Three seismic lines were acquired over the Minneola complex for this experiment (Fig. 5). Line 1 was designed to run along the center of a channel defined by both previous seismic data and well control (Figs. 4, 6, 7, 8, and 11), and to extend from a relatively thick producing sandstone to a part of the channel containing no sandstone (Figs. 4, 9,10,11). Because of Fresnel zone effects and sideswipe from the sides of the channel, as well as a possibility of improper migration due to an oblique cut of the channel margins, two other lines were acquired which extend perpendicular to the channel associated with Line 1 and other channels (Figs. 4, 6, 7, 8, and 11). These lines were designed to compare the seismic signature associated with sandstones of relatively equal thickness, as well as an area of a channel which contains sandstone versus an area of the channel which contains none (Figs. 4, 9,10,11).
A mini-hole dynamite source was used for economic reasons and also because small dynamite charges typically yield good high frequency data. Line 1 was acquired using a 2.5# Seisgel charge at 20 ft. depth, tamped with a sand-gravel mixture to prevent blowouts. Lines 2 and 3 were acquired at the same depth but with a 2# shaped Pentalite charge. Deeper holes would have been better but would have raised acquisition costs considerably. All three lines were acquired using 30 Hz geophones (12 to a group) arranged in a small linear array on line 1 (12 over 20 ft. centered on the station) and a small circular array on lines 2 and 3 (10 ft. diameter centered on the station). The small arrays were used to prevent cancellation of high frequency data. All the geophones were buried to enhance coupling and reduce any wind noise. Line geometry was a 1 10 ft. group interval, 220 ft shot interval between stations, 55 ft. minimum offset, 6655 ft. maximum offset. A low cut filter of 36 Hz and high cut of 250 Hz were thought to be used but may not have been. Data were also acquired without a 60 Hz notch filter to prevent phase problems and frequency holes. The system used to acquire the data was a 21 bit, 120 channel Bison system. It was hoped that the large dynamic range of the system would allow the lower amplitude higher frequencies to be recorded even though the 60 and 120 Hz electrical noise was large. Unfortunately, that was not the case.
Some test shots that were acquired prior to acquisition of Line 1 are shown in Figures 16 and 17. These test shots were acquired to determine the affect of varying charge size, charge type, depth of hole, and hole packing material. As can be seen on every shot gather, 60 Hz power line noise and its first octave at 120 Hz overwhelmed the signal on many traces. Because of the 21 bit system, signal was still recorded, but at much lower amplitudes. Winds were relatively calm during acquisition and the geophones were buried so wind noise should not be a problem. However the ground was saturated due to snow melt and rains. This combined with moisture in the cables and phones probably enhanced the electrical noise. Although the saturated ground conditions probably enhanced the source and receiver coupling with the ground, thus resulting in more transmitted and recorded energy, the benefit was overshadowed by the increased electrical noise in the cables. In addition to isolated electrical noise in individual channels, cross coupling of noise between channels is indicated by horizontal bands of noise on many channels and may have been due to moisture in the cable (Figs. 16 and 17). This cross coupling may have also affected recorded signal, particularly degrading the higher frequency component of the signal. In hindsight, delay of acquisition until the ground dried up and the various ponds that formed in low areas evaporated may have been prudent. However, economic and time considerations prevented that.
Comparison of a 2# directed, quick-bum Pentalite shot at 20 ft. with a 20 ft. sand-gravel pack (Figs. 16a and 17a) with a 2# directed, quick bum Pentalite shot at 20 ft with just the cuttings filling the hole (Figs. 16b and 17b) shows that there is not much improvement in signal quality with using a gravel-sand pack, at least for the small charge sizes used in this experiment. There is a slight reduction in ground-roll and a slight improvement in reflection signal with the sand-gravel pack, but possibly not enough to justify the cost. However, since all the holes were drilled on the production lines before the tests were run (due to economic considerations), they were sand-gravel packed just to make sure there were no blow-outs. Comparison of the 2# Pentalite shot at 20 ft. with a cuttings pack (Figs. 16b and 17b) with a 2.5# conventional Seisgel charge at 20 ft. with a cuttings pack (Figs. 16c and 17c) suggests that although the Seisgel charge gave off more energy, it created more ground roll, higher amplitude direct and refracted waves, and a lower frequency signal. However, the Seisgel signal was slightly stronger on the farther offsets. Because Line I was shot with the 2.5# Seisgel charge and Lines 2 and 3 were shot with the 2# Pentalite charge, Line I should have a higher power signal weighted more towards the lower frequency end than Lines 2 and 3. Line 1 should also have had higher amplitude ground roll masking the reflection data. However Lines 2 and 3 probably had a lower amplitude signal. Comparison of the 2.5# Seisgel shot (Figs. 16c and 17c) with a 5# Seisgel shot (Figs. 16d and 17d) show that the 5# shot has considerably more energy than the 2.5# shot. Reflections are even clearly visible down to I second on the unguided shot gather for the 5# shot (Fig. 16d). Reflections from the 5# shot also show up on the noisy traces in the gained shot gather (Fig. 17d). The only drawback of the 5# shot though is higher amplitude ground-roll, direct arrival, and refraction's, as well as a reflection amplitude spectrum that seems weighted more towards the low end than the 2.5# shot. In summary, if economic conditions (down time costs for the recording crew) allowed testing of these shot sizes prior to drilling out the production lines, a 5# or more Seisgel or 4# or more Pentalite charge would have been used to boost the signal-to noise ratio and allow a higher amplitude signal to be recorded on the noisy traces.
A typical processing flow was used for each line. It is worth noting here though that spectral balancing substituted for notch filters. Notch filters were not used in order to prevent phase problems. Spiking deconvolution was also applied to the data to boost the high frequencies and change the wavelet from minimum to zero phase. Final display of both the stacked and migrated sections are reverse the field polarity because the sections tie better with the previously acquired seismic data and with a normal polarity synthetic better. This suggests that peaks on the final sections represent acoustic impedance increases while troughs indicate decreases. Despite spectral balancing and deconvolution, there still did not appear to be much signal above 60 Hz. This could have been a result of a combination of high amplitude electrical noise at 60 and 120 Hz swamping the much lower amplitude higher frequencies, crossfeed, or simply the lack of generation or recording of frequencies much above 60 Hz due to attenuation at the near surface. The fact that the near surface was composed of Loess, and the shots at the 20 ft depth could have crushed a lot of material creating an extreme loss of higher frequencies, may have contributed to the problem. It is also possible that the source did not generate enough power to allow the higher frequencies to reach the depth of investigation and return before loosing so much amplitude as to be useless.