivity retrieval is significant. Because of the lack of appropriate data for vegetation cover and single scattering albedo, the effects of the vegetation had to be estimated. Within a reasonable parameter range they were comparable to the effects caused by soil moisture changes. To quantify the effect of surface emissivity changes on integrated water vapor retrieval, brightness temperatures were modeled using actual soil and atmospheric parameters. The radiative transfer equation was then inverted with respect to the atmospheric contribution using an average value for the surface emissivity. An uncertainty of 5% in volumetric soil moisture caused an error of 30 kg m−2 in integrated water vapor.