P-Wave Velocity Characterization of the Boise Hydrogeophysical Research Site

Publication Date


Type of Culminating Activity


Degree Title

Master of Science in Geophysics



Major Advisor

Michael D. Knoll


Many environmental and engineering problems require knowledge of subsurface structure. This structure is often mapped using P-wave reflections off of interfaces between materials with different seismic properties. In some coarse-grained sediments, however, the strong velocity contrast between the vadose zone and the saturated zone prevents energy from reaching reflectors below the water table. In these cases, borehole seismic methods may be useful.

Borehole seismic methods consist of VSPs, with the source on the surface and the receiver in the borehole, RVSPs, with the source in the borehole and the receiver on the surface, and crosswell tomography surveys, with sources and receivers in different boreholes. The traveltimes of the first arrivals from any of these surveys can be inverted to image the seismic velocity structure of the subsurface. In this study, I have inverted the traveltimes of RVSPs and crosswell tomography data at the Boise Hydrogeophysical Research Site, a research well field in a coarse alluvial unconfined aquifer near Boise, ID.

I inverted the RVSP data using a curved-ray iterative linearized inversion. This routine used a ray-shooting forward model and an analytical approximation to the Jacobian matrix. I chose the regularization parameter using Morozov's discrepancy principle. The final velocity models produced by the inversion were generally consistent with prior information about the site.

I inverted the crosswell tomography data using Pronto, a commercially available curved-ray tomography routine. I chose the inversion parameters using L-curve analysis. The lateral structure of the tomograms depended heavily on the initial model used, but the average vertical velocity profile of the tomograms was insensitive to the initial model. The decision to interpret the tomograms in a 1-D sense was consistent with the poor horizontal resolution predicted from the wavelength of the seismic signal and the geometry of the problem. The vertical velocity profile derived from the tomogram was consistent with prior information about the site.

The RVSP- and tomography-derived velocity profiles suggest that the BHRS has a generally layered structure. One of the boundaries between these layers of different velocities is slightly deeper than predicted by neutron porosity logs. The comparison of the two types of velocity profiles suggest that RVSP-derived velocity models have high vertical resolution, but tomography-derived velocity profiles have low uncertainty.

Files over 30MB may be slow to open. For best results, right-click and select "save as..."