Refraction Seismic Survey to Determine Rippability
by Jim Hasbrouck

NOTE: The following is a summary of the Results of the Geophysical Investigation. This paper can be found in its entiretyj at the National Technical Information Service (NTIS) under the identification number DOE/ID/12584-37.

A refraction seismic survey was conducted to measure compressional-wave seismic velocities to aid in the evaluation of the rippability and/or excavability of the subsurface. Refraction seismic data were acquired along five lines selected by the client. Stakes were placed at the end of each seismic line and stations along the lines were marked with flagging, but land surveying to the client’s coordinate grid was not performed, therefore no location map is included with this letter report.

The 24-channel refraction seismic data were acquired with 30-Hz geophones, and a 16-pound sledgehammer source. The geophones were located 10 feet apart and source impacts were made at various distances offset and along the seismic profile. The geophones were located on a straight line and distances were measured with a tape. Relative elevations were surveyed with a level and stadia rod. The seismic data were stacked, nominally, eight times at each source point to increase the signal-to-noise ratio. Stacking, or signal enhancement, involved repeated source impacts at the same point into the same set of geophones. For each source point, the stacked data were recorded into the same seismic data file, or record, and, from each impact and thus was enhanced while noise was random and tended to be reduced or canceled. Overall, the quality of the seismic data was excellent and easily identifiable first breaks (first arrival of seismic energy) were present.

The refraction seismic data were processed and interpreted. The general processing and interpretation flow consisted of the initial selection, or "picking", of the seismic first breaks, creation of data files for input into the interpretation program, and interpretation of the data using modeling and iterative ray-tracing techniques. The program uses the delay-time method to obtain a first-approximation depth model, which is then trimmed by a series of ray-tracing and model-adjustment iterations to minimize any discrepancies between the picked arrival times and corresponding times traced through the 2.5-dimensional cross-sectional model. For the direct arrivals through the first layer, the velocity is computed by dividing the distances (relative to elevation and horizontal, versus slope, distance) from each source point to each geophone by the corresponding arrival times. These individual velocities are averaged for each source point, and a weighted average is computed. For layers beneath the first layer, velocities are computed by two methods: 1) Regression, in which a straight line is fit by least squares to the arrival times representing the velocity layer and average velocities are computed by taking the reciprocals of the weighted average of the slopes of the regression lines, and 2) the Hobson-Overton method wherein velocities are computed if there are reciprocal arrivals from two opposing source points at two or more geophones. The final velocities are computed by taking an average of the two methods.

Figures 1 through 5 are the relative elevation versus distance refraction seismic depth models, with annotated average velocities for each layer, for lines 1, I extension, 2&5, 6, and 7, respectively. Figures 1 through 5 were constructed using the depth model data, and the estimated total depth of investigation was computed by simply subtracting 60 feet from the relative surface elevation. In refraction surveys, depth of investigation is related both to the length of the surface spread of geophones and source points, and the expected subsurface velocities. Since basement in the survey area probably consists of relatively fast velocity material (assumed greater than 8000 feet/second), the first geophone to "see" a refraction from that layer would be at a distance of 3 to 4 times the expected depth (if 60 feet is assumed, then that geophone would be at 180 to 240 feet along the spread, but probably closer to the 180 feet because of a relatively large velocity contrast between the basement and overlying sediment velocities). Since a refraction was not apparent within the data from a third layer along any of the lines, only an estimate of depth of investigation can be made. For the figures, a conservative estimate of 60 feet was chosen (total spread length of 240 feet divided by 4), but the depth of investigation could be deeper (i.e., 80 feet or 240 divided by 3). Again, however, without a refraction from the third layer, the depth is only an estimate.

Figure 1

As discussed above, only two layer refraction seismic depth models were computed for each line since no refraction from a third layer is present within any of the data. The layer I velocities range from 1500 to 1928 feet/second, which is consistent with that expected from unconsolidated sediments, while the second layer velocities range from 3390 to 4237 feet/second, which is indicative of the Gila Conglomerate. However, due to the averaging nature of the computation of the seismic velocities, as previously discussed, and minor changes in the surface or near-surface (a few feet), it would be more geophysically correct to state that the first layer velocities are about 1700 feet/second, and second layer velocities are around 4000 feet/second. Using these geophysically estimated velocities for the subsurface in conjunction with tables prepared by the Caterpillar Tractor Company, it should be possible to rip to the estimated depth of investigation with a D9, D8, or D7 ripper. However, marginal rippability occurs for conglomerates at about 4500 feet/second for a D7 ripper, so it would probably be more cost effective to use a D8 or D9 since the velocities of the second layer approach the limits of the D7 ripper.

Figure 2

Figure 3

Figure 4

Figure 5