Vertical Time and Observed Travel Time Inversion (vfitw.m, vplot.m, bvsp)

The BSU distribution includes a procedure, vfitw.m, which reads the picks in the trace headers of a BSEGY data set. The procedure projects the picks to the vertical, and prompts the user to select the top and bottom of an interval for a least squares linear fit. The fit curve is drawn and the user clicks at a point where the interval velocity solution will be printed (on the figure). This velocity is an estimate of the vertical group propagation velocity. If the results of vfitw.m are written to disk, a second program, vplot.m, may be used to plot the solutions with more user control over scales. Exporting the plot to xfig format will allow one to draft on the final figure.

Figure (19A) is the vplot.m generated plot of a two interval analysis of the T-component data. A significant increase in group velocity occurs at the transition from silt-clay soil to the sand (soil types determined by Soil Behavior Type (SBT) of the ConeTec survey made available at GeoLogan97). An alternative to working with vertical times is program bvsp which does an inversion based on 2 layers over a half-space. Bvsp works with the observed times, and solves for refracted and direct raypaths in a non-linear ray tracing scheme (horizontal boundaries). The bvsp result is shown in Figure (19B) (after 40 iterations). Program bvsp fits the observed arrival times by ray tracing, and is hardwired to do a 3 layer fit.

Figure 19: T-component Data Travel Time Inversions (a) Vertical Time (b) Observed Time

It should be remembered that the ConeTec and seismic survey were done at different times of year (ConeTec in November 1996, seismic in July 1997). Further, the two surveys were in the same general area, but not at exactly the same location. The lateral separation of the two could easily be 50 meters or more. For that reason, exact depth correlations are not possible.