GPR Imaging of Prehistoric Animal Bone-beds

Abstract

This research investigates the detection capabilities of Ground-penetrating radar for imaging prehistoric animal bone-beds. The first step of this investigation was to determine the dielectric properties of modern animal bone as a proxy for applying non-invasive ground-penetrating radar (GPR) for detecting prehistoric animal remains. Over 90 thin section samples were cut from four different modern faunal skeleton remains: bison, cow, deer, and elk. One sample of prehistoric mammoth core was also analyzed. Sample dielectric properties (relative permittivity, loss factor, and loss-tangent values) were measured with an impedance analyzer over frequencies ranging from 10 MHz to 1 GHz. The results reveal statistically significant dielectric-property differences among different animal fauna, as well as variation as a function of frequency. The measured sample permittivity values were then compared to modeled sample permittivity values using common dielectric-mixing models. The dielectric mixing models were used to report out new reported values of dry bone mineral of 3-5 in the frequency range of 10 MHz to 1 GHz. The second half of this research collected controlled GPR experiments over a sandbox containing buried bison bone elements to evaluate GPR detection capabilities of buried animal bone. The results of the controlled GPR sandbox tests were then compared to numerical models in order to predict the ability of GPR to detect buried animal bone given a variety of different depositional factors, the size and orientation of the bone target and the degree of bone weathering. The radar profiles show that GPR is an effective method for imaging the horizontal and vertical extent of buried animal bone. However, increased bone weathering and increased bone dip were both found to affect GPR reflection signal strength. Finally, the controlled sandbox experiments were also utilized to investigate the impact of survey design for imaging buried animal bone. In particular, the effects of GPR antenna orientation relative to the survey line (broad-side mode versus end-fire mode) and polarization effects of the buried bone targets were investigated. The results reveal that animal bone does exhibit polarization effects. However, the polarization results are greatly affected by the irregular shape and size of the bone, which ultimately limits the potential usefulness of trying to utilize polarization data to determine the orientation of buried bone targets. In regard to antenna orientation, end-fire mode was found to have little difference in amplitude response as compared to the more commonly used broad-side mode and in fact sometimes outperformed the broad-side mode. Future GPR investigations should consider utilizing multiple antenna orientations during data collection.