Velocity variations and water content estimated from multi-offset, ground-penetrating radar

The common midpoint (CMP) processing technique has been shown to be effective in improving the results of ground-penetrating radar (GPR) profiling. When radar data are collected with the CMP multioffset geometry, stacking increases the signal-to-noise ratio of subsurface radar reflections and results in an improved subsurface image. An important aspect of CMP processing is normal-moveout velocity analysis. Our objectives are to show the effect of multiple velocity analyses on the stacked radar image and particularly, to demonstrate that this velocity information can also be used to determine subsurface water content. Most GPR surveys are very limited in spatial extent and assume that within the survey range, radar velocity structure in the shallow subsurface can be adequately approximated by a single velocity function in data processing. In this study, we show that variation in radar velocity can be quite significant and that the stacked profile improves as the number of velocity analysis locations is increased. Interval velocities can be calculated from the normal moveout velocities derived in the CMP velocity analysis. With some reasonable assumptions about subsurface conditions necessary for radar propagation, interval velocity can be converted to an estimate of volumetric water content. Therefore, by collecting GPR data in the multioffset CMP geometry, not only is the radar profile improved but it also allows for an interpretation of subsurface variation in water content. We show the application of these techniques to multioffset GPR data from the Chalk River test area operated by Atomic Energy of Canada Limited.