Michaël Antoine

Electromagnetic inversion of GPR signals and subsequent hydrodynamic inversion to estimate effective vadose zone hydraulic properties

We combine electromagnetic inversion of ground penetrating radar (GPR) signals with hydrodynamic inverse modeling to identify the effective soil hydraulic properties of a sand in laboratory conditions. Ground penetrating radar provides soil moisture time series that are subsequently used as input in the hydrodynamic inverse procedure. The technique relies on an ultrawide band (UWB) stepped frequency continuous wave (SFCW) radar combined with an off-ground monostatic transverse electromagnetic (TEM) horn antenna. Ground penetrating radar signal forward modeling is based on the exact solution of the three-dimensional Maxwell equations for describing free wave propagation and on linear systems in series and parallel for describing wave propagation in the antenna. Water flow in the sand is described by the one-dimensional Richards equation using the Mualem-van Genuchten parameterization. Both model inversions are formulated by the classical least-squares problem and are performed iteratively using advanced global optimization techniques. Compared with time domain reflectometry (TDR), results demonstrated the appropriateness the GPR integrated approach to measure soil moisture remotely. In particular, the approach was found to be less sensitive to the inherent small-scale heterogeneities. Hydrodynamic inversion of soil moisture data led to hydraulic parameters agreeing reasonably well with direct measurements. The observed discrepancies were attributed to the different characterization scales and samples. The overall integrated approach offers great promise to map the effective hydraulic properties of the shallow subsurface at a high spatial resolution.

Effect of soil roughness on the inversion of off-ground monostatic GPR signal for noninvasive quantification of soil properties

We report on a laboratory experiment that investigates the effect of soil surface roughness on the identification of the soil electromagnetic properties from full-wave inversion of ground-penetrating radar (GPR) data in the frequency domain. The GPR system consists of an ultrawide band stepped-frequency continuous-wave radar combined with an off-ground monostatic horn antenna. Radar measurements were performed above a rectangular container filled with a loose sandy soil subject to seven water contents and four random surface roughnesses, including a smooth surface as reference. Compared to previous studies, we have reduced the modeling error of the GPR signal for the smooth surface case thanks to improved antenna transfer functions by solving an overdetermined system of equations based on six model configurations instead of only three. Then, the continuously increasing effect of surface roughness on the radar signal with respect to frequency is clearly observed. In close accordance with Rayleighs criterion, both the radar signal and the inversely estimated parameters are not significantly affected if the surface protuberances are smaller than one eighth of a wavelength. In addition, when this criterion is not respected, errors are made in the estimated parameters, but the inverse solution remains stable. This demonstrates the promising perspectives for application of GPR for noninvasive water content estimation in agricultural and environmental field applications.