An Improved Microwave Semiempirical Model for the Dielectric Behavior of Moist Soils

Abstract

Soil semiempirical dielectric models (SEMs) are powerful, and they are generally considered a useful hybrid of both empirical and physical models. In this paper, the Wang–Schmugge dielectric model is improved to more accurately estimate the relative complex dielectric constants (CDCs) of moist soils. Instead of the Debye relaxation spectrum of liquid water located outside of the soil (i.e., free out-of-soil water) adopted in the Wang–Schmugge model, the Debye relaxation formula related to the free-water component inside the soil [i.e., free soil water (FSW)], which is correlated with the soil texture, is employed in the improved SEM. In addition, the effective conductivity loss term related to both soil texture and soil moisture is introduced to explain the ionic conductivity losses of FSW. Since the soil moisture influence is reduced at high frequencies, the effective conductivity loss term related to only the soil texture is also analyzed for 14–18 GHz. As in the Wang–Schmugge model, the relative CDC of bound soil water varies with the soil volumetric moisture content when the soil moisture is lower than the maximum bound water fraction in the new model, which takes a different approach than the Mironov mineralogy-based SEM. The proposed model obtains better fitting results than the three most widely employed SEMs. The improved model exhibits a significantly improved accuracy with a higher correlation coefficient ( $R^{2}$ ), a closer 1:1 relationship, and a lower root-mean-square error, including in the L-band, and especially in the imaginary part of the L-band.