VOL. GEOPHYSICS, 59, NO. 11 (NOVEMBER 1994); P. 1650-1665, 22 FIGS.
A 3-D seismic case history evaluating fluvially deposited thin-bed reservoirs in a gas-producing property
Bob A. Hardage* Raymond A. Levey*, Virginia Pendleton, James Simmons*, and Rick Edson*
ABSTRACT We conducted a study at Stratton Field, a large Frio gas-producing property in Kleberg and Nueces Counties in South Texas, to determine how to best integrate geophysics, geology, and reservoir engineering technologies to detect thin-bed compartmented reservoirs in a fluvially deposited reservoir system. This study documents that narrow, meandering, channel-fill reservoirs as thin as 10 ft (3 m) and as narrow as 200 ft (61 m) can be detected with 3-D seismic imaging at depths exceeding 6000 ft (1800 m) if the 3-D data are carefully calibrated using vertical seismic profile (VSP) control. Even though the 3-D seismic images show considerable stratigraphic detail in the interwell INTRODUCTION The Bureau of Economic Geology at The University of Texas at Austin has completed several research studies to better understand the internal architecture of complex, heterogeneous oil and gas reservoir systems (Finley et al., 1992; Levey et al., 1992b). One of these research efforts is the Secondary Gas Recovery (SGR) project, which is an ongoing study to determine if reservoir compartmentalization in older producing properties creates gas accumulations that have either not been contacted, or not been effectively produced, by the perforated intervals in production wells (Sippel and Levey, 1991). The primary focus of this SGR project is to determine how the depositional process and diagenesis, rather than structure, contribute to reservoir compartmentalization. Consequently, the field studies are done in producing intervals that have minimal faulting because faulting introduces a reservoir compartmentalization that overprints and complicates any compartmentalization
spaces and indicate where numerous thin-bed compartment boundaries could exist, the seismic images cannot by themselves specify which stratigraphic features are the flow barriers that create the reservoir compartmentalization. However, when well production histories, reservoir pressure histories, and pressure interference tests are incorporated into the 3-D seismic interpretation, a compartmentalized model of the reservoir system can be constructed that allows improved development drilling and reservoir management to be implemented. This case history illustrates how realistic, thin-bed, compartmented reservoir models result when geologists, engineers, and geophysicists work together to develop a unified model of a stratigraphically complex reservoir system. effects inherited from the depositional system. The studies are also done in older producing properties because such fields usually have enough well-by-well production history and pressure documentation to confirm whether or not reservoir compartment boundaries are present (Jirik, 1990; Kerr, 1990; Kerr and Jirik, 1990; Levey et al., 1992a). This case history summarizes the results of a SGR field study that analyzed how fluvial deposition affects gas reservoir compartmentalization. We performed this study in a portion of Stratton Field in Kleberg and Nueces Counties of South Texas. The stratigraphic interval we studied was the Oligocene Frio Formation, a thick, fluvially deposited sandshale sequence that has been a prolific gas producer in Stratton Field and in several other fields along the FR-4 depositional trend (Figure 1). The regional structure and stratigraphy of the Frio system are well documented by Nanz (1954), Galloway (1977, 1982), Han and Scott (1981), and Galloway et al. (1982) and will not be repeated here.
Manuscript received by the Editor April 18, 1994; revised manuscript received June 15, 1994. *Bureau of Economic Geology, The University of Texas at Austin, Austin, TX. Formerly Bureau of Economic Geology, The University of Texas at Austin, University Station, Box X, Austin, TX; presently, Integrity Geophysics, 1503 Palma Plaza, Austin, TX 78703-3434. 1994 Society of Exploration Geophysicists. All rights reserved. 1650
3-D Seismic Thin-bed Case History
1651
THE STUDY SITE We confined our study to a 7.6-mi2 area (Figure 2) where 3-D seismic data were acquired in this SGR effort. A large number of wells, some as old as 40 years, existed inside this 3-D grid, and the logs recorded in these wells allowed us to make a reasonably thorough geologic analysis of the Frio reservoirs. As shown in Figure 2, we acquired additional data in several wells (the circled dots) to supplement the historic well log, production, and reservoir pressure data bases. These supplemental data consisted of modern well logs, cores, and various pressure tests. Vertical seismic profile (VSP) data were recorded in two closely spaced wells inside the triangle shown near the center of the 3-D grid. 3-D SEISMIC DATA ACQUISITION The 3-D seismic data were recorded across the study area in four overlapping swaths (Figure 3). Each swath consisted of six east-west receiver lines spaced 1320 ft (402 m) apart. The source lines were oriented north-south, spaced 880 ft (268 m) apart, and extended from receiver line 2 to receiver line 5 of each swath. The recording began at the south end of the study area (swath 1, Figure 3a) and rolled northward. When all of the vibrating points (VPs) shown in swath 1 were recorded, swath 2 was created by dropping the southern three receiver lines and adding three receiver lines on the north side of swath 1. This modification allowed the VPs to continue northward as continuous source lines (Figure 3b). In the southern part of the grid (swaths 1 and 2), the source lines were straight and uniformly spaced because bulldozers could be used to clear lanes through the mesquite-covered property. In the northern portion (swaths 3 and 4), permitting restrictions prohibited the use of bulldozers, and the source lines followed irregular paths along existing roads and lanes (Figures 3c and 3d).
We stationed receiver groups at intervals of 110 ft (34 m). Each array consisted of 12 inline geophones spanning a distance of 110 ft (34 m) centered on the receiver flag. Source flags were positioned at intervals of 220 ft (67 m). At each VP, eight linear sweeps (l0-120 Hz) were generated and summed using a 4-vibrator array symmetrically positioned relative to the source flag. This geometry created a grid of 110 ft x 55 ft (34 m x 17 m) stacking bins in which a stacking fold of 20 existed over most of the image area. Before migration, a trace interpolation was done in the source line direction to reduce the bin size to 55 ft x 55 ft (17 m x 17 m). The final processed data had an effective bandwidth of 10 to 80 Hz in most of the Frio interval.
FIG. 1. Map of the prolific Frio FR-4 gas trend in south Texas showing the location of Stratton Field.
FIG. 2. Generalized map of the study area. The solid dots show existing production wells, many drilled at 40-acre spacings. The circled dots locate wells where additional geologic and engineering control was acquired in the form of modern logs, cores, or 2 ressure tests. The 3-D seismic area is approximately 7.6-mi . VSP control data were recorded in two closely spaced wells inside the triangle near the center of the 3-D seismic grid.
1652
Hardage et al.
FIG. 3. The 3-D seismic data were recorded in four overlapping receiver apertures, referred to as swaths 1, 2, 3, and 4. Each aperture consisted of six east-west receiver lines spaced 1320 ft (402 m) apart. In each swath, north-south source lines spaced 880 ft (268 m) apart, extended from receiver line 2 (circled) at the south to receiver line 5 (circled) at the north. Receiver groups were spaced 110 ft (33 m) apart, and source points were spaced at intervals of 220 ft (67 m). Data recording began at the south end of the area using swath 1 (panel a) and ended at the north end with swath 4 (panel d). To convert from one swath to the next, the three southernmost receiver lines of the recording aperture were dropped, and three receiver lines were added to the north side. The solid dots show only key wells, not all wells.
3-D Seismic Thin-bed Case History
1653
THIN-BED INTERPRETATION PROCEDURE The emphasis in this case study was to demonstrate how geologic and engineering data are essential in interpreting depositionally generated reservoir compartment boundaries in 3-D seismic images. The seismic interpretation at Stratton Field was particularly challenging because most of the Frio reservoirs were thin [< 15 ft (5 m)], and they were closely stacked, in some areas separated only 10 ft-15 ft (3 m-5 m) vertically. These conditions required precise calibration of stratigraphic depth-versus-seismic traveltime to extract a depositional strata1 surface from the 3-D data volume that would reliably depict the area1 distribution of a particular Frio thin-bed reservoir. Thin-bed depth-versus-time calibration We used VSP as the primary measurement to define where a specific thin-bed reservoir was positioned in the 3-D seismic data volume. The locations of the two VSP calibration wells we used are shown in Figure 2. The zero-offset VSP data recorded in one of these wells were used to establish the precise depth-versus-time control needed for the thin-bed interpretation. These VSP data are shown in Figure 4, where the zero-offset image is spliced into a north-south vertical slice from the 3-D data volume passing through the VSP well. Also shown in the figure is a graphic representation of the stratigraphic column penetrated by the VSP well. Only producing or potentially producing Frio reservoirs are shown in this diagram, and not all of the reservoirs are labeled by name. The top and base of each reservoir are accurately positioned in terms of two-way VSP traveltime, and since there is no difference in the VSP and 3-D time datum in this instance, the reservoirs are also correctly positioned vertically inside the 3-D seismic data volume at the VSP well. Using these VSP traveltime control data, we knew exactly where each thin-bed reservoir belonged in the 3-D seismic reflection waveform at the VSP well. We then extended this thin-bed calibration away from the VSP well and across the entire 7.6mi2 area imaged by the 3-D data. Defining chronostratigraphic depositional surfaces The fundamental assumption we made in our seismic interpretation was that seismic reflections follow chronostratigraphic depositional surfaces (Vail and Mitchum, 1977). This assumption means that if we map a continuous seismic reflection event over the entire 7.6-mi2 area imaged by the 3-D seismic data, we define a geologic surface that corresponds to a fixed, constant depositional time. In other words, we define a depositional strata1 surface. We were able to find two such areally continuous reflection events in the Frio interval. These two surfaces are shown on the east-west vertical section crossing the VSP well in Figure 5. At the VSP control well, the apex of the peak associated with the shallower strata1 surface (the orange surface in Figure 5) corresponded to the thick C38 reservoir (Figure 4), and the apex of the peak at the deeper strata1 surface (the green surface in Figure 5) correlated with the F11 reservoir. Thus we assumed that the seismic time surface following the apexes of all of the peaks of the orange event defined the
ancient topographic Frio surface at the time when the C38 reservoir sediments were deposited. Likewise, we assumed that the seismic time surface following the apexes of the peaks of the deeper green event defined the ancient depositional surface associated with the F11 reservoir. Once the 3-D data volume was flattened relative to one of these two reference strata1 surfaces, we further assumed that any horizontal time slice in this flattened data volume also followed an ancient Frio depositional surface, as long as the seismic reflection character in the immediate neighborhood of the time slice was time-conformable with the reflection character in the immediate vicinity of the reference surface used to flatten the data volume. In our opinion, the entire Frio section inside the 7.6.mi2 3-D grid was seismically conformable to one of the two seismic reference surfaces we created. The thin-bed interpretational philosophy we followed can be summarized as follows: 1) Choose one or more areally continuous Frio seismic reflections that can be used as reference surfaces to flatten the 3-D seismic data volume. Specifically, we picked two such reflection events: the green and orange surfaces shown in Figure 5. 2) Use VSP data to define the stratigraphic depth corresponding to the peak (or trough) time of each selected reference strata1 surface. We found that the shallow (orange) reflection peak time correlated with the C38 reservoir at our VSP control well, and the deeper (green) reflection peak time correlated with the F11 reservoir. 3) Flatten the 3-D seismic data volume relative to one of the reference strata1 surfaces. In the immediate neighborhood of the reference surface, this procedure restores the flat, no-dip, surface topography that likely existed at the time the meandering, fluvial channel environment deposited the Frio sediments associated with that seismic reference reflection. 4) Make horizontal time slices through this flattened seismic data volume only in time windows in which the seismic reflections are conformable to the seismic reflection associated with the flattened reference surface. These time slices are then assumed to follow individual, fixed, depositional surfaces because the conformable reference reflection event was assumed to be a strata1 surface (Vail and Mitchum, 1977). 5) Use the same VSP control that defined where the reference seismic surface was positioned in stratigraphic depth to define where each time slice is positioned in stratigraphic depth. If two thin-bed reservoirs and relative to the reference occur at times strata1 surface at the VSP control well, then the depositional surfaces for these two reservoirs are horizontal and in the flattened time slices at times seismic data volume. Following this procedure, we found that in this specific interpretation problem, where we had many closely spaced (vertically) thin-beds, the VSP-defined position of a particular thin-bed reservoir was rarely at the apex of a reflection peak or trough. Invariably, each thin-bed of interest was
1654
et al.
FIG. 5. An east-west vertical slice from the migrated 3-D data crossing the VSP control well, which is at crossline coordinate 118. The green and orange surfaces follow the two continuous Frio reflection events that were used as reference strata1 surfaces for flattening the 3-D data volume. The yellow surface is the deepest Frio reservoir level. Immediately below this yellow surface is the severely faulted Vicksburg section.
3-D Seismic Thin-bed Case History
1655
positioned at some intermediate, often nondescript, phase point in the reflection waveform at the VSP control well. To create a seismic image that emphasized the internal complex architecture of a given thin-bed reservoir system, we time-shifted the migrated 3-D data volume first so the proper predefined reference strata1 surface was flat. Then we made a horizontal time slice through this flattened data volume at the exact VSP-defined time for the targeted thin-bed, regardless of where that time slice was positioned in the reflection waveform at the VSP control well. We then assumed that the seismic time surface contained in this horizontal slice was the fixed depositional strata1 surface
where that thin-bed unit was deposited. We also assumed that any seismic anomalies seen on this surface would be related directly to stratigraphic heterogeneities within the targeted thin-bed, and to a lesser degree, would be related to stratigraphic variations in thin-beds positioned immediately above and below the target thin-bed. We now show the results of this thin-bed interpretational procedure at Stratton Field and support the interpretations with geologic and engineering control. F39 RESERVOIR The F39 reservoir was the deepest Frio reservoir we studied. The depositional surface for the F39 reservoir (defined by the steps described above) is shown by the yellow horizon in Figure 5. This surface is immediately above the severely faulted Vicksburg section, and we anticipated some faulting effects might extend into the lowest Frio and create some nondepositional reservoir compartmentalization at the F39 level. The reflection amplitude behavior on our interpreted F39 depositional surface is shown in Figure 6. We assumed the red, linear north-south trends near the center of the image to be residual effects from the deeper Vicksburg faults. A magnified view of this F39 surface in the vicinity of four key wells is displayed in Figure 7. These wells were critical to our study because we were able to acquire F39 reservoir pressure measurements in all four wells at the same time (Figure 8), and the differences in these static pressures told us that each well was in a different F39 compartment. We then examined the 3-D seismic image and the available geologic control to see if these data indicated where the boundaries were that segregated the F39 reservoir into these distinct compartments. Figure 9 displays our interpretation of the available geologic control. Our interpretation of the log curves inferred that the F39 reservoir in each well was deposited in a channel environment that showed some evidence of splay deposition. These log data, by themselves, do not provide
FIG. 6. Seismic reflection amplitude behavior on a strata1 time surface that passes through the F39 reservoir at the VSP calibration well.
FIG. 7. Magnified view of the reflection amplitude behavior on the F39 strata1 time surface in the vicinity of four critical information wells.
1656
Hardage et al.
much information about where compartment boundaries may exist. However, the seismic image does indicate some possible compartment boundaries. For example, the most likely cause of the compartment boundary that separates well 197 from the other wells is the depositional variation that created the red/blue (positive/negative) amplitude changes that trend north-south between crossline coordinates 130 and 140 (Figure 7). Similarly, a probable seismic indication of the compartment boundary that segregates well 75 from the other wells is the positive-to-negative (red-toblue) amplitude variations trending north-south between crossline coordinates 110 and 120. However, this same logic of looking for inter-well reflection amplitude changes does
FIG. 8. Static pressures in the F39 reservoir measured during January 1992 in the four wells shown in Figure 7. The bottom-hole pressure in well 75 was 300 psi at abandonment.
not explain why there is a compartment boundary between wells 175 and 202, which are only 200 ft (61 m) apart, since there is no appreciable change in the reflection phase between these two wells (i.e., there is no color change in the reflection amplitude plot between wells 175 and 202 in Figure 7). We relied on offset VSP imaging to verify that there was a significant change in the F39 reservoir seismic reflection character between wells 175 and 202. One of these VSP images is shown in Figure 10. The classical VSP-CDP image (right side of Figure 10) revealed a subtle change in the F39 reflection waveform in the vicinity of the 175 well, but the most definitive indication of a compartment boundary was found by examining the individual reflection traces before they were binned. These traces and the VSP-CDP stacking bin overlay are shown on the left side of Figure 10. When the F39 reflection peaks are followed across the binning corridors toward well 175 (which is located in the eighth stacking corridor from the left edge), we find the peaks terminate in the sixth corridor, at least 50 ft (15 m) short of well 175, and they do not resume in a robust fashion until corridor 12, some 100 ft (30 m) beyond well 175. Thus, we believe the VSP data provide seismic evidence of a compartment boundary between wells 175 and 202 and forces us to reconsider why this boundary is not evident in the 3-D image (Figure 7). We propose the following explanation. The well positions in Figure 7 show only where the wellheads are located, not where the wellbores penetrated the F39 reservoir. We did not run deviation surveys to determine the true bottom-hole locations of the wells. All available information simply indicated the holes were vertical, and we now believe the assumption of true vertical wells can lead to interpretational difficulties when the depositional stratigraphy varies as rapidly in the lateral direction as it does in this example. For example, the F39 reservoir in these wells is at a depth of approximately 6700 ft (2042 m), and at
FIG. 9. Stratigraphic cross-section of the F39 reservoir showing the depositional environments interpreted from log shapes and the initial bottom-hole pressure (BHP) observed in each well. The date when each well was drilled is shown above the BHP value. The log depths are measured from KB and the curves are shifted to align on a stratigraphic datum.
3-D Seismic Thin-bed Case History
1657
this depth, a 1 degree deviation from vertical a is horizontal movement of approximately 120 ft (37 m). If a wellbore is within 3 or 4 degrees of vertical, most people define the well to be vertical. An inspection of Figure 7 shows, however, that the borehole for either well 175 or 202 has to deviate eastward only 100 ft (30 m) or so to move that wellbore from a red zone (positive reflection amplitude) to a blue zone (negative reflection amplitude). Thus, the 3-D seismic image may be revealing the compartment boundary between wells 175 and 202 if we knew exactly the inline/crossline seismic coordinates where the wellbores intersected the F39 reservoir . Our conclusion from analyzing the seismic, geologic, and engineering data associated with the F39 reservoir is that we can seismically detect F39 reservoir compartments, at least in the vicinity of wells 75, 175, 197, and 202, but we must interpret the seismic image with the assistance of reservoir pressure data to infer which of the many stratigraphic changes revealed in the seismic image are most likely to be the compartment boundaries.
F37 RESERVOIR The F37 reservoir was approximately 20 ft (6 m) above the F39 reservoir in our VSP calibration well, which is a two-way traveltime difference of only 4 ms. Using the previously described thin-bed interpretation procedure, we created a time slice through our flattened 3-D data volume 4 ms above the F39 strata1 surface. This F37 surface is displayed in Figure 11. Comparing this image with the F39 surface (Figure 6), we saw red, linear north-south channels in the central part of the F37 image similar to those observed in the F39 image, implying that Vicksburg faulting was still controlling sedimentation in this part of the field. However, there was a significant difference in the southeast quadrant of the F39 and F37 images. Specifically, meander channel features occurred at the F37 level but were not present at the deeper F39 surface. We focused considerable attention on this F37 depositional topography and show an enlarged plot of the meander features in Figure 12.
FIG. 10. Offset VSP imaging of the interwell space between wells 175 and 202. The F39 reservoir is imaged by the low-amplitude black peaks immediately above the dashed line labeled F39. The VSP geometry allowed stacking bins only 25 ft (8 m) wide to be used, so well 175 is positioned at the ninth trace in the VSP-CDP image, which is 200 ft (61 m) away from well 202, the receiver well. In the VSP-CDP image, the F39 peaks exhibit some change in reflection character near the 175 well, but a more definitive indication of an interwell stratigraphic discontinuity in the F39 reservoir is provided by the display on the left, which shows each individual receiver trace before the traces are summed to create the VSP-CDP image. The VSP-CDP stacking corridors are defined by the superimposed grid of lines sloping up to the right. In this prestack display, the F39 reflection peak disappears in stacking corridor 6, about 50 ft (15 m) short of well 175, implying an interwell stratigraphic break of some type.
1658
Hardage et al.
FIG. 11. Seismic reflection amplitude behavior on a strata1 time surface that passes through the F37 reservoir at the VSP calibration well. This surface is conformable with, and only 4 ms above, the F39 strata1 surface (Figure 6).
We made a log-based stratigraphic cross-section of the F37 reservoir across the meander features and continuing southward beyond the seismic grid (Figure 13). The depositional environment (either channel or splay) at each well is an interpretation based on log curve shape and was made before the 3-D seismic data were recorded. This initial geologically based interpretation of the F37 depositional environments indicates that the meander feature seen in the F37 seismic surface is indeed a depositional channel. Specifically, the log interpretation (Figure 13) implies the F37 reservoirs found in wells 189 and 185 were deposited as channel fill, and the seismic image shows these wells to be directly atop a meander feature. The initial depositional interpretation for the extremely thin F37 reservoir in well 211 was that this wellbore could have penetrated either splay or channel fill. The splay option is indicated in Figure 13. Because the 211 wellhead is approximately 300 ft (91 m) north of the meander feature (Figure 12), the logbased interpretation of the F37 depositional environment at the 211 well is also supported by seismic evidence, since the bottomhole location could be either in channel fill or in a splay. We analyzed pressure histories recorded in several F37 reservoirs near these seismic meander features to determine if reservoir compartmentalization existed. These pressure histories, summarized in Figure 14, show there are at least three, and perhaps four, individual F37 reservoir compartments in this area of the field. We relied heavily on these pressure data to guide the interpretation of the thin-bedded F37 reservoirs. A reservoir model that honors all three data bases-the seismic, the geological, and the reservoir engineering-is proposed in Figure 15. This model assumes that the F37
FIG. 12. Magnified view of the reflection amplitude behavior on the F37 strata1 time surface in the vicinity of four critical information wells. An interpretation of the stacked thin-bed channel features revealed in this image follows as Figure 15.
3-D Seismic Thin-bed Case History
1659
reservoir in the southeast quadrant of the 3-D grid is composed of three intermeshed channels, labeled A, B, and C, and a grid overlay of seismic inline and crossline coordinates is provided so these channels can be correlated with features in the 3-D seismic image. The location of the F37 stratigraphic cross-section (Figure 13) is shown, but this geologic information defines channel locations along only a single 2-D profile of the model. The important information again is the reservoir pressure data, because without this engineering data there would be no reason to conclude that a 3-channel model would be appropriate. Thus, the model places well 129 in channel A and well 185 in channel B, which allows these two wells to be in different F37 pressure regimes; i.e., in different compartments (channels). Wells 127 and 161 are proposed to be in channel C, south of the 3-D seismic coverage. Only one meander loop of this hypothesized channel C extends into the 3-D seismic grid. The rapid F37 pressure decay observed in well 189 (Figure 14) implies that this well is not in pressure communication with well 185, even though both wells are in channel B. For this reason, we show a stratigraphic variation in channel B where there may be an intrachannel compartment boundary. We wish to emphasize that the reservoir model in Figure 15 is hypothetical and may not yet be the correct picture of the compartmentalized nature of these F37 reservoirs. However, we do know that the F37 reservoir in this portion of Stratton Field is segregated into distinct compartments, that this compartmentalization must be caused by the
FIG. 14. F37 reservoir pressure histories observed in wells near the seismically imaged meander features. These pressure decline curves indicate that these wells are positioned in at least three different F37 compartments, labeled as channels A, B, and C (same labeling notation used in Figures 13 and 15). There may be an additional, intrachannel, F37 compartment boundary in channel system B that segregates well 189 (the bold Xs) from well 185 (the open circles).
FIG. 13. Stratigraphic cross-section of the F37 reservoir showing the depositional environments interpreted from log shapes and the initial bottom-hole pressure (BHP) observed in each well. All log curves are depth-shifted to a marker datum. The channels labeled A, B, and C refer to meander features shown in map view in Figure 15. The northern three wells are inside the 3-D seismic grid; the southern three wells are not.
1660
Hardage et al.
fluvial deposition because the seismic data show no evidence of faulting, and that the proposed reservoir model honors all existing data that provide any information about the F37 reservoir system. We concluded that the seismic image in Figure 12 revealed not just one meander channel system but at least three intermeshed thin-bed channels, and that again by using common existing reservoir engineering data (i.e., pressure histories), we were able to use 3-D seismic images to define where compartment boundaries most likely existed in the interwell spaces.
D11 RESERVOIR Our third example of a compartmented fluvial reservoir iS the D00 system, which is one of the very thin reservoir immediately above the D35 reservoir shown in Figure 4 Again, using our VSP calibration technique to define the appropriate flattened time slice that equated to thE D11 stratigraphic level, we generated the reflection image ir Figure 16. The information in Figure 17 summarizes the D11 geologic and engineering data available in this portion of the 3-C
FIG. 15. Proposed model for the F37 reservoir system imaged in Figure 12. This model honors the stratigraphic cross-section shown in Figure 13 and the pressure history shown in Figure 14. The channels are arbitrarily drawn as A being the deepest and C the shallowest. An intrachannel barrier is proposed near crossline, inline coordinates (125, 35) to explain the different pressure behaviors in wells 185 and 189 (see Figure 14). South of the 3-D seismic coverage area, the boundaries of channels A, B, and C are drawn as dashed lines to show that the channel shapes and positions are highly speculative.
3-D Seismic Thin-bed Case History
1661
FIG. 16. Seismic reflection amplitude behavior on a strata1 time surface that passes through the D11 reservoir at the VSP calibration well.
FIG. 17. Stratigraphic cross-section of D11 reservoirs showing the depositional environments interpreted from log shapes and the initial D11 reservoir pressures observed in two critical wells (77 and 150). These pressure data show that well 150, drilled in 1983, penetrated a D11 compartment with virgin reservoir pressure. This compartment evidently has no effective communication with the D11 compartment where well 77 was drilled because well 77 was abandoned in 1972 because of depleted pressure. Consequently, an intermeshed two-channel reservoir system is proposed for the D11 reservoir level in this portion of Stratton Field.
1662
Hardage et al.
FIG. 18. Proposed model for the D11 reservoir system imaged in Figure 16. This model honors the geologic and engineering data summarized in Figure 17 and mimics the meander pattern shown in Figure 16.
seismic grid. The D11 reservoirs in these four analysis wells were all interpreted from log shape analysis to be channel fill; some log character indicated splay deposition in the younger channel system. The pressure information was again the key information that told us how many reservoir compartments should be seismically imaged. As noted in Figure 17, the oldest production well (77) was drilled in 1956 and encountered a reservoir pressure of 1740 psi. This well was abandoned in 1972 when the pressure declined to uneconomic levels. Wells 135, 150, and 156 were drilled in the 1980s, and all three wells encountered virgin D11 reservoir pressure, implying these wells were in a single D11 compartment that was depositionally segregated from well 77. We thus interpreted our D11 seismic image with the knowledge that we were dealing with at least two intermeshed or juxtaposed thin-bed channel systems. The reservoir model we inferred from Figure 16 is shown in Figure 18. This model honored the geologic information by placing each of the four wells in a seismically defined D11 feature that appeared to be a channel and honored the pressure data by placing wells 135, 150, and 156 in the same channel complex and well 77 in a separate channel system. We concluded again that fluvial deposition can create compartmented reservoirs without the stratigraphic section being faulted, that individual reservoir compartments can be seismically imaged when geologic and reservoir engineering data are used to guide the seismic interpretation process, and that detailed VSP data or their equivalent should be used to precisely calibrate the location of thin-bedded reservoirs in the seismic reflection response. This VSP calibration is critical because when many thin-beds are stacked close together, a particular thin-bed seismic response can be
FIG. 19. Stratigraphic cross section of the F21 reservoirs showing the depositional environments interpreted from log shape analysis and the initial reservoir pressures (BHP).
3-D Seismic Thin-bed Case History
1663
completely missed if the time slice is mispositioned by only 3 or 4 ms. F21 RESERVOIRS For our fourth and last example of seismic thi...