Comparison of Petrophysical Relationships for Soil Moisture Estimation using GPR Ground Waves

Abstract

Soil water content measurement using ground‐penetrating radar (GPR) requires an appropriate petrophysical relationship between the dielectric permittivity and volumetric water content of the soil. The suitability of different relationships for GPR soil water content estimation has not been thoroughly investigated under natural field conditions for a complete range of seasonal soil conditions. In this study, we examined the ability of various empirical relationships, volumetric mixing formulae, and effective medium approximations to predict near‐surface volumetric soil water content using high‐frequency direct ground wave (DGW) velocity measurements for three soil textures. The estimated water contents were compared with values obtained from gravimetric sampling. The accuracy of soil water content predictions obtained from the various relationships ranged considerably. The best predictions for the overall data set in terms of RMSE were obtained with a differential effective medium approximation based on a coated sphere model (RMSE = 0.045 m3 m−3); however, an empirical relationship (RMSE = 0.052 m3 m−3) and a volumetric mixing formula (RMSE = 0.048 m3 m−3) also performed well. These best‐fitting relationships do exhibit some degree of textural bias that should be considered in the choice of petrophysical relationship for a given data set. Further improvements in water content estimates were obtained using our best‐fit third‐order polynomial relationship (RMSE = 0.041 m3 m−3) and our three‐phase volumetric mixing formula with geometric parameter α = 0.36 (RMSE = 0.042 m3 m−3); these optimized relationships were developed using the DGW permittivity and soil water content data collected in this study.

Current uses of ground penetrating radar in groundwater-dependent ecosystems research

“…In practice, the GPR θ depends mainly on the validity of selected petrophysical relationships (Slater and Comas, 2009). Steelman and Endres (2011) reported accuracy levels after analysing the behaviour of some petrophysical relationships for several types of geological materials and antenna frequencies ( Table 2). The early-time technique estimates GPR θ without v while changes in amplitude attributes of the combined air wave and ground wave of the reflected signal collected through common offset configurations are used to map ε r (Prado et al, 2011;Ferrara et al, 2013;Algeo et al, 2016).…”

Validation of ground penetrating radar full-waveform inversion for field scale soil moisture mapping

“…Considering the whole field, the use of a unique petrophysical relationship may be affected by some limitations as this relationship may depend on the specific local soil conditions (e.g., soil texture, porosity). In addition, the use of different petrophysical models, even when calibrated with field data, can result in differences in terms of volumetric soil moisture up to 0.10 m 3 m −3 (Sambuelli, 2009;Steelman and Endres, 2010). Despite these important limitations, the comparison with soil sampling measurements is still the only method for quantifying the accuracy of the GPR method, in addition to its precision.…”

Detecting Moisture Damage in Archaeology and Cultural Heritage: a Brief Introduction

“…However, other antenna parameters, such as the signal amplitude of the wavelet, can be used to evaluate the EM properties of the material in the region near the antennas, where the signal propagates through a surface material [10] and [11]. The potential of estimating the soil's dielectric parameters using the high attenuation and temporal stretching of the EM signal amplitude has been demonstrated experimentally and validated by numerical simulations and on-site measurements [12]. In particular, the dielectric permittivity affects both the amplitude and duration of the GPR signal, so higher amplitudes and shorter wavelets are associated with lower permittivities [9, 13, and literature therein].…”

“…Note that as discussed above, the signal amplitude is affected by variations in the EM parameters and particularly by permittivity changes, so it is also possible to determine the soil's volumetric water content using an appropriate petrophysical relationship [12].…”