Numerical modeling of GPR wavefields using ray-based, Fourier, and finite-difference algorithms with applications to field data

Ground-penetrating radar (GPR) has gained wide attention in the past few years because of its use in a wide range of applications for revealing near-surface structure with high resolution. GPR data processing is important to maximize the advantage of information in field data, and essential to interpretation of GPR data from complicated environments. GPR data processing methods are often borrowed from seismic data processing, but also have their own characteristics in many aspects.; Ray and Fourier algorithms are viable for numerical simulation of 2-D monostatic GPR data, with different strengths and limitations. They are used to simulate the main features in two field data sets, one from a Quaternary fluvial/aeolian environment, and one from a Cretaceous marine carbonate environment.; Ray-based numerical simulations of monostatic and bistatic GPR responses for several tank and pipe configurations reveal the potential for noninvasive diagnostic evaluations. Simulations can reproduce the salient features of field GPR data recorded over a metal pipe, and over plastic pipes field with air, fresh water, and salt water.; A field feasibility test over a collapsed-paleocave system in the Lower Ordovician Ellenburger dolomites in central Texas shows the ability to image both large and small scale features displayed in adjacent quarry walls. This shows potential feasibility for detailed study of carbonate facies and features from GPR data and suggests 3-D surveys as a desirable next step. Potential applications include 3-D characterization of analogs of collapsed-paleocave hydrocarbon reservoirs.; Amplitude versus offset (AVO) curves are extracted from 2.5-D finite-difference simulations of GPR reflection responses, of a layered model, to horizontal transverse-electric dipole excitation at the air-earth interface. The model parameterization is in terms of Cole-Cole relaxation mechanisms to allow explicit separation of the effects, of conductivity and the frequency-dependent components of the complex dielectric permittivity and of the complex magnetic permeability, in reflection responses as a function of antenna separation. The potential is shown for extraction of detailed electromagnetic properties from GPR AVO observations. It is necessary to use frequency-dependent simulations/predictions in analysis of field GPR data when amplitude information is to be interpreted in terms of the distribution of electromagnetic properties.