Sensitivity analysis of soil heterogeneity for ground-penetrating radar measurements by means of a simple modeling

Higher-frequency ground-penetrating radar (GPR) is becoming more common for various applications. While it can achieve a higher resolution than low-frequency measurements, it also becomes more sensitive to heterogeneous soils. The increased sensitivity in small-scale measurements often creates the problem of unwanted scattering in the data. Unwanted scattering is commonly called clutter. If clutter contaminates the data significantly, data analysis and interpretation become difficult. Our study analyzes the sensitivity of soil heterogeneity by evaluating the amount of clutter caused by soils. Clutter is calculated by simple modeling that takes into account soil heterogeneity. The method constructs a dielectric sphere model that uses statistical properties of permittivity distribution and calculates backscattering power from the sphere that emulates the backscattering observed from heterogeneous soil. The modeling was carried out for a range of heterogeneities. The results show that the level of permittivity variation of soil mostly dominates the clutter power. However, the influence of correlation length becomes greater when the correlation length of soil permittivity distribution is a multiple of wavelength. Therefore, to observe the influence of heterogeneous soils on GPR measurements, the spatial distribution of soil permittivity must be taken into account in addition to the variation.