Luncheons

Multicomponent Case Study One Company’s Experience in Eastern Alberta

Photo Unavailable

Paul Anderson & Ron Larson
Apache Canada Ltd.

Monday, November 27th, 2006 – 10:30 AM
Telus Convention Centre 8th Ave SE, Calgary

Abstract

In 2004, Apache embarked on an investigation of multicomponent technology. One of the projects undertaken was a target of upper-Mannville channel sands, specifically the Rex member. Some of this work was presented my Monk et. Al. in the May 2006 Recorder, however additional efforts have been put into the analysis recently, in order to better understand what the multicomponent data is revealing. For background on the acquisition design and initial processing, we refer you to the May 2006 Recorder article (Monk et. al.). One point of interest is that the reservoir quality sands are very shaley, relative to conventional reservoirs, with shale content as high as 60% in producing reservoir sands.

Simple mapping carried out very early in the project showed positive results beyond expectations. Figure 1a shows a two way time map, made from the second processing version of the PP data, of the Rex coal. The drape anomalies associated with the channel system are immediately evident. We consider this to be an indirect way of mapping the channel system. The same figure also shows an RMS amplitude map, made from the first processing version of the PS data, of a 19 ms window located beneath the Rex Coal. The map clearly shows the presence and absence of the channel system. We consider this to be a direct measurement of the channel system, given that PS data should respond only to lithology, not to the fluid content within any particular lithology.

Fig. 01
Figure 1 - Left map created using PP data version 2 of 2 way time to the Rex coal. Positive structure in yellow is inferred to be drape associated with the presence of grain supported sandstone in the channel system. Right map created using PS data version 1 of RMS amplitude in a 19 ms window below the Rex coal. High RMS amplitude is believed to be a direct PS seismic response to the presence of channel lithology.

As this data was an experiment with multicomponent technology, what seismic experiment would be complete without an additional effort on reprocessing? As such a small test area was used for additional testing. Figures 2 & 3 show the newly processed PP and PS data respectively, after rotating the data to zero-phase.

Fig. 02
Figure 2 - Comparison of PP-data from 2 different contractors. The recently processed data is on the left.
Fig. 03
Figure 3 - Comparison of PS-data from 2 different contractors. The recently processed data is on the left.

Correlation of the seismic data was performed, first on the PP data then on the PS data using a PS synthetic. Well ties on the newly processed data proved to be very good over a large window for both the PP (cc: 85%) and the PS (cc: 73%) data after rotating the phase to zero relative to well control (figures 3 & 4). Well ties to the PP and PS data are shown in figures 4 and 5 respectively.

Fig. 04
Figure 4 - PP well tie. The synthetic is shown in blue and the pilot trace from the seismic is shown in red. PP-well tie shows correlation of 85% between 640 and 910 ms. The log reack to the right of the seismic shows the P-wave impedance as well as the Converted wave impedance tracks calculated from well logs.
Fig. 05
Figure 5 - PP well tie. The synthetic is shown in blue and the pilot trace from the seismic is shown in red. PS-well tie shows correlation of 73% between 600 and 1560 ms (PS-time). The log track to the right of the seismic shows the P-wave impedance as well as the Converted wave impedance tracks calculated from well logs.

REGISTRATION

Using these well ties to correlate events between the PP and PS data, registration was undertaken. Not unexpectedly, a key horizon, the Paleozoic unconformity, is a difficult reflector to pick and requires several iterations to finalize upon its picks. The difficulty being that, even though the Paleo is an weathered unconformity between Paleozoic carbonates and cretaceous clastics. As a result, the seismic response is very heterogeneous within a 3D volume, usually being a peak, but can also occur as a zero-crossing or trough, due to tuning artifacts and relative compressibility and rigidity of the weathered carbonates. As a result, a deeper reflector has been chosen at this time to use for the preliminary registration. Figure 6 shows a cross-section of the registered PS-data in PP-time and the corresponding Vp/Vs model. The Vp/Vs ratio from well data is also shown at the well location.

Fig. 06
Figure 6 - Cross-section of PS data registered to PP-time. The Vp/Vs ratio derived from horizon-based registration is shown in colour. Also shown is the Vp/Vs ratio from the well data.

As a result of the registration process, the apparent band width of the PS data, using a cut-off of 20 dB down, has increased from 8-32 Hz (figure 7) to 11-52 Hz (figure 8), an increase of approximately 63% a the high-end. As a comparison, the PP-data bandwidth, over the same geologic interval and using the same cut-off, is 11-98 Hz (figure 9).

Fig. 07
Figure 7 - PS amplitude spectrum in PS-time.
Fig. 08
Figure 8 - PS amplitude spectrum in PS-time.
Fig. 09
Figure 9 - PP amplitude spectrum in PP-time.

PS1 & PS2 ANALYSIS

Additionally, rotation analysis of the radial and transverse data was performed to evaluate the presence of shear-wave splitting. The basic principle is that as shear waves probagate through an anisotropic medium, for example a thick coal with cleats, the shear waves, Sv and Sh, become polarized in the direction of the cleats, resulting in time delays that are dependant upon azimuth for a particular CCP gather. When it is present, this time-delay can appear to be residual statics or velocity errors on the radial CCP gather as shown in figure 9. It also results in residual, coherent energy on the transverse component. Upon investigation of transverse component, residual coherent energy was found, so PS1-PS2 rotation analysis was performed. Figure 10 shows a sample CCP gather radial and transverse components for this rotation analysis. The transverse component show characteristic polarity reversals which correspond to the fast and slow shear wave azimuths. By examining the radial component we can determine that the fast shear is oriented 55° E of N and the slow is perpendicular to that at 35° E of S. The radial data shows a time delay of approximately 13 ms down to the top Mannville, however, when measuring the time delay at the Rex Coal, is is approximately 17 ms, indicating an increase in anisotropy within the upper Mannville, above the Rex coal.

Fig. 10
Figure 10 - Cross-section of PS data for the Transverse (left) component and Radial (right) component at a single CCP location for different rotations. Note that the azimuths of the polarity reversal on the transverse component match to the fast (PS1) and slow (PS2) orientations of the shear waves.

CONCLUSIONS

To date, the PS data has proven very valuable in aiding the risking of potential drilling locations. A look-back was performed to determine how many of the original wells on the program would have been drilled had the multicomponent data been incorporated. Surprisingly, the success rate would have improved dramatically, resulting in fewer dry-holes drilled, and to date, no false negative locations at this time. Work continues on this project, including post-stack inversion of the PP-data as well as Joint inversion, where the multicomponent data will be included in inverting for P-impedance, S-Impedance and Density, based upon similar work published by Hampson et. al. (2005). As well, neural network analysis will be attempted to determine if such an analysis can provide an estimate of sand quality (Anderson & Gray, 2005).

Biography

Paul Anderson received a Bachelor degree in geophysics from the University of Calgary and is currently working as a geophysicist in the Exploration and Production Technology group at Apache Canada Ltd. Paul's current responsibilities include seismic processing QC, multicomponent processing & analysis, AVO analysis and inversions, in Canada and abroad. Before joining Apache in January 2006, Paul was with Veritas GeoServices in Calgary (1998-1999, 2000-2005) and Houston (1999-2000), where his responsibilities included AVO and LMR analysis, Post stack inversions, modeling, multicomponent interpretation and fracture detection work. He has a wide range of experience in a number of geologic basins, including; Western Canada, the onshore Texas, Gulf of Mexico, Beaufort Sea & Mackenzie Delta. Paul is currently a member of APEGGA, CSEG, CSPG & SEG.

Ron Larson received a Bachelor degree in geophysics from the University of Calgary and is currently working as an interpretation geophysicist at Apache Canada Ltd. Other than "drilling lots of wells" his current responsibilities are vague – a situation he finds agreeable. Before joining Apache in August 2001, Ron was with Boyd PetroSearch (1997-2001), Pulsonic (1994-1997), Teknica (1990-1993), Canadian Orient (1987-1999) and Dome Petroleum (1983-1986). A self described itinerant geophysicist, he has experience in a number of geologic basins, including; Western Canada, Argentina, Libya, and China, and similar experience with a range of seismic methods including AVO, inversion and multicomponent seismic. Ron is currently a member of APEGGA, CSEG, and SEG.