3.1 P-Wave Interpretation
Compressional surface seismic data from Vacuum Field were interpreted in order to provide a structural and stratigraphic framework for multicomponent analysis. This interpretation is focused primarily on RCP Phase VII P-wave data, but also relies on the results of Talley (1997) from the regional Maljamar 3D P-wave volume.
The Maljamar 3-D P-wave seismic dataset provides a regional structural and stratigraphic view of the subsurface at Vacuum Field (Figure 3.1). The change in horizon dip at the Permian margin is visible. The N-S extent of RCP 3-D seismic volumes is represented by dashed black lines in Figure 3.1. The RCP surface seismic volumes are shelfward of the margin present at San Andres time, although the Glorieta shows dip changes associated with the slope.
The RCP multicomponent surface seismic volumes interpreted in this thesis were acquired in December of 1997. The multicomponent datasets were acquired as a pre-injection baseline for RCP Phase VII time-lapse analysis. Figure 3.2 is a basemap of the survey area. This interpretation focuses on the central portion of the seismic survey area ("Study Area" in Figure 3.2, lines 30-110, traces 30-110), as this area is minimally affected by low-fold and migration-edge effects. Phase VII P-wave surface seismic picks
were derived from correlation with synthetic seismograms and VSP data (see Chapter 2, this volume). Figure 3.3 shows a N-S line extracted from the RCP Phase VII surface seismic volume with key horizons annotated.
3.1.1 P-Wave Resolution
Bandpass filter and spectral analyses were performed on the Phase VII P-wave surface seismic. These analyses show a dominant frequency of approximately 50 Hz within the reservoir interval (Figure 3.4). Spectral analysis highlights the limitations of the P-wave surface seismic in resolving the stratigraphic complexities of the reservoir. Using the l /4 approximation for vertical resolution (Sheriff, 1985), and assuming a compressional velocity of 4500 m/s (15000 ft/s) within the reservoir, the required separation for resolvable events at Vacuum is 22.5 m (75 ft). Kerans, et al. (1994) describe three levels of depositional cyclicity present within the San Andres from outcrop study. Two low-order sequences (Upper San Andres and Lower San Andres) are identified. Intermediate-order, unconformity-bounded cycles range in thickness from 12-75 m (40-250 ft) in thickness, while high-order cycles range from 1-12 m (3-30 ft) in thickness. Thus, even if resolution in portions of the reservoir is higher than indicated by the l /4 approximation, all of the high-order cycles, and several of the intermediate-order cycles, are probably unresolvable from the VII surface seismic. Wang, et al. (1998b) show from synthetic studies that frequencies up to 200 Hz are necessary to successfully image
the rock-fabric units that correspond to these high-order cycles, and, to a large extent, control production within Permian shelf-margin reservoirs. Many of the petrophysical-based flow units mapped by Matt Pranter (1998) from well control at Vacuum are similar in thickness to the high-order cycles of Kerans, et al (1994). Figure 3.5 shows a cross section through the study area, with these flow unit correlations. The thickest flow unit within the San Andres averages approximately 85 ft thick. The thinnest flow unit correlated in the San Andres section is the Lovington Siltstone, which averages 15 ft in thickness. Analysis of P-wave synthetic seismograms indicates that distinct seismic character is associated with the thickest flow units. In general, however, individual seismic events represent multiple flow units (Figure 3.6).
3.1.2 Stratigraphic Interpretation
Although inadequate for resolving small- and intermediate-scale packages, P-wave surface seismic does provide a general stratigraphic framework for the reservoir interval at Vacuum Field. Figure 3.1 is a north-south line extracted from the Maljamar 3D dataset, and from this broader view, many of the large-scale stratigraphic features associated with the Permian Shelf Margin are visible. A general progradation of sediment is suggested by wedge geometries between the Glorieta and Grayburg in the southern portion of the survey.
The RCP Phase VII compressional seismic volume provides better resolution within the Grayburg-San Andres interval than the Maljamar 3-D survey, and allows for a
more detailed stratigraphic analysis within the study area. Major reflections in the Grayburg/San Andres interval are genetically related to unconformity-induced effects. The Yates, Queen, Grayburg, Upper San Andres, Lower San Andres, and Glorieta horizons all represent to major unconformity-bounded sequences (Borer and Harris, 1991; Wang, et al., 1998a; Kerans, et al., 1994; Sonnenfeld, 1991; Silver and Todd, 1969). Impedance contrasts associated with these sequence boundaries are driven by abrupt facies changes and associated petrophysical variations (Ruppel and Holtz, 1995). The Cycle 1 seismic horizon is related to the impedance change between a tight, anhydrite-filled karst breccia zone and underlying reservoir rock. The karst breccia is related to exposure due to higher-order cyclicity within the Lower San Andres. Other seismic horizons within the Lower San Andres interval are related to higher-order cyclicity as well, although limited vertical seismic resolution and limited deep well control largely obscure the issue. Within the low-impedance-contrast Guadalupian section, seismic response is largely controlled by unconformity-related lithologic changes.
P-wave isochron mapping provides stratigraphic insight into the Guadalupian section. Isochron mapping between the Upper San Andres and Glorieta shows a general trend of time-thickening to the south (Figure 3.7). This trend is probably evidence of backstepping associated with relative sea-level rise during earliest Lower San Andres time,
and subsequent progradation. This assumption is consistent with the observations of Ruppel and Holtz (1995) and others, who place the Leonardian-Guadalupian boundary within the Lower San Andres, just above an interpreted maximum flooding surface. Insufficient deep well control within the study area prevents confirmation of this trend from log data. Isochron maps from the Yates to Queen (Figure 3.8) and Queen to Upper San Andres (Figure 3.9) do not show discernible time thickening in the southern portion of the study area, and these relationships are taken as evidence that, during mid- to upper-Guadalupian time, the margin was basinward of the study area. Figure 3.10 is an isopach map from Grayburg to Upper San Andres; lack of thickening to the south is supportive of isochron trends.
3.1.3 Structural Interpretation
During RCP Phase VI, coherency analysis was undertaken on the regional Maljamar 3D P-wave dataset. These coherency attribute maps suggest the presence of structural deformation within the reservoir interval at Vacuum Field (Talley, 1997; Duranti, 1999). Coherency analysis of the RCP VII P-wave data, performed by Miguel Galarraga (1999), shows similar discontinuity trends within the study area (Figure 3.11). Fault interpretations were made from the coherency discontinuities. Figure 3.12 is a seismic line extracted from the VII P-wave volume, and shows the faults interpreted from the coherency attribute in Figure 3.11. Fault trends interpreted within the San Andres
from coherency analysis are represented laterally by time structure maps from the Lower San Andres interval (Figure 3.13), and by RMS amplitude extractions taken within the Lower San Andres (Figure 3.14). The faults labeled MF1 and MF2 in Figures 3.13 and 3.14 were interpreted from RCP Phase VI seismic analysis and well control, and incorporated into the Phase VII reservoir model (Pranter, 1999). These faults are evident in Phase VII P-wave seismic, along with additional structural features. The southern modeled fault (MF1) is represented by synthetic seismograms computed within the study area. Figure 3.15 shows a comparison of synthetic traces from CVU 196 and CVU 203. The traces from CVU 203, which is located south of the southern fault (MF1 in Figure 3.13) show the development of a peak not evident in synthetic traces from CVU 196, which is located north of the southern fault. This synthetic cross section transects the southern fault in the region of greatest time displacement, and demonstration of displacement is less clear, presumably due to resolution issues, in other portions of the study area. Lower San Andres structure maps constructed from log correlations also support the interpreted fault trends. Figure 3.16 is a depth structure map of the Lower San Andres top, contoured without structural displacement. Interpreted faults correspond to areas of closely spaced contours. Well control shows that displacement across the faults varies laterally, with a maximum well-to-well offset of approximately 15 m (50 feet). This offset range is below the calculated vertical resolution of the surface seismic, but higher frequencies occur in the upper portion of the Lower San Andres. It is also possible
that the faults may be detectable at lower thresholds than indicated by l /4f considerations (Yilmaz, 1987).
In addition to the E-W trending faults discussed above, a N-W trending anomaly is visible from the P-wave surface seismic. This zone is interpreted to represent a fractured zone within the reservoir because displacement, apparent in time analyses, is not evident from depth structure mapping. The trend is displayed on Figures 3.11, 3.13 and 3.14. The orientation of this fractured zone is similar to the trend of an interpreted erosional scarp at the Glorieta interval (Figure 3.17). The Glorieta-San Andres contact is unconformable, with substantial erosional removal of Glorieta-aged sediment indicated from regional studies (Skinner, 1946; Sonnenfeld, 1991). The interpreted fracture zone may result from differential compaction across the erosional scarp.
The interpreted structural features discussed above are seismically represented only over a limited interval of the reservoir at Vacuum Field. Frequency attenuation with depth probably plays a role in the loss of subtle structural representation in the lower portion of the Lower San Andres. Figure 3.18 (top) shows a spectral analysis of the upper reservoir interval (Grayburg to Cycle 1, 70 ms) on a line extracted from the Phase VII P-wave seismic volume. Dominant frequency from this interval is approximately 58 Hz. Over an interval in the lower Lower San Andres, spectral analysis reveals a dominant frequency of approximately 42 Hz (Figure 3.18, bottom). The trend of frequency attenuation with depth is clear, and is the inferred cause for loss of resolution in the lower section. In
addition to resolution-related absence of structural offset in the lower portion of the Lower San Andres, structural patterns visible in the upper portion of the San Andres are,
in general, not visible seismically above the Upper San Andres top. The top of the Upper San Andres is an unconformity, with outcrop studies suggesting erosional removal of up to 100 feet of section (Sonnenfeld, 1991; Sarg and Lehman, 1986; Skinner, 1946). This
erosion could have leveled the Upper San Andres prior to Grayburg deposition, effectively removing evidence of pre-Grayburg structural relief. Figure 3.19 shows an isochron map constructed for the Upper San Andres-Lower San Andres interval. Thickening in the northern and southern portions of the study area is apparent, and suggests preferential removal of Upper San Andres section along the upthrown (central) fault block. The isochron trend is supported by isopach data over the Upper San Andres interval, which clearly show thinning of the Upper San Andres interval across the upthrown block (Figure 3.20). This interpretation suggests decreased vertical fault motion during and after Grayburg deposition.
Interpretation of Phase VII P-wave data indicates the existence of significant structural deformation within the San Andres interval. The interpreted faults parallel the Permian Shelf margin, and coincide with faults interpreted by Talley (1997), who discussed differential compaction along the margin as a possible cause of faulting. Horizontal log analysis from well CVU 110 (south of study area) shows evidence of a parallel, E-W fault trend (Pranter, 1999), and suggests the existence of a fault system
associated with the Permian margin. Evidence of laterally-variable offset across the northern and southern faults indicates the possibility of a wrench fault system within the San Andres interval, with most recent relative motion along faults being primarily strike-slip. Fault patterns interpreted by Galarraga (1999) from the regional Maljamar 3-D dataset also support a wrench fault interpretation. Lateral variations in fault offset may also be related to differential erosional patterns during sea level lowstand, and ensuing karst effects.
P-wave amplitude maps were constructed for several intervals within the Grayburg-San Andres section. RMS amplitudes extracted from a narrow window around the Cycle1 horizon display the structural patterns interpreted within the San Andres (Figure 3.14). These patterns are also expressed on the RMS amplitude map constructed over the Grayburg-Cycle1 interval, which includes most of the reservoir at Vacuum Field (Figure 3.21). The amplitude values from this map were cross-plotted with mean porosities over the interval (Figure 3.22). The crossplot shows considerable scatter, with no distinct relationship between mean porosity and amplitude apparent.
3.1.5 Significance of P-wave Interpretation
Analysis of RCP Phase VII compressional seismic provides significant insight into the structural and stratigraphic character of the Guadalupian section at Vacuum Field.
Although resolution limitations prevent seismic delineation of intervals at the flow unit scale, major stratigraphic interfaces and structural features are discriminated. The establishment of this structural and stratigraphic framework provides a crucial geological base for subsequent multicomponent seismic analyses.
3.2 Shear Wave Interpretation
Shear wave seismic is useful as a characterization tool because of an increased sensitivity, relative to compressional data, to a variety of reservoir parameters (Tatham and McCormack, 1991). This sensitivity, however, also complicates S-wave interpretation considerably. Experience has shown that S-wave surface seismic at Vacuum Field is not suitable as a stand-alone discriminator of reservoir structural or stratigraphic patterns; the benefit of shear data is realized in a multicomponent sense, after calibration with compressional data.
Phase VII shear wave surface seismic was acquired in December, 1997. The acquisition grid for shear data was the same as that used for P-wave acquisition (Figure 3.2). Shear wave data were rotated in processing to S1 (N 118° E) and S2 (N 28° E) orientations, in accordance with maximum horizontal stress estimations from borehole breakout considerations at well WS 2-26 and multicomponent VSP polarization analyses at wells CVU 200 and VGWU 127 (Scuta, 1997; Mattocks, 1997; Reservoir Characterization Project, 1998). Horizontal stress considerations provide a means of predicting dominant open fracture and low aspect ratio pore directions. However, maximum horizontal stress directions may change locally in response to structural and stratigraphic features (Roche, 1997; Bruno and Winterstein, 1993). Thus, in some areas of the reservoir, the processed S1 component does not necessarily correspond to the orientation of the fast shear wave. The analyzed shear wave volumes will be referred to as "S1" and "S2" in the discussion that follows, and the natural shear coordinates will be referred to as the fast and slow shear wave directions.
RCP Phase VII S-1 and S-2 surface seismic volumes were calibrated to surface compressional data through VSP analysis, the CVU 200 shear wave synthetic, and character matching at the Yates Formation (see Chapter 2, this volume). Figure 3.23 is a N-S line extracted from the S-1 volume, with picked horizons annotated. The shear wave seismic shows less continuity within the reservoir than is seen from the P-wave (compare Figures 3.23 and 3.3). The discontinuous character is related to both structural and petrophysical changes within the interval.
3.2.1 Shear Wave Resolution
The dominant frequency of shear wave seismic within the reservoir is approximately 21 Hz (Figure 3.24). Taking a representative velocity of 2800 m/s (8500 ft/s), and using a l /4 approximation, vertical resolution is 30 m (101 ft). Figure 3.25 shows the shear wave synthetic constructed from dipole sonic data at CVU 200, with reservoir flow units displayed. As with P-wave data, even for locally higher resolution within the reservoir, seismic character on the shear wave traces is controlled by contributions from multiple flow units.
Seismic amplitudes from the Phase VII shear wave data were analyzed to asses relationships with geological information. Figure 3.26 is an RMS amplitude map from the S1 volume over the Grayburg-Cycle1 interval. This interval includes most of the reservoir at Vacuum Field. The S1 amplitude map shows high values in the northern and southern portions of the study area. The high amplitude anomalies show a general correspondence with structural trends. S1 amplitudes show considerable scatter when cross-plotted with mean porosities over the Grayburg-Cycle1 interval (Figure 3.27). S2 amplitudes over the Grayburg-Cycle1 interval differ markedly from S1 amplitudes, but show a similar range of values (Figure 3.28). A crossplot of S2 amplitude with mean porosity does not reveal distinct trends over the Grayburg-San Andres interval (Figure 3.29).
3.3 Integrated, Multicomponent Interpretation
A variety of multicomponent techniques have been applied to RCP Phase VII surface seismic volumes. These methods provide information on petrophysical properties not available from traditional characterization methods. From the Phase VII volumes, seismic time measurements, specifically Vp/Vs and seismic anisotropy analyses, give reasonable correlations with geological features and production data. Unfortunately, Vp/Vs and anisotropy measurements from seismic require large intervals for robust measurement. Time-based multicomponent analyses at Vacuum also require measurement between high amplitude events, as weaker horizons are typically discontinuous and subject to picking errors. For anisotropy maps, intervals of approximately 300 ms (S-wave data) were used. Vp/Vs measurements can be computed over smaller intervals (approximately 150 ms, S-wave time, within the reservoir) with reasonable results, as these measurements consider both S-wave picks and P-wave horizons. P-wave velocities are higher, and compressional isochrons are less affected by minor pick variations.
Shear wave amplitude measurements were considered as a means of increasing the vertical resolution of multicomponent measurements. Arestad (1995) found S1 amplitudes to be dominantly controlled by porosity/lithology variations, and S2 amplitudes to be influenced by both porosity/lithology and fracture variations. Amplitude difference maps at Joffre Field indicated possible changes in permeability trends. A similar differencing technique was applied to shear wave amplitudes over the reservoir at Vacuum Field. These analyses did not determine any relationship with porosity or permeability data, and thus are not considered further in this thesis. Shear wave amplitude analysis is an excellent area for further study at Vacuum Field.
3.3.1 Vp/Vs Measurements
Vp/Vs ratio measurements take advantage of P- and S-wave velocity differences (Section 1.3.1, this volume). Vp/Vs measurements can be affected by numerous petrophysical and reservoir properties, but ratio values are influenced most heavily by lithology and porosity (Tatham and McCormack, 1991).
Figure 3.30 shows a Vp/Vs1 map over the Grayburg-Cycle1 interval. This interval spans the Grayburg, Upper San Andres, and a portion of the Lower San Andres, and includes most of the reservoir at Vacuum Field. Mapped values range from 1.4 to 2.4, with generally lower ratio values in the southern portion of the study area. The Vp/Vs2 map over the same interval shows differences interpreted as related to anisotropy, but similar overall trends (Figure 3.31). Cycle1 picks are poor in some areas of the S2 volume, and the map is heavily edited.
The Vp/Vs values plotted in Figures 3.30 and 3.31 cover the range expected for carbonate reservoirs (Section 1.3.1), and are in the same range as values obtained from sonic log Vp/Vs measurement at CVU 200 (Figure 3.32). Velocity ratio ranges from the
Phase VII data are also comparable to results from Phase VI seismic and petrophysical analysis (Roche, 1997; Talley, 1997; Capello de Passalacqua, 1995).
Vp/Vs1 values shown in Figure 3.30 were cross plotted with mean neutron porosity values computed for wells within the study area over the Grayburg-Cycle1 interval (Figure 3.33). This crossplot displays two trends; a trend of increasing Vp/Vs1
with increasing mean porosity (Trend 1), and the counter trend of decreasing Vp/Vs1 with increasing mean porosity (Trend 2). Wells which correspond to Trend 1 are concentrated in the northern portion of the study area, while wells which follow Trend 2 occur primarily in the southern portion of the study area, along and south of the major southern fault (Figure 3.34). Wells with large deviations or unscaled porosity logs have been removed from these analyses.
The relationship between Vp/Vs1 and mean porosity for wells corresponding to Trend 1 is shown in Figure 3.35. The coefficient of correlation for this crossplot is 0.721. An analogous trend was noted by Roche (1997) over a similar interval at Vacuum Field (Queen-Cycle1), and was interpreted to reflect variability in low aspect ratio pore space. Low aspect ratio pores result in decreased rigidity within the sampled medium, which decreases shear wave velocity relative to compressional velocity, and leads to higher Vp/Vs values (Roche, 1997; Wang, 1997). Adams (1997) states, from petrographic analysis, that most pores examined from available cores are irregularly shaped, with
measured aspect ratios concentrated between 0.2 and 0.7. Thus, in general, relationships with low aspect ratio pores are assumed to be indicative of mean porosity trends.
Wells in the study area which correspond to Trend 2 show decreasing Vp/Vs1 values with increasing mean porosity (Figure 3.36). The coefficient of correlation for this crossplot is –0.682. This trend is interpreted to result from increases in volume percent anhydrite over the reservoir interval in the southern portion of the survey area. Roche (1997) offers a similar interpretation at Vacuum Field, and Arestad (1995) notes regional decreases in seismically-derived Vp/Vs in association with increased volume percent anhydrite over the dolomitic Nisku reservoir at Joffre Field, Alberta, Canada.
Mineralogical analysis at Vacuum Field supports this interpretation. r maa-Umaa crossplots were analyzed for 14 wells with requisite log suites in the study area (Pranter, 1999). The r maa-Umaa technique uses photoelectric index, neutron porosity, and bulk density logs to correct for pore fluid influences and provide information on matrix mineralogy (Doveton, 1994). r maa is an apparent matrix density factor, and Umaa is an apparent matrix volumetric photoelectric factor. Figure 3.37 shows the r maa-Umaa crossplot for CVU 196 in the central portion of the study area (top), and the crossplot from CVU 203 in the southern portion of the study area (bottom). A higher percentage of log samples fall above the 20% anhydrite cutoff for CVU 203 than for CVU 196. The lateral distribution of percent log samples above the 20% anhydrite cutoff from r maa-Umaa analysis is shown in Figure 3.38 (Pranter, 1999). Higher percentages of anhydrite
are indicated in the southern portion of the study area, generally along the trend of the southern fault. The distribution of high percent anhydrite corresponds to low Vp/Vs trends over the same reservoir interval in the southeastern portion of the study area (Figures 3.30 and 3.31).
Figure 3.39 shows a crossplot of percent log samples above the 20% anhydrite cutoff from r maa-Umaa analysis against Vp/Vs1 ratio values for wells with the requisite log suite in the study area. The crossplot shows a trend of decreasing Vp/Vs1 value with increasing anhydrite; this relationship is supportive of trends interpreted from Vp/Vs and from r maa-Umaa analysis.
At Joffre Field, increases in anhydrite over the reservoir interval, as indicated by local decreases in Vp/Vs, coincided with porosity occlusion and reservoir degradation (Arestad, 1995). At Vacuum, however, wells in the southern portion of the survey area, which show increased volume percent anhydrite, also have relatively high mean porosities
(Figure 3.33). Increased anhydrite in the southern portion of the study area is concentrated in thin intervals, with relatively anhydrite-free intervals interspersed (Figure 3.40). These less anhydritic intervals drive the high mean porosities apparent in the southern portion of the survey area.
Laboratory measurements suggest higher Vp/Vs values for pure anhydrite than for dolomite (Capello de Passalacqua, 1995; Sarmiento, 1994). Thus, the observed relationship between increasing percent anhydrite and decreasing Vp/Vs ratio values
appears counterintuitive. Arestad (1995) noted this at Joffre Field. However, most if not all of the anhydrite present within the reservoir at Vacuum Field occurs in a void-filling habit; no intervals of bedded anhydrite are apparent within the Grayburg-San Andres interval (Adams, 1997). In low aspect ratio pores and along fracture planes, infilled anhydrite is interpreted to locally change pore shapes and increase rigidity. This increased rigidity would lead to locally higher shear wave velocities, and accordingly decreased Vp/Vs values.
The trend of low Vp/Vs values, interpreted as representing increased volume percent anhydrite over the Grayburg-Cycle1 interval, corresponds closely to the trace of the southern fault zone. This correspondence suggests that increased percent anhydrite may be related to faulting within the reservoir at Vacuum. Additional data is necessary to further formulate geological interpretations of the observed relationships.
An alternative explanation for the trend of decreased Vp/Vs values in the southern portion of the study area could be an increased clastic component within the reservoir near the shelf margin. Laboratory results indicate decreased Vp/Vs ratios for quartz relative to dolomite (Table 1.1). Substantial clastics, however, are not noted in the CVU 100 core or in r maa-Umaa crossplots from wells in the study area.
Vp/Vs1 maps over the Grayburg-Cycle1 interval show a region of low velocity ratio values and low percent anhydrite in the south-west-central portion of the study area (Figure 3.30). The region is characterized by relatively high gas/oil ratio (GOR) values (Figure 3.41), oil production (Figure 3.42), and mean porosities (Figure 3.43). This area was discussed in Phase VI multicomponent analysis by Roche (1997), who interpreted the region to represent Vp/Vs response to locally higher GOR. High percentages of gas within the pore space would increase compressibility, and thus drive down P-wave velocity relative to shear velocity. The low Vp/Vs values could also be controlled by the predominance of high aspect ratio pore structures over the interval. Approximately spherical pores would not slow shear wave velocities to the extent of low aspect ratio voids, and would result in lower Vp/Vs values. The region corresponds to an area of structural complexity, and may represent a compartmentalized zone within the reservoir. Sufficient data is not available to definitively interpret this region of low Vp/Vs.
Vp/Vs relationships correlate well with porosity and anhydrite trends over the Grayburg-Cycle1 interval. The identification of dual Vp/Vs vs. porosity trends strongly suggests that Vp/Vs values are driven by different pore structure effects in different portions of the survey area. Sufficient core data is not available at Vacuum Field to quantify changes in porosity percentages or pore-filling material. However, Vp/Vs trends noted from the Grayburg-San Andres tie well with borehole data. This tie indicates a strong link between velocity ratio measurements and variations in pore structure and lithology. This relationship is not apparent from P-wave seismic measurements, and provides porosity and lithology information for the interwell space which is not available from log analysis.
3.3.2 Shear Wave Anisotropy Measurements
Shear wave anisotropy measurements take advantage of shear wave birefringence, and provide information on aligned discontinuities within the subsurface (Section 1.3.2, this volume). Because S1 and S2 shear wave surface seismic volumes at Vacuum Field were rotated in processing to constant azimuthal directions of N118° E and N28° E (Section 1.3.2), localized variations in horizontal stress can result in negative values of anisotropy. This single rotation complicates shear wave interpretation within the reservoir considerably.
Figure 3.44 is an anisotropy map constructed over the Grayburg-Glorieta interval at Vacuum Field. This interval is approximately 300 ms thick on shear wave surface seismic volumes. The plotted values range from -15% to 15% anisotropy, with most values concentrated between +/- 10% (Figure 3.45). Negative values are indicative of areas where the S1 isochron values are larger than S2 values over the sampled interval, and suggest that the fast shear wave polarization direction locally approaches the processed S2 orientation. The high positive anisotropy anomalies highlighted on Figure 3.44 are interpreted to represent northwesterly-oriented fracture sets within the reservoir.
Figure 3.46 is an anisotropy map computed over the Tansill-Grayburg interval. Time thickness of this interval is approximately 300 ms on the shear wave volumes. The range and magnitudes of anisotropy values over this interval are similar to those seen within the reservoir (Figure 3.45). This observation suggests that anisotropic features
persist above the Grayburg Formation at Vacuum Field. The anisotropy map over the Tansill-Grayburg interval also shows a subtle "mirror effect" when compared with the reservoir anisotropy map in Figure 3.44. Positive anisotropy values from the reservoir interval are, in some cases, represented by negative values from the overlying interval. This effect is interpreted to represent seismic distortion due to anisotropy within the Tansill-Grayburg section.
Figure 3.47 is a shear wave anisotropy map computed over the interval above the reservoir from the Rustler to the Tansill Formations. The time thickness of the interval is approximately 500 ms on the S1 and S2 volumes. The map is plotted at the color scale used in Figures 3.44 and 3.46 for comparison, and shows a more narrow range of anisotropy values than is seen within the reservoir interval. This narrow range of measured anisotropy suggests that this upper section is more weakly anisotropic than the lower intervals analyzed. This trend is geologically reasonable, as the late Permian was a time of relative tectonic quiescence. Additionally, regional studies and the limited shallow log data available at Vacuum Field suggest that the upper Permian section is dominated by evaporitic units, which are less likely to sustain fractures than carbonates.
Anisotropy values over the reservoir interval at Vacuum correspond visually with high total fluid production anomalies in the central portion of the study area (Figure 3.48). Specifically, two high positive anisotropy anomalies (indicated on Figure 3.44) correspond with locally accelerated total fluid production trends. These two anisotropy anomalies are
interpreted to represent fracture sets within the reservoir interval. The western anomaly on Figure 3.44 shows a similar alignment to the interpreted erosional scarp at the Glorieta (Section 3.1.3, Figure 3.17), and fractures could be genetically related to differential compaction effects over this feature. The eastern anomaly indicated on Figure 3.44 occurs between the interpreted northern and southern faults within the study area, and associated fractures could be related to wrenching along the fault planes. Determination of definitive
causes of these anomalies is not feasible given current data, but both occur in locations which make a fracture interpretation reasonable.
The orientations of the anomalies are approximately parallel to measured horizontal stress, and thus associated fractures are likely to be open (Major and Holtz, 1997). Open fractures locally enhance permeability along fracture strike. Zones of increased fracture permeability provide preferential flow paths for EOR injection fluids. Reservoir zones outside the fractured areas present targets for bypassed pay strategies.
The visual link between high positive anisotropy anomalies and total fluid production is in general supported by statistical analysis. Figure 3.49 show a crossplot of total fluid production, over the entire produced interval, and anisotropy for producing wells within the study area. Two wells in the northern portion of the study area (wells a=CVU 88 and b=CVU 87 in Figure 3.49) show high production and relatively low anisotropy, and thus skew the correlation. If these wells are removed from the analysis, a trend of increasing total fluid production with increasing anisotropy is apparent (Figure 3.50).
The coefficient of correlation for this plot is 0.768, but this value is largely driven by two highly productive wells located within regions of high anisotropy. Production from additional wells would assist in verification of this trend.
Wells a (CVU 88) and b (CVU 87) in Figure 3.49 display anomalously high production with very slight anisotropy. There are two possible explanations for this relationship. The high fluid production from these wells may be due to high matrix permeabilities from unfractured rock. This interpretation is unlikely, as neither well shows anomalously high matrix permeability from the neural network estimation (Figure 3.51). Additionally, both wells occur near the interpreted northern fault (MF2). A preferred interpretation is that conjugate fracture sets exist in the vicinity of wells a and b. Conjugate fracture sets would lower seismic anisotropy, as shear wave polarizations would fall between fracture directions (Winterstein, 1992). This interpretation is geologically reasonable for wells a and b. Well a occurs near an interpreted bend in the northern fault. Fractures with variable orientations are expected in areas of fault plane curvature (Talley, 1997). Well b occurs along the trend of the northern fault, just north of a high positive anisotropy anomaly interpreted to represent a northwest oriented fracture trend. The anisotropy anomaly terminates abruptly against the northern fault (Figure 3.44), and this termination could represent the development of a conjugate fracture trend north of the fault.
Conjugate fracture sets have previously been interpreted at Vacuum Field. Scuta (1997) saw evidence for northeasterly and northwesterly trending fractures from FMI analysis at well WS 2-26, along the interpreted southern fault in the southwestern portion of the study area. Talley (1997) discussed conjugate fracture sets as an explanation for a time lapse anomaly south of well CVU 97, also along the southern fault zone. Thus, wells a and b in Figure 3.49 are interpreted to represent zones of multiple fracture sets.
Despite the statistical and visual link between high positive anisotropy and enhanced fracture permeability, localized anomalies are visible in Figure 3.50. Wells in the southeastern portion of the study area which correspond to positive anisotropy trends (wells k=CVU 203 and i=CVU 204 in Figure 3.50) fall within the region of increased anhydrite (see Section 3.3.1). Partial anhydrite filling of fracture planes may locally reduce fracture permeability, and is interpreted as the cause of low fluid production from these wells. Negative anisotropy values also require further consideration. Negative values occur where isochron values from the processed S1 volume are larger than corresponding values from the S2, and are interpreted to represent fracture sets or elongate pores with roughly northeasterly orientations. These features could be related to localized variations in maximum horizontal stress along fault planes. Scuta (1997) observed fracture sets with east-northeast orientations from FMI analysis at well WS 2-26, along the fault zone in the southwestern portion of the study area.
Anisotropy mapping at Vacuum Field is complicated by shear wave data concerns and limited production data. Expected values of anisotropy range from 0%-15%, and these values partially fall within the noise range of the shear wave seismic. The limitations imposed by single rotation angles in the processing of shear wave seismic prevent the discrimination of fracture trends with varying orientations. Shear wave data quality deteriorates to some extent near the edges of the study area as a result of decreased fold and migration issues. Migration velocity functions are currently being analyzed by RCP researchers. In addition, playa lakes at the surface present statics problems in processing, and may affect data quality locally, specifically in the southwestern and northeastern corners of the study area. These near-surface effects are also being examined. The subsurface interval immediately above the Grayburg-San Andres shows considerable anisotropy, and may distort the shear seismic representation of the reservoir. The Cycle1 event, discussed at length in Section 3.3.1, is discontinuous on the S2 volume, preventing the use of this reflector as an isochron boundary for anisotropy analysis. Accordingly, anisotropy measurements were created over the entire Grayburg-Glorieta interval. Production data within the study area is available for the entire Grayburg and Upper San Andres, but spans only the upper portion of the Lower San Andres section. Thus, production data discussed above is computed from a more limited reservoir interval than shear wave anisotropy measurements. Additional production data from a larger number of wells within the survey area would lend increased validity to the relationships observed.
Despite these issues, the anisotropy relationships discussed in this thesis are considered valid. Visual and statistical correlations of anisotropy anomalies and production data along with borehole and core evidence indicate the presence of aligned fractures within the Grayburg-San Andres interval which serve as preferential flow paths for reservoir fluids. Although anisotropy relationships are derived over a larger interval than that used for production computations, interpreted faults and fractures are approximately vertical (Talley, 1997). This suggests that the spatial distribution of fractured regions interpreted from anisotropy mapping should be approximately consistent with the trends indicated by production data. In addition, preliminary analysis of Phase VII time lapse seismic shows similar anisotropy patterns to those identified from the baseline data. Shear wave anisotropy mapping thus potentially provides information regarding lateral fracture permeability trends. Such data is crucial to efficient EOR programs, and is not available from conventional P-wave seismic and well log characterization techniques alone.
Multicomponent interpretation of Phase VII surface seismic provides a great deal of insight into the Grayburg-San Andres interval at Vacuum Field. P-wave data provides a structural and stratigraphic framework for the reservoir. Compressional seismic horizons correspond to unconformities within the reservoir. P-wave coherency, time structure, isochron, and amplitude mapping provides indications of structural trends.
Vp/Vs and shear wave anisotropy analyses provide information not derivable from compressional seismic analysis. Figure 3.52 is a direct comparison of the Vp/Vs1 (Grayburg-Cycle1) and anisotropy (Grayburg-Glorieta) maps discussed in preceding sections. The two maps show different character, and therefore indicate different reservoir properties. Vp/Vs values are sensitive to pore structure, lithologic trends, and fluid effects. Shear wave anisotropy trends are indicative of directions and intensities of
aligned fracture sets and low aspect ratio pores, and thus give information on fracture permeability pathways.
Figure 3.53 is a compilation of important trends and reservoir regions identified within the study area from multicomponent interpretation of Phase VII seismic data. Fault and fracture trends, indicated by P-wave attributes, are represented by solid and dashed black lines. Fault zones locally compartmentalize the reservoir, and, by providing flow paths for reservoir fluids, channel anhydrite emplacement and karst effects. Fault motion may have been instrumental in fracturing the reservoir.
High positive anisotropy anomalies are interpreted zones of open fractures paralleling the dominant direction of maximum horizontal stress (Figure 3.53). These anomalies occur between major fault zones, and correspond to increases in total fluid production from the reservoir. They are interpreted to represent azimuthally high fracture permeabilities in the northwestern direction.
Statistical analysis shows that, in the northern portion of the study area, high mean porosity values over the Grayburg-Cycle1 interval are represented by high Vp/Vs ratios (Figure 3.53).
The trend of increased anhydrite over the Grayburg-Cycle1 interval is interpreted from low Vp/Vs1 values in the southern portion of the study area. This trend is supported by mineralogical analysis from wells within the survey area. The anhydrite is concentrated
within specific reservoir zones. Anhydrite emplacement along fault planes in the southern portion of the study area may locally impede fluid transmissibility.
The area of low Vp/Vs1 values and low anhydrite percent in the southwestern corner of the study area is indicated on Figure 3.53. This portion of the study area may reflect fluid effects, pore structures, or a combination of effects.
The multicomponent relationships derived from RCP data are expected to contribute to development of bypassed pay and EOR patterns. In the northern portion of the study area, bypassed pay considerations should focus on areas represented in white on Figure 3.53. These are regions of lower mean porosity, as interpreted from Vp/Vs analysis over the Grayburg-Cycle1 interval. The areas of anhydrite fill also provide potential storage for bypassed pay. Injectivity should be focused into specific intervals of high anhydrite, as these intervals are interspersed with porous, relatively anhydrite-free
Multicomponent Seismic Attributes
Table 3.1: Summary of reservoir properties and responding multicomponent methods for Vacuum Field.
units. Portions of the reservoir between open fracture zones, as interpreted from shear wave anisotropy analysis, also provide potential for bypassed pay storage. EOR patterns could be established to focus on these areas and avoid enhanced fracture permeability pathways.
Potential contributions to EOR and development efficiency suggest the economic benefits of multicomponent seismic at Vacuum Field. The multicomponent surveys acquired by RCP at Vacuum Field represent a cost increase of approximately 35% relative to identically-sized P-wave surveys (Davis, 1999). Because of increased acquisition and processing costs, incremental benefit studies will be necessary to assess the economic viability of multicomponent techniques at other fields. However, multicomponent methods can substantially improve characterization of heterogeneous reservoirs because they provide a much more sensitive measure of porosity and permeability patterns in the interwell space than compressional seismic (Table 3.1).