The detection and characterization of fractures is important in many engineering practices e.g., rock fall assessment, quarry rock quality determination and mine roof and tunnel stability. Fractures may provide suitable contrasts in electrical properties for detection by GPR instruments. Their ability to reflect radar waves and the dependence of this reflectivity on fracture properties (aperture and fill) makes the GPR method a promising tool for rock fracture characterization. Doing so successfully requires quantification of the reflectivity, reflection coefficient, R, of individual fractures and its variation with incidence angle (AVA), or in practice, offset. Measuring the full AVA response using Common Mid Point (CMP) surveys enables simultaneous estimation of fracture aperture and fill permittivity, difficult to achieve with Common Offset (CO) profiling which does not evaluate angle dependence. This paper reports on AVA analysis carried out on CMP data acquired with 500 MHz antennae over Carboniferous Limestone with horizontal bedding plane fractures in Yorkshire, United Kingdom. Using the transverse electric (TE) polarization mode, data were collected at a sampling interval of 0.05 ns for offsets at 0.04 m steps from 0.37–16.05m. The recorded amplitude is related to the reflection coefficient but is also influenced by other factors we eliminated by making some simplifying assumptions and amplitude corrections. We assume frequency‐independent electrical properties, constant antenna coupling with the ground surface, constant losses associated with transmission losses through any interfaces above the target reflection for the range of offsets considered and a flat homogeneous surface over which measurements are made. Amplitudes were corrected for conductive attenuation, spherical spreading and antenna patterns, which we measured through transillumination surveys across limestone boulders and numerical modelling using the GprMax 3D modelling code. To constrain fracture aperture and fill permittivity, we use the least squares fit of normalized reflection coefficient curves to corrected CMP amplitudes. The analysis allowed characterization of the (bedding plane) fracture fill relative permittivity (~7.8) and aperture (~0.043 m, ~0.2λ at 500 MHz). The values obtained are consistent with field observations of fracture fill, corresponding to a mixture of clays and calcite.

No abstract