Estimation of groundwater level by GPR in an area with multiple ambiguous reflections

Abstract

Ground-penetrating radar (GPR) was applied to estimate the groundwater level in Ulaanbaatar City, Mongolia. Multiple reflectors occur at different levels. Any one of these was a possible water table reflector. In order to identify the groundwater reflection amongst a set of ambiguous reflectors and to increase the signal-to-noise ratio, the common-midpoint (CMP) method was used. The operation frequency of the radar system was 100 MHz. The start position of a survey line was 10 m away from a building that influenced GPR data. However, a reflection from the building can be suppressed significantly by two-dimensional f– k filtering. By noise reduction, velocity analysis, normal moveout correction, stacking and migration using estimated velocities, the subsurface images are obtained. The signal-to-noise ratio of the stacked CMP profile is much higher than that of the common-offset profile. The usual detectable range of GPR is several meters, however, we can measure more than 10 m in depth because of the improvement of the signal-to-noise ratio by CMP and due to the very dry soil. Also, we can estimate the vertical dielectric constant distribution from the interval velocities obtained from the CMP. Then we can locate a groundwater reflection within a multitude of ambiguous reflectors, and find that the depth of the groundwater level is approximately 8.0 m. The experiment results show that GPR data collected by the CMP method not only improves the quality of the radar profile but also enables us to estimate the subsurface vertical profile of the dielectric constant.