GPR Pulse Attenuation in a Fine-Grained and Partially Contaminated Formation

Abstract

We have determined attenuation rates and relative dielectric permittivities of a fine-grained sediment formation, part of which was contaminated by hydrocarbons. Our purpose was to test previous laboratory documentation of the effects of probable adsorption and∕or capillary tension on the electrical properties of water at a common ground-penetrating radar center frequency of [Formula: see text]. The formation is a saturated, and dense glacio-marine diamicton that had a small area partially contaminated by jet fuel. Surface and trench observations showed saturation near 30% by volume, uniform and high density, resistivity at [Formula: see text], and a total hydrocarbon content of less than 0.1% within a small zone. The local hydrocarbon contamination and a spatial gradient in grain size allowed the possibility of seeing if an increase of adsorbed water and∕or relief of capillary tension within interstitial water affected the attenuation rates. We calculated the rates from the amplitude decay of ground surface wave pulses within a series of WARR (moveout) profiles recorded along a single transect. Our rates of [Formula: see text] in the coarser-grained and contaminated zone can be accounted for by the resistivity values, while those of [Formula: see text] in the uncontaminated and finer-grained zone require additional dielectric loss. The relative permittivities varied from about 17 to 22, for which modeling strongly suggests a free water value of dielectric permittivity throughout despite the fine grain content, with a mild trend toward lower values in the finer-grained zone. A strong correlation was found between attenuation rates and clay content, and only a weak correlation with hydrocarbon content. An empirical mixing theory suggests that a [Formula: see text] relaxation frequency can explain the attenuation rates in the uncontaminated zone and that DC resistivity alone falls far short of predicting the correct rates. The hydrocarbon content analysis suggests that if a threshold level for the hydrocarbon attenuation effect exists, it is not higher than about [Formula: see text]. In addition, offset surveys appear to be a useful field approach for measuring field attenuation rates.

Bayesian frequency-domain blind deconvolution of ground-penetrating radar data

“…The assumption that the embedded wavelets are stationary within the estimation and application data window is critical for reliable deconvolution. Whereas dispersive attenuation was considered to be insignificant for our field data example, wavelet dispersion and frequency-dependent attenuation are observed in GPR data, for example for clay-rich sediments (e.g., Arcone et al, 2008) or contaminated formations (e.g., Arcone et al, 2003). In cases of non-stationary wavelets, either dispersion removal techniques (e.g., Irving and Knight, 2003) can be applied beforehand or a time-varying deconvolution procedure can be carried out (e.g., van der Baan and Fomel, 2009).…”

10. Permittivity Structure Derived from Group Velocities of Guided GPR Pulses

“…Most of these studies focus on guided waves in single profiles. Few studies highlight the analysis of guided-wave phenomena in a roll-along profile, as is done by Arcone et al (2003a) to estimate attenuation rates. Here, we briefly review previous research on GPR guided waves.…”