A microwave-optical/infrared disaggregation for improving spatial representation of soil moisture using AMSR-E and MODIS products

This study validated the Advanced Microwave Scanning Radiometer E (AMSR-E) soil moisture products developed by the Vrije Universiteit Amsterdam (VUA) in collaboration with the National Aeronautics and Space Administration (NASA) using ground measurements at six Rural Development Administration network sites during the 2007 growing season (May 1 through September 30) in Korea. In order to overcome current key validation issues from the spatial scaling mismatch between the ground measurements (point scale) and the AMSR-E soil moisture products (25 km scale), a synergistic approach from 25 to 1 km spatial resolution was performed using auxiliary data from a Moderate Resolution Imaging Spectroradiometer (MODIS). The 25 km VUA-NASA AMSR-E soil moisture data had error statistics (biases = -0.099 to -0.170 m(3) m(-3), root-mean-squared error = 0.131 to 0.179 m(3) m(-3), and correlation coefficients = 0.127 to 0.725) that were worse than the suggested goal of accuracy (less than or equal to root-mean-squared error of 0.06 m(3) m(-3)) determined by the several previous validation studies. This deficiency is theoretically due to the spatial scaling mismatch and different measurement depth between the ground measurements and the VUA-NASA AMSR-E soil moisture products. The disaggregated 1 km soil moisture was reasonably similar to the 25 km VUA-NASA AMSR-E soil moisture at spatial and temporal scales with better error statistics (biases = -0.024 to -0.158 m(3) m(-3), root-mean-squared error = 0.059 to 0.171 m(3) m(-3), and correlation coefficients = 0.348 to 0.658). Although additional studies are needed under a range of field conditions, the synergistic approach used in this study appears to be a feasible method to improve the spatial distributions of VUA-NASA AMSR-E soil moisture products as well as future microwave soil moisture products. (c) 2012 Elsevier Inc. All rights reserved.