Dielectric Constant Assessment and Monitoring of Backscattering Radar Data in Saline Soil Using Semi-Empirical Models

Soil salinity is one of the serious environmental hazard and agriculture problems in arid and semiarid areas. Monitoring and assessing this phenomenon in affected areas is restricted and has been associated with poor results in optical remote sensing. Using potential of radar remote sensing can be effective approach for detecting salinization in salt affected areas due to the sensitivity of signal to dielectric properties of the target. The dielectric constant, which is considered as inherent characteristics of the surface, impacts the radar backscattering coefficient and is an important factor for simulating the radar backscattering coefficient. This paper deals with modeling the backscattering coefficient of salt affected soils by evaluating its dielectric constant behavior. For this purpose, semi-empirical model, which proposed by Wang were used and dielectric constant was estimated in both real and imaginary parts, as a function of salinity and volumetric soil moisture content for different types of soil texture in different frequencies. We also used two backscattering models, Dubois and Oh, to simulate backscattering coefficient of the Sentinel-1 data in VV and VH polarization. The results showed that real part of dielectric constant is influenced by soil moisture and increases with amount of water content up to 50. Imaginary part is also influenced by both salinity and moisture content and increase up to 40. The results also indicated that both real and imaginary parts of dielectric constant decrease sharply with increasing frequency, which reveal the necessity of using low frequencies to detect salt affected soils due to its higher sensitivity to salinity and moisture content in these frequencies. Evaluating the results of simulating the backscattering coefficients in VV and VH polarization, showing the RMSE=2.9554, Bias=-1.5138 for Dubois model in VV polarization and RMSE=2.0704, Bias=0.9839 for Oh model in VH polarization. It can be concluded that used models have great potential for modeling radar backscattering in salt affected soils.