A method to improve satellite soil moisture retrievals based on Fourier analysis

[1] Knowledge of the spatial distribution and temporal changes of the global soil moisture for a long period of time is crucial to the understanding of climate changes and hydrological processes. By applying Fourier analysis to the time-series observations from the space-borne passive microwave sensors, this paper proposes a method to extract the high-frequency part of the satellite observed signals that reflect the soil moisture changes and help to generate the historical soil moisture datasets with an improved accuracy. The method is applied to the observations from Advanced Microwave Scanning Radiometer for the Earth Observing System (AMSR-E). Validations using field sampled soil moisture from two watersheds in the U.S. indicate that the method can effectively improve the sensitivity of current National Aeronautics and Space Administration (NASA) soil moisture products to the temporal changes of the surface soil moisture, with the correlation coefficients of the retrievals and measurements increased from 0.462 to 0.595 and 0.403 to 0.613 for the two watersheds, respectively.