| Tomographic ground penetrating radar (GPR) traveltime data has been widely used for hydrologic parameter estimation in the shallow subsurface. However, hydrogeological information extracted from tomographic GPR data is subject to great uncertainty because of the nonlinear and non-unique relationships between hydrogeological and geophysical attributes, and the spatial heterogeneity of these attributes. Good data acquisition design can make the hydrogeological parameter estimation problems less ill-posed. However, few efforts have been made to systematically evaluate and compare the impacts of different data acquisition and field parameters (borehole geometry, antenna locations and spacing, soil type, soil water saturation conditions, etc.) on the borehole radar responses.;Here, a stochastic approach is adopted and multiple flow fields are simulated by generating various hydraulic conductivity fields, and varying hydraulic flow parameters as well as initial and boundary conditions. Given different data acquisition parameters (field geometry, antenna settings, etc.), the tomographic radar responses are computed. A finite-difference method is used to solve the Eikonal equation to obtain accurate radar travel times, by taking all possible wave paths into account through a local traveltime computation algorithm. The Eikonal solver method computes the corresponding GPR responses for each field given a particular set of data acquisition factors and field parameters. Besides the influences of data acquisition parameters (antenna spacing, antenna location) on radar responses, we also studied the impacts of noise levels in radar arrivals times, time of data collection following infiltration events. Meanwhile we evaluated the effects of flow parameters such as hydraulic conductivity and water retention model parameters, as well as various flow boundary conditions. The impact of data acquisition parameters on the applicability of the borehole radar data can be systematically evaluated and compared with flow parameters, thus provide useful information on borehole radar data acquisition design. The hypotheses of this study include: (1) design parameters have comparable impacts on tomographic radar responses to hydrologic parameters of interest in hydrogeological characterization; (2) optimum/reasonable data acquisition design enhances the sensitivities of radar responses to hydrological parameters, and thus improves hydrological parameter estimation using radar data. |
