Interim Report : Hill AFB OU-2 High-Resolution Seismic Study

11 March 1999

Sponsored by DoE and the AATDF/DoD through the EESI at Rice University.

A. Levander,C. Zelt,D. Dana, W.W. Symes


Eight personnel from the Rice University Geophysics Program conducted a high resolution seismic study at OU-2 at Hill Air Force Base from 24 August through 28 August 1998. We acquired three seismic reflection lines each 36 to 42 meters (m) in length crossing the buried channel region defined by drillholes, which acts as a collection trough for DNAPLs. The location of the seismic lines is shown in Figure 1.

Field acquisition :

Each seismic line was collected using two 24 bit 60 channel portable seismograph systems, Rice's 24060 Bison system, and IRIS/PASSCAL's Geometrics Strataview system, the former recording from sixty single 28 Hz Geospace geophones, the latter recording from sixty single 40 Hz Mark Products geophones. Both recording units were run with preamplifier gains of 24 db. Geophone spacing was 30 cm, giving line lengths of 36 to 42 m, depending on the presence of surface obstacles (concrete well pads) along the seismic lines. Each line was collected using three seismic sources: a .22 caliber rifle fired at every geophone location, an accelerated weight drop repeated 10 times at every other geophone location, and an 8 gauge shotgun fired every 10 geophone intervals. These source parameters produce CMP data with nominal fold of 60, 30, and 6, respectively. All the sources were recorded for 500 milliseconds (ms) at 0.5 ms sample rate. The seismic recording units were simultaneously triggered by the sources.

Each seismic line was surveyed from known locations such as survey markers and well heads using a surveyor's level. The first line is located in the AATDF demonstration area, and runs 42 m approximately East-West across the centerline of the channel. Line 2 was also shot in the AATDF demonstration area to the north of Line 1 and runs 36 m approximately WSW. Line 3 was located along the East side of Perimeter Road across the Griffith Pool, and extended 36 m in a North-South direction (Figure 1).

Figure 1 - 2-D Seismic Survey -- Hill AFB Seismic lines are shown with local monitoring wells. Wells specified in the text are labeled.

Data Processing:

Source records recorded on the Bison system were combined with the records from the Geometrics Strataview into seismic records of the full length of each line (120 stations). Processing of the combined records was performed using the ProMAX system, a commercial seismic reflection processing. The results presented in this report are from source records from the slide hammer (an accelerated weight drop device) only.

Trace editing of the records removed noisy traces and corrected traces with polarity reversals. Source and receiver locations from survey measurements taken during the experiment were used to assign geometry to each record. Elevation statics were applied to compensate for elevation changes between stations. The reference datum was chosen to be 1431 m (4695 feet (ft)).

Analysis of the frequency spectrum of the records shows the majority of the source energy lies within the 35-250 Hz band. This range was used as the basis for a series of narrowband filter tests to determine the best filter to minimize ambient noise (pump noise, traffic, jet noise, etc.). A Butterworth filter with a passband of 64 - 240 Hz provided the best result, revealing near surface reflections and reducing non-coherent noise.

Coherent (linear) noise remaining after trace-by-trace filtering was removed by trace mixing within each source gather along the velocities of the linear features. Finally, air blast energy was attenuated, reducing the high amplitude sound in the gathers. Minimum phase spiking deconvolution was applied to the filtered data to remove ringiness in the data, and resulting from the filtering process. To compensate for amplitude variations between the individual shots and to balance reflections from structural features in each seismic section, trace equalization was applied to each trace prior to stacking. For Lines 1 and 2, trace normalization (time variant gain) was also applied to balance trace amplitudes in time.

Brute stacks were constructed for each line using a constant normal moveout (NMO) correction velocity. Although NMO velocities between 400 and 600 m/s appear to work, an NMO correction of 600 m/s produced reasonable seismic sections for Lines 1 and 2 and 400 m/s was used for Line 3.


The seismic processing sequence that we have applied to the data at this point is still extremely simple, yet on both Lines 1 and 2 we clearly see the presence of the buried channel at about the correct depths known from drilling (Figures 2a-b,3a-b green line, alternate interpretation = red dashed line). Line 2 particularly shows the base of the channel extending from X=13 to 19 m along the line at 62 ms or about 15.5 m (51 ft) depth. Line 3 shows a continuous reflector at about 45 ms (12 m / 38 ft) (Figure 5b, blue line); we do not know what geology to expect here and do not present a geologic interpretation of the event. Based on this level of processing and the Line 2 section we believe that we have a successful result.

Well U2-050 is adjacent to Line 2 (approx station 155 - 10.5 meters into the line). The contact between the clay and the overburden was logged at 13.2 m (43.2 ft). The reflector in the section is at ~45 ms at that point in the line. Using the logged depths, we would estimate a bulk velocity of 520 m/s above the reflector, roughly the same as the constant stacking velocity we have used for these sections. First break velocities for Line 2 agree with the bulk velocity estimate (375 - 600 m/s). Well SB-7 is less than 3 m from the northeast end of Line 2. The clay contact at this well is logged at 14 ft (4.3 m). Using the calculated bulk velocity, two-way time for the logged depth would be 16 ms. This suggests that the clay layer shallows rapidly off the east end of the line.

Brute stacks from the three lines show a number of reflections. On Line 1, the slide hammer section (Figure 2a-b) shows two reflections at about 50 ms (depth converted to 11.5 m / 38 ft) at X=4 m along the line and 70 ms (18 m /55 ft) at X=10 m. These reflections correspond in depth and location to the center of the channel and its eastern flank. Here the channel appears to be very irregular, with a steep western flank which shoals at X=20 m at 25 ms (6.5 m / 21 ft), and then deepens gradually to about 37 ms (9.6 m / 32 ft) at X=40 m. Well U2-060 is adjacent to station 78 (X = 13.8 m) on Line 1. The logged depth to the clay layer is 45.1 ft (14 m). Using the previously calculated bulk velocity of 520 m/s, the top of the clay layer should lie at 58 ms in the section. This places it near the interpreted base of the channel.

Figure 2a - Line 1 Slide Hammer - Brute stack 600 m/s uninterpreted seismic section. Adjacent well is marked as is the point at which the line crosses the treatment line (pipe).

Figure 2b - Line 1 Slide Hammer - Brute stack 600 m/s interpreted seismic section. The clay layer is marked in green, with an alternate interpretation marked with a dashed red line. The interpreted base of the channel is also labeled.

In contrast on Line 2 we see a continuous reflection event at 30 ms (8 m / 26 ft) at the eastern end, deepening to about 55-62 ms (14-16 m / 47-53 ft) at the deepest point beneath the deepest spot at X=16 m, and then shallowing to 22ms (6 m / 19 ft) at about X=24 m and staying at about that depth to the western end.

Figure 3a - Line 2 Slide Hammer - Brute stack 600 m/s uninterpreted seismic section. Adjacent well is marked as is the point at which the line crosses the treatment line (pipe).

Figure 3b - Line 2 Slide Hammer - Brute stack 600 m/s interpreted seismic section. The clay layer is marked in green, with an alternate interpretation marked with a dashed red line. The interpreted base of the channel is also labeled.

The west end of Line 2 is within 1.5 meters of Line 1 at approximately X=35.5 m along Line 1. Figure 4 is a side by side comparison of the end of Line 2 with the adjacent portion of Line 1. Both sections agree, showing the clay reflector at 30 ms (8 m / 26 ft).

Figure 4 - Comparison of Lines 1 and 2 Portions of Lines 1 and 2 are shown side by side for comparison. The nearest points on each line are marked.

Line 3 includes a fairly flat reflector at about 40-47 ms (10.4-12 m / 34-40 ft) extending across the entire line (Figure 5a-b).

Figure 5a - Line 3 Slide Hammer - Brute stack 400 m/s uninterpreted seismic section.

Figure 5b - Line 3 Slide Hammer - Brute stack 400 m/s interpreted seismic section. The unidentified layer is marked in blue.

At this point we have not rigorously examined the seismic velocity field, this is the next step in the seismic analysis. Reflector depths may be off by 25% due to uncertainty in seismic velocity.

Results of traveltime inversion of first arrival times.
back to CCG
back to Diana's home page