Multicomponent georadar data: Some important implications for data acquisition and processing

Many seismic reflection processing techniques are applied routinely to grou nd-penetrating radar (georadar or GPR) data Although similarities exist bet ween seismic (acoustic) and radar wave propagation there are some significa nt differences some of the most important of which an associated with the d ipole nature (1) of georadar sources and receivers and (2) of elemental sou rces used to represent scattering bodies. Neglecting the dipole character o f electromagnetic surveys may result in incomplete or biased images of the subsurface. In an attempt to understand better the consequences of recordin g dipolar wavefields, we have simulated numerous multicomponent georadar da ta sets. These simulations an based on the weak scattering (Born) approxima tion, such that point heterogeneities in the subsurface can be represented by infinitesimal dipoles with moments parallel and proportional to the inci dent georadar wavefields. The effects of depolarization and dispersion are not included. Nevertheless, many subsurface structures can be modeled by su ites of appropriately distributed infinitesimal dipoles. Georadar images of even the simplest subsurface structures are shown to depend strongly on th e relative orientations and positions of the source and receiver antennas. A positive aspect of dipolar wavefields is that multicomponent georadar pro files contain information on the locations of both in-plane and out-of-plan e structures. Furthermore, "pseudoscalar" wavefields can be simulated from coincident georadar data sets acquired with two pairs of parallel source-re ceiver antennas one oriented perpendicular to the other. Pseudoscalar geora dar data, which are characterized by low degrees of directionality, can be processed (including migration) confidently using standard seismic processi ng software (assuming that dispersion is not a major problem). To illustrat e the advantages of multicomponent georadar data, two field examples are pr esented. One demonstrates the value of recording dual-component georadar da ta along isolated profiles; the other shows the benefits of combining 3-D g eoradar data sets acquired with dual component source-receiver antenna pair s to form pseudoscalar wavefield images.