In this article, the impact of volumetric moisture from 0% to 40% (vertical moisture profiles), dry bulk density from 0.4 to 1.8 g/cm3, clay content from 0.15 to 0.55 g/g (agrosoil type), standard deviations heights of surface roughness from 0 to 4 cm of soil on the variations of reflection coefficient in the ultra-wide frequency band from 100-400 MHz to 1.26-2.4 GHz were investigated. The reflection coefficient was calculated for a smooth and rough soil surface. Mironov's two-relaxation dielectric model was used (input parameters: dry bulk density, clay content, volumetric moisture, frequency of electromagnetic wave) to calculating the reflection coefficient. It is shown that the reflection coefficient is an ambiguous function of the clay content and the dry bulk density; the error in retrieving high values of soil moisture can be 5 times higher than for dry soils. In relation to the clay content and the dry bulk density, the soil roughness is the dominant parameter, which uncertainty, significantly effects on the error of soil moisture retrieving. For the considered various vertical profiles of soil moisture, it is shown that the effective thickness of topsoil, which forms the reflection coefficient, does not exceed 2 cm at a sensing frequency above 1 GHz. When the volumetric moisture content of soil surface is more than 28%, the reflection coefficient in the frequency range from 433 MHz to 1.26 GHz does not depend on the vertical distribution of moisture in soil. The conducted studies establish quantitative limitations on the accuracy of soil moisture retrieval for a given error in the initial parameters of the model: the clay content, the dry bulk density, soil surface roughness, which should be taken into account when developing algorithms for remote sensing of agrosoil moisture.
Read full abstract