Evaluation of Dielectric Mixing Models for Passive Microwave Soil Moisture Retrieval Using Data From ComRAD Ground-Based SMAP Simulator

Abstract

Soil moisture measurements are required to improve our understanding of hydrological processes and linkages between the Earth’s water, energy, and carbon cycles. The efficient retrieval of soil moisture depends on various factors among which soil dielectric mixing models are considered to be an important factor. The main objective of this work focuses on testing different dielectric models—Mironov et al. , Dobson et al. , Wang and Schmugge, and Hallikainen et al. —for soil moisture retrieval using the combined radar/radiometer (ComRAD) ground-based L-band simulator system, which serves as a simulator for the instruments on NASA’s soil moisture active passive (SMAP) mission scheduled for launch in early next year. The single-channel algorithm at H polarization ( SCA-H ) version of the tau-omega model was used for soil moisture retrieval. A summer field experiment was conducted in 2012 at the United States Department of Agriculture (USDA) test site from which ComRAD measurements and validation samples of soil moisture were collected using theta probes and in situ sensors. The highest performance statistics combination in terms of high correlation ( ${r}$ ), low root-mean-square error ( RMSE ), and least bias has been obtained with SCA-H using the Mironov dielectric model ( $r = 0.79$ ; $RMSE = 0.04\;{m^3}/{m^3}$ ; $bias = 0.01$ ) followed by Dobson ( $r = 0.76$ ; $RMSE = 0.04\;{m^3}/{m^3}$ , $bias = - 0.01$ ), Wang and Schmugge ( $r = 0.79$ ; $RMSE = 0.04\;{m^3}/{m^3}$ , $bias = 0.02$ ) and Hallikainen ( $r = 0.76$ ; $RMSE = 0.06\;{m^3}/{m^3}$ , $bias = 0.04$ ). Although the performance of the four dielectric models is relatively comparable, this analysis indicates that the Mironov dielectric model is marginally better than others for passive-only microwave soil moisture retrieval and could be a useful choice for SMAP satellite soil moisture retrieval.