|
 DvC2-05
Figure DvC2-05 shows three EI-27 EPL #1 pre-drill maps, over approximately the same area of Eugene Island Block 27 (entire), and EI-46 (northwestern portion, below EI-27). These maps graphically compare:
(LEFT SIDE) - A conventionally processed and (Landmark Graphics workstation) interpreted 450-acre "amplitude anomaly" prospect map, in EI-27, with a gas-productive, 244-acre "analog" amplitude anomaly less than 3 miles to the south, in EI-46. This analog production was interpreted to belong to the same geological layer (named after the CIBECIDES CARSTENSI micro-fossil it contains).
(RIGHT SIDE) - Two D3DSP map-view renderings from VoxelGeo (VG). The upper map shows EPL's 2002 "exploratory well target" (here 630 acres) Common-Impedance Object, or CIO (in red and yellow), representing the same CIB CARST sand target shown on the left-hand conventional map (upper anomaly). Likewise, the lower map shows the same (26+ billion cubic feet gas and equivalent condensate [bcfge, cumulative production] CIB CARST gas-reservoir) "analog" CIO, as the lower anomaly on the left-hand conventional prospect map, but now covering 923 acres.
For the EI-27 and EI-46 wells (EPL #1 and Norcen #2, respectively), the TDs and logged depths were converted to seismic reflection two-way-travel-time (TWT) using time-depth relationships from velocity surveys shot in the two light-blue-dot-annotated wells (in Figure DvC2-04) in EI-24 and EI-26, respectively. The blue "cross-hairs" through each of the EPL #1 and Norcen #2 wells, are EW-inline and NS-crossline tracks, marking the two separate VG Starting VOXEL "detection seed points" that were planted and allowed to grow to the interpreted maximum reservoir sizes, shown. The two seed points were chosen along the EPL #1 proposed (and drilled in 2002) well path, and the Norcen #2 drilled (1996) well path, seen on the upper and lower D3DSP maps, respectively. It is significant that the Norcen #2 discovery well (EI-46) was drilled and produced after Fairfield Industries had shot these multi-client ("spec", radio-telemetry recording, not streamer-cable) 3D data, in 1995. Therefore, what VoxelGeo sees in the D3D-impedance data, is truly (D3DSP-interpreted to be) the gas sand that Norcen discovered ... un-depleted at the time the seismic data were recorded. Note, also, the difference between the sizes of the smaller (244-acres) lower-left, conventional amplitude anomaly (produced by using layer-earth-assumption processing and mapping), and the larger (923-acres), lower-right D3D-CIO. The lower-right map is a top-down view of the correlated (by logs and paleo) CIB CARST sand CIO, as identified by an excellent synthetic tie to the sonic-and-density-logged Norcen #2 well perforated depth, and this D3DSP-recognized increase in the size (area, volume, and seismically estimated reserves) of the analog field "anomaly", probably helped reduce the EI-27 prospect risk, in some manager's minds. More of the D3DSP's EI-27 story is told in the captioned discussions, for the other Figures (DvC2-06 through 12), that follow.
As for this Figure, the "amplitude map" on the left side is the conventional (composite of an EI-27 map and an EI-46 map), time-structural-contours-on-seismic-reflectivity-amplitudes prospect map. The white color, within both of the pink-surrounded-by-green features, represents the "highest amplitude" (loudest echo) areas, and the TWT-structural contours tend to wrap around the down-dip (easterly) portion of the upper anomaly. This was a conventional amplitude map indicator of a possible, flat-lying, oil-water or gas-water contact. So, even though there was no clear conventional evidence that either high-amplitude anomaly was a gas sand (such as the upper, yellow meandering channel shape, and some edge-contact with known gas pay wells, seen on the right-hand D3DSP maps), the proximity to a 20+ bcfge CIB CARST gas well and the EI-27 amplitude anomaly's down-dip limit conforming to structure, were enough to bid on the lease and (once having won it) start the search for a drilling partner. An early sign of some of the risks involved with this prospect, was the fact that the conventional prospect map was created by interpreting the TWT to two separate seismic reflection peak events (one at the northern anomaly and another one at the southern Norcen well, but nearly the same), both considered to represent the seismic wavelet interference pattern near the top of the geologists' CIB CARST sand layer. These two, nearly continuous, mappable seismic events were interpreted to contain both the northern and southern CIB CARST sand amplitude anomalies, so the final prospect map (left-hand CIB CARST sand amplitude map) was created as a composite (virtually splicing together) of the two maps, because it was impossible to map both anomalies on the same seismic (horizon) surface, without "jumping a leg".
{Once upon a time, seismic interpreters worked with colored pencils on paper records containing wiggle- or variable-area-traces, with the recorded traces placed on the desk so that the positive amplitude "peaks" (often with the variable area under the wiggle-trace curves, colored solid black) pointed upward, and the negative amplitude "troughs" pointed downward. The downward pointing troughs looked like "legs", and structural interpretation of a seismic section consisted of manually drawing a colored lined connecting the side-by-side (often overlapping) peak or trough layers (also called horizons or events). Once the event had been interpreted ("picked") on all of the 2-D lines in the survey grid, the TWT values were measured with a ruler, under selected Shot Point locations, then tabulated, "corrected" (for line-to-line 2D-line processing misties), posted by hand on a Shot Point map, and contoured to show the TWT "topography" of this assumed subsurface layer. High areas (anticlines) were more prospective because they made good petroleum "bubble" traps, and low areas (synclines) were less prospective because they made poor trapping areas for lighter-than-water petroleum. If, when connecting troughs on adjacent traces, the interpreter drew the colored pencil across a positive peak event, in order to connect two negative trough events (such as between two separate, perhaps "shingled" events), he or she was committing a serious interpretation error. If it was noticed, he or she would be accused of "jumping a leg" and (intentionally or not) creating a false time-dip picture of the subsurface. It was (and is still) important to correctly interpret and map the dip, because it was clear that oil and gas did not normally "float" down-dip. An inexperienced or unethical interpreter could create a dip-closed "trap" on the prospect map ... either accidentally or purposefully ... in order to raise the money to get a wildcat well drilled.}
On the left-hand map in this Figure DvC2-05, is a squiggled line. It is drawn just above the southern amplitude anomaly (along the northern block line of EI-46), and marks the area of an intentional "leg-jump", which was clearly annotated on this map as a (common and useful) "change in mapped horizon". The thin, white ribbons in the upper-left and lower-middle of the map are fault traces, indicating the "heave zone" (lateral separation) where the otherwise continuously overlapping peaks seemed to be "broken" by a fault that offsets the mapped event interface, both vertically and horizontally. The contoured CIB CARST sand structural contours (overlaying the colored amplitudes) show these two ribbon-like fault zone polygons, as areas where the mapped horizon (called the "Top/CIB CARST sand", as correlated on well logs) would be found to be missing, or "faulted out", in any well penetrating the CIB CARST sand in the white fault-heave zone. In fact, the entire CIB CARST sand (not just the depositional top) might be absent in such a (poorly placed CIB CARST sand test) well. The 10 milliseconds (ms) TWT contours (approximately 40-feet), show that the regional time-dip is to the east-southeast, in this area.
The northwestern-most (upper-left) key well is the Odeco #4, drilled and logged as a deep test by Texaco, using the Offshore Drilling Company (ODECO) as their drilling contractor. Odeco took over as operator when Texaco's deep objectives were all wet sands. This key well is marked by the uppermost red circle, and it logged a thin gas zone in a CIB CARST sand interval, less than 15-feet thick, but apparently full-to-base (no water leg, as shown on the log in Figure DvC2-08). But the 15-feet of only slightly resistive sand, found in this well, created a substantial drilling risk for the EI-27 prospect. Using this closest "known CIB CARST sand thickness, the prospective 450-acre conventional-data Bright Spot represented roughly 6,750 acre-feet of reservoir, which equated to about 8.1 bcfge, recoverable (when multiplied by the engineer-accepted 1,200 mcfge per acre- foot, Recovery Factor for this sand in this area). It was an unfortunate coincidence (it is now known) that this up-time-dip, watered-out gas well, had already produced 8.2 bcfge, before it was plugged (from 1977 to 1984). This Odeco #4 CIB CRST sand cumulative production value, was close enough to the estimated 8.1 bcfge reserves for the conventional amplitude anomaly, that the anomaly was easily mistaken for a "footprint" of a water-swept, depleted-gas-sand reservoir. The implication being that the was a good chance that additional drilling would find only low-velocity, residual gas ("fizz-water") in this depleted-gas-sand.
It seemed puzzling (and it made the prospect more risky) that the down-time-dip amplitude anomaly in EI-27 (if it was a gas reservoir) had not been drained by the Odeco #4 well, during its (1977-1984) aquifer-driven productive life. It certainly should have been drained, if the perforated CIB CARST sandstone interval was accurately portrayed by the conventional layered-earth-model.
{Unless, of course, (and it is still possible that) the Odeco #4 production-curtailing water came from a break-down in the up-hole casing cement. One of the main purposes of this cement was to keep any nearby normally pressured, water-bearing sands from forcing brine back into the production perforations, if and when the pressure was drawn down by the gas production. Records have not been located that would firmly establish the source of the final water-cut.}
And, if the water that caused the Odeco #4 well to be abandoned in 1984, did not come from the CIB CARST sand, then it was another risky puzzle as to why the amplitude anomaly failed to reach west to the Odeco #4 location on the conventional amplitude map? The bottom-hole pressure in the EPL #1 discovery well was found to be the same as the reported abandonment pressure in the Odeco #4 assumed-water-out sand, which strongly suggested (after the EPL well was drilled) that there was at least some communication between the "new" EPL #1 gas reservoir and the old 1984 well's watered-out reservoir. In any case, the technical story that is being introduced by this Figure DvC2-05, and further told in the following DvC2 figures, is worth reviewing. These Figures provide good examples of how the D3DSP might be used to help get a "subtle trap" prospect drilled, even if it was not used to "generate and lease" the prospect ... this time.
The D3DSP images on the right side of Figure DvC2-05 show part of the "prospect assistance" offered by the D3DSP. Screen captures of two map-view renderings of the D3D-impedance Time Cube (T'ube), are shown. They were created using Paradigm Geophysical's commercial 3D volume visualization computer program, VoxelGeo (VG). Each of these D3D-VG renderings was created using very few layered-earth assumptions, and they allow the opaque contents (three-dimensional objects, or CIOs) of the virtual D3D-impedance time-cube, to be observed the same way that fish, hovering in an aquarium full of clear water, can be observed by simply looking down into the top of the tank. These right-hand maps are not "interpreted" amplitude maps, extracted from an interpreter's smooth (possibly faulted) 2D horizon time-picks on tens or hundreds of inlines and crosslines. Instead, these pink, red, and yellow CIOs were "grown", separately, by planting VG "detections seeds" in low-impedance "starting VOXELs", corresponding to the depths of the prospective (un-drilled at the time, in EI-27) and the logged (in EI-46) gas sand reservoirs. Each "Starting VOXEL" is a D3D-reprocessed impedance-trace-sample, sitting at a TWT (converted from its logged depth) on the well path that penetrated either the upper EPL #1 CIO, or the lower Norcen #2 CIO, respectively.
So, where did these extremely useful D3D seismic data come from? In 1995, a large non-streamer-cable (telemetry, single-sensor per recording station, compact 2x7 airgun source array), speculative 3D survey was recorded by Fairfield Industries, using a vertical sampling interval of 3 ms, and then processed using a (Gulf Coast conventional) 4 ms re-sampled interval, to create a (55-feet x 55-feet) square grid of post-stack-time-migrated traces. With the airgun source array using one “pop” per shot record, the acquisition specs fit reasonably well to the D3DSP’s known-point-source-and-point-receiver position, field acquisition requirement. After EPL purchased a non-exclusive license to use these data, in 2002, they were D3D-reprocessed using a sample interpolator to convert each (pre-stack) trace from a 3 ms to a 2 ms vertical sample interval. The high resolution delineation of the CIB CARST gas reservoir sand (note the yellow clean-sand meander and flat gas-water-contact [GWC] base in Figure DvC2-06) proves the value of this unconventional, interpolated-re-sampling step. And it was motivated by the desire to image buried "objects" rather than any assumed layers. Conventional "signal-to-noise" filter-testing invariably sees the laterally continuous (layer-like), overlapping-peaks and troughs as seismic "signal", and demanded that the non-layered peaks and trough be branded as "noise" and removed. Conventional recording and processing form a self-fulfilling prophecy that seismically transforms real, buried objects (CIOs) into expected geological layers, and thereby blurs much of the most valuable, high-resolution information.
Once again, it is crucial to this story that when these Fairfield data were recorded, the southern (Norcen #2) gas reservoir had not yet been drilled, and the northern prospective (EPL #1) reservoir had already had its reservoir pressure (only slightly) drawn down, presumably due to the (1977-84) production of 8.2 bcfge, from the Odeco #4 well. This Odeco well is now thought by some to have tagged the up-dip edge of the EPL #1 discovery well gas sand, as shown schematically by the red dashed well path in Figure DvC2-02. But it was abandoned, in 1984, due to high water cut, eleven years before the 3D seismic survey was shot. It is quite possible that the Odeco #4 well's perforations spanned two pressure-isolated sand units, with some of the gas and all of the water coming from a separate (lower CIB CARST?) sand, analogous to the lower orange sand (labeled "E" in Figure DvC2-02). This encroaching lower CIB CARST sand water contact may have effectively shut off the few feet of gas-productive perforations (from a separate sand lens), just above it. Low volume (one or two perforations?) gas production could have been contributed by a pressure-separated, very thin gas sand analogous to the up-dip, overbank edge of a thicker reservoir sand package. See the lower, yellow turbidite-channel sand that can be seen on the Hollywood Quarry cross-section (Figure DvC2-02). The actual, complex CIB CARST sandstone (probably not a turbidite sand), over which the Odeco #4 well was perforated, can also be seen by zooming in on the bottom-center log in Figure DvC2-08. This well's Spontaneous Potential (SP) log gives a good measure of rock permeability, and is to the left of the Deep Induction resistivity, used to indicate fluid resistivities in a good (SP) sandstone. Petroleum products have much higher resistivities than salty brine. The area under the induction (resistivity) curve is colored red across the 34-feet of perforated casing (10,456'-10,490'), but only 20-net-feet (by SP) of gas sand is measurable on the SP log, in two zones separated by an impermeable shale break.
Prior to the D3DSP work, it had been recognized that Norcen had drilled their #2 well (1996, EI-46), and made a CIB CARST well (southernmost red circle on the same map in Figure DvC2-05), with its gas pay sand shown in red on the Norcen #2 log (lower-left corner of Figure DvC2-08). It was a (conventionally) puzzling well because this log (lower left of Figure DvC2-08) shows less than 20 feet of resistive sand, above a possibly connected wet sand, and the conventionally processed amplitude anomaly (left-hand map, southern anomaly, Figure DvC2-05) measures an anomalous area of 244 acres. Using a maximum of 20 feet of sand, over the full amplitude anomaly yields only 4,900 acre-feet for the conventional seismic volume estimate. This gives an estimated ultimate recovery of less than 6 bcfge, when multiplied by the accepted 1,200 mcfge/acre-ft "Recovery Factor", supplied by engineering data. At the time of the D3DSP analysis, this well had already produced over 26 bcfge, and by the time EPL's discovery had been drilled and logged, it had begun making significant water and its estimated ultimate recovery was put at 29 bcfge. Using the transit time and bulk density logs from the (thin pay sand) Norcen well, a synthetic seismogram could not produce the high amplitude seismic response seen on the conventional map, around the well. These high quality logs, run by Norcen, made this puzzling synthetic seismogram work hard to refute, and these (8,000 feet per second) gas sand velocities were later used to help calibrate the thickness of the EI-27 D3D-impedance response.
The question in the minds of many (EI-27 prospect potential partner) geoscientists was, "Where was all the EI-27 gas coming from?" Without a good, conventional-data answer to this question (and others), higher risk had to be assigned to the EI-27 prospect. In this post-3D-recording well (a little more than 14,000-feet south of the EPL #1 discovery well), there were less than 20-feet of surprisingly (weak high-resistivity) productive gas sand, located within the conventionally mapped “bright spot” amplitude anomaly (bottom, left-hand map, Figure DvC2-05). It was impossible to reproduce a high amplitude synthetic seismic trace that would match the conventional response seen at the Norcen well, using the 20 Hz wavelet that was measured on the originally processed spec data (lower left panel, Figure DvC2-09). At EI-27, the watered-out gas sand (after a little fluid replacement work) produced a synthetic trace signature consistent with what was seen on the conventional data, but the 8.1 versus 8.2 bcfg coincidence loomed large and risky. A hypothesized velocity gradient (with depth-conversion velocities increasing to the west, to produce a conventional TWT artificial pull-up that would put the abandoned Texaco/Odeco #4 well [with an uncertain KB] down-dip from the prospective EI-27 "attic" gas) was not supported by any of the surrounding Velocity Surveys or the reprocessed stacking velocities. Because of all the risks and unanswerable questions, the EI-27 prospect was placed near (or at?) the bottom of a ranked list of Ready-To-Drill prospects. In the end, the EI-27 Prospect "sales presentation” proved too risky and too small to attract a drilling partner.
In December, 2002, in EI-27, the EPL #1 vertical well discovered approximately 35 feet of net (50 feet gross) pay in the CIB CARST sand at a -10,420-foot subsea elevation. The D3DSP had produced precise (and now known to be quite accurate) predictions, using analyses of the D3D-impedance-reprocessed seismic over and around the vertical-hole that was eventually drilled. As mentioned, the actual drill site was chosen prior to the D3DSP analyses, from the spec-processed and conventionally interpreted seismic amplitude (Bright Spot) anomaly, seen in the upper left of this Figure (DvC2-05). Depending on the compressional velocity of the gas-filled reservoir, the D3DSP predicted that from 32- to 40-feet of gas sand would be encountered at –10,450 feet, subsea elevation. It also predicted that approximately 18 BCF of recoverable gas would be found, DOWN-DIP to the Odeco #4 well that had produced 8.2 BCF, before apparently watering-out in 1984. Its recoverable-gas reserves were confidently supported by the D3DSP analyses at the Odeco #4 and Norcen #2 (which had cum'ed 26+ bcfge, from 1995-2002), but a more up-dip, to the west, location was recommended by the D3DSP. A very D3DSP-significant result of the EI-27 "subtle trap" discovery (or re-discovery of by-passed pay?) well is that it was drilled (a) without a confident structure map, because of the uncertain Odeco well KB, (b) down-dip to a well that was thought to have watered out (the Odeco well), with no obvious time-structural, fault, or stratigraphic barrier interpretable on the conventional data, and it (c) found a highly commercial gas reservoir. During one month in 2003, its cash-flow was more than three times that of the next best well. This EPL #1 exploratory well at EI-27 is a classic example of the D3DSP principle of drilling into three-dimensional reservoir objects (low-D3D-impedance CIOs) that can be visualized in three dimensions (e.g., using VoxelGeo), even if a conventional "trapping mechanism" is difficult to understand.
In Figures DvC2-10, -11, and -12, the first frames of two animations (movies, that can be seen elsewhere in this website) that show:
DvC2-10 - Two spinning D3D-impedance-CIO reservoirs (in EI-46 and EI-27), created after the EPL discovery well logs were available, and used to calibrate the CIO-impedance "cutoff" value (-68) for both reservoirs.
DvC2-11 - The "Seeded-Growth Evolution" movie, created to show how continuously changing the D3D-impedance detection criterion (or impedance "cutoff") allows the "Starting VOXEL", planted in the CIB CARST gas sand in the EPL #1 well, to grow from a very compact, highest quality (and hence, lowest impedance) common-impedance object channel sand, to the maximum-size CIO, based on using a "cutoff" derived from the analog producing CIO nearby.
DvC2-12 - A montage of all 36 frames of the DvC2-11 Evolution movie, showing the basis for choosing -68 as the D3D-impedance "cutoff" for the EI-27 and EI-46 CIB CARST gas reservoirs.
And, applying VoxelGeo's volume visualization and analyses capabilities to this buried (virtual, three-dimensional, common-impedance) "object", the D3DSP made mostly confirmed estimates of:
Structural Elevations (depth structure map)
Area of Pay (reservoir outline)
Thickness of Pay (isopach map),
Physical Volumetrics (approximate volume), and
Estimated Ultimate Recovery (using an engineer-supplied Recovery Factor).
|
