GPR reflection and dispersion analysis methods for water content: Multi-year study of GPR estimates and soil core measurements of water content

A multi-year series of ground penetrating radar (GPR) measurements, complemented by contemporaneous soil cores, were collected at a single location in south-eastern New England, USA. The shallow subsurface is characterized by a 0.9 m thick sandy soil layer which overlies a gravelly-sand layer. Over a number of months, and different soil moisture conditions, 30 common midpoint (CMP) soundings were collected, and on each of the days, co-located soil cores were taken for analysis and comparison with the soil water content from the GPR velocities. GPR velocities were estimated using two independent methods: standard normal moveout (NMO) analysis of reflected traveltimes, and the analysis of frequency-dependent velocity dispersion of shallow GPR waveguide modes. A comparison of GPR estimated water content in the field versus gravimetrically measured water content in the lab provides a site-specific Topp-like empirical basis for predicting soil water conditions from GPR data. Although such site-specific relations may be useful, we find that for our site comparing the results of this study with the conventional Topp relation reconfirms the continuing utility of the latter.