Initial Evaluation of Structural and Stratigraphic Compartmentalization in the Pony-Knotty Head Field, Green Canyon, Deep-Water Gulf of Mexico
Published:December 01, 2012
- PDF LinkChapter PDF
Bill Kilsdonk, C. Robertson Handford, 2012. "Initial Evaluation of Structural and Stratigraphic Compartmentalization in the Pony-Knotty Head Field, Green Canyon, Deep-Water Gulf of Mexico", New Understanding of the Petroleum Systems of Continental Margins of the World, Norman C. Rosen, Paul Weimer, Sylvia Maria Coutes dos Anjos, Sverre Henrickson, Edmundo Marques, Mike Mayall, Richard Fillon, Tony D’Agostino, Art Saller, Kurt Campion, Tim Huang, Rick Sarg, Fred Schroeder
Download citation file:
Reservoirs at Pony-Knotty Head Field consist of stacked, middle Miocene (Serravallian) turbidites deposited as high-frequency low-stand successions within an increasingly ponded basin. Depositional elements include: (1) high to moderate permeability channel axes, channel margins, channelized lobes, and amalgamated lobes; and (2) those having low-permeability, such as marginal to distal lobes, levee-overbank debrites, slumped mudsheterolithics, and pelagic/hemipelagic muds. Fluid pressure data demonstrate that the Pony-Knotty Head Field is segmented into pressure compartments at multiple scales.
Although the field is a low-dip, faulted, four-way turtle structure, interpreted faults are neither long enough nor have sufficient throw to segment reservoirs into observed pressure cells. Analyses of individual reservoir units indicate that variations in fluid potential are often greater vertically within wells than laterally between wells. This pattern indicates that at least some segmentation at this scale is due to low-dip stratigraphic barriers between depositional elements rather than to steeply dipping barriers, such as faults.
At the field scale, both fluid pressures and depositional elements change vertically. Excess pressure was used to help define compartments at Pony-Knotty Head Field. (“Excess pressure” is the difference between pressure measured in a well and pressure calculated using a datum with an expected fluid gradient.) The deepest reservoirs have the lowest excess pressures. They are dominated by laterally continuous, unconfined depositional elements that bled excess pressure laterally. Progressively shallower reservoirs have progressively higher excess pressures in progressively more confined depositional elements. Between reservoirs of different depths and ages, stratigraphic complexity increased with time as increasing structural confinement of the depocenter above mobile salt drove stratigraphic evolution from a lobe-dominated system to a channelized lobe and levee-channel complex system. We propose that compartmentalization at this scale results directly from stratigraphic responses to the structural evolution of depocenters.
The Pony-Knotty Head Field lies in Green Canyon, 150 miles south of New Orleans and 30 miles northwest of the Sigsbee Escarpment (Kilsdonk et al., 2010, 2011). Middle Miocene (Serravallian) reservoirs are buried to more than 30,000 ft: below 3400 ft of water, a thick allocthonous salt canopy, as well as post reservoir clastics both above and below the canopy. The trap is a low-relief, faulted turtle structure that trends broadly northeast-southwest (Fig. 1). It formed during evacuation of deep salt into the overlying canopy after reservoir deposition, from the Tortonian through Pliocene.
The Pony-Knotty Head oils fill middle Miocene (Serravallian) turbidite sands developed near the base of fourth-order fining-upwards sequences (Fig. 2A). Depositional elements indicate an approximately 1.5-2 Ma history of third- and fourth-order cycles showing an overall evolution from unconfined turbidite lobes to partly confined lobe and channel elements (Fig. 2B).
During reservoir deposition, the Pony basin was rimmed on three sides by rising salt structures and bordered by a salt-cored fold to the southeast. Both semiregional and local section restorations indicate that compressional shortening accelerated in the Pony basin and the surrounding area during the middle Miocene. Stratigraphic analyses of our middle Miocene reservoirs indicate increases in both basin confinement and stratigraphic variability during the Serravallian.
Fluid pressure and geochemical data from exploration and appraisal wells in the Pony-Knotty Head Field clearly show reservoir compartmentalization at multiple scales. We discuss reservoir compartmentalization at two scales of interest for field development and planning: (1) field scale and (2) reservoir scale. At the field scale there are three broad pressure compartments (Fig. 3). Each is separated from the next by hundreds of psi’s (pounds per square inch) of excess pressure in hydrocarbons and brine; each corresponds to one of three main reservoir intervals; and each is further segmented into smaller compartments at the reservoir scale. At the second, or reservoir scale, there are secondary pressure cells separated by tens of psi’s of hydrocarbon excess pressure within each reservoir interval.
We propose that at both scales, compartments result largely from stratigraphic architecture and, even though the stratigraphic architecture is influenced by structural and tectonic processes, structural features play only supporting or indirect roles in reservoir compartmentalization. Our analyses of these early stage data indicate that reservoir compartments in this field are largely formed and separated by stratigraphic barriers characteristic of stacked turbidite lobes. In contrast, faults within the field have neither sufficient length nor sufficient seal capacity to separate large pressure compartments. Rather, non-sealing faults commonly juxtapose different reservoirs against one another to bridge stratigraphic compartments that might have existed in their absence.
Primary Scale Compartmentalization
Each of the three main reservoir packages: Ser 1 (~13.8 Ma), Ser 2 (~13.2 Ma), and Ser 3 (~12.4 Ma) has excess hydrocarbon and brine fluid pressure, but each to a differing degree (Fig. 3). In broad terms, the deepest and oldest reservoir package (Ser 1) has the lowest excess fluid pressure of the three, the youngest and shallowest (Ser 3) has the highest excess pressure, and the middle reservoir package (Ser 2) has intermediate excess pressure. These primary scale pressure differences are clearly controlled by hydrocarbon seals and low permeability shales that are part of the stratigraphic architecture between reservoir units. The observed pattern of excess pressure with depth is opposite to what would be expected from purely depth or temperature dependent processes (disequilibrium compaction, etc.) operating in one dimension:i.e., the deepest reservoir package has the lowest excess fluid pressure; the youngest and shallowest has the highest excess pressure; excess pressure in the middle package is intermediate.
Additionally, petroleum geochemical data indicate differences between oils reservoired in the Ser1, Ser 2, and Ser 3 sandstones. The differences are manifested by a general increase in hydrocarbon maturity from the shallowest reservoir unit (Ser 3) to the deepest (Ser 1), although there is some overlap. This pattern indicates, in broad terms, that the shallowest reservoir received earlier generated, less mature, oils than the middle reservoir, which in-turn received earlier generated, less mature, oils than the deepest reservoir (Fig. 4).
We propose that this pattern of upward increasing reservoir overpressure resulted from structural/tectonic influences on turbidite depositional architecture. An upward decrease in 3D sandstone connectivity through 1.5-2 Ma corresponded to increasing depositional confinement driven by increasing structural isolation of the depocenter.
Structural movement and depositional confinement occurred during a middle Miocene semi-regional transition in the dominant structural processes. The transition passed from: (1, beginning middle Miocene) nearly symmetric depocenter growth balanced largely by evacuation of underlying salt into depocenter flanking salt stocks; to (2, end middle Miocene) asymmetric depocenter growth influenced by compressional folding near the down-slope limit of salt detached deformation (Figs. 5 and 6).
We infer that the onset and acceleration of compressional folding lead to increased confinement, smaller depocenter size, more heterogeneous facies distribution, and decreased area of continuous sand deposition. The deeper more connected sandstones are more likely to extend over structural highs where they may have bled-off some excess pressures (see Mann and Mckenzie, 1990). In contrast, the same data imply that the shallower reservoirs may be less homogeneous and more highly compartmentalized by stratigraphic barriers, limiting hydraulic connections to shallower pressure relief points.
Reservoir depositional elements and stratigraphic interpretation
The 3D seismic image of the subsalt field is sufficient to define structural geometry and faulting, but it lacks the resolution and frequency content required for detailed stratigraphic interpretation. As such, we formulated a stratigraphic depositional facies framework for the middle Miocene reservoirs by combining core analyses and observed stacking patterns with log based facies interpretations; dipmeter data; stratigraphic correlations; and analog data.
This approach led to a hierarchical subdivision of reservoirs into depositional elements and element complexes. Depositional elements were mapable packets of genetically related strata (Allen, 1983) that formed building blocks to interpret stratigraphic expressions of depositional systems. Each represented the cumulative depositional product of many dynamic events over periods ranging from tens to thousands of years (Miall, 1988). Depositional elements interpreted from the Pony-Knotty Head core data represent: channels and channel-margins; channelized lobes; overbank and levee deposits; amalgamated lobes; marginal-distal lobes; deformed muds/slumped heterolithics; and hemi-pelagic muds.
Middle Miocene sands entered the Pony-Knotty Head depocenter in confined, turbidity flows from the north via relatively narrow passes between salt walls and ridges. The flows then spread out as fans onto a less-confined basin floor. A local increase in net sand in Pony-Knotty Head suggested that a local, partially confined, depocenter or ponded mini-basin developed during deposition in response to evacuation or withdrawal of underlying salt. The narrow fairway of thickened sand between salt welds supporteed partial confinement of the depocenter by structurally elevated basin flanks to the east and west. Net sand decreased to less than 400 ft south of the depocenter, indicating a component of ponding there.
The stratigraphic pile in the Pony-Knotty Head area can be subdivided into a hierarchical arrangement of small- to large-scale depositional elements (Fig. 7). Recognition of the various scales comes from sedimentological analysis of the core tied to well-log curves and e-facies stacking patterns. For example, large-scale depositional elements in Pony-Knotty Head are lobe complexes, which are interpreted as mainly fourth-order lowstand systems tracts bounded by chronostratigraphic surfaces. Lobe complexes, generally 100-200 ft thick, are composed of individual lobes, which may be ~50 ft or less, as well as channels and associated elements. Individual lobes and channels may be bounded by fourth-order surfaces and autocyclic boundaries. Smaller scale features include lobe and/or channel elements, at the parasequence or high-frequency cycle scale, which are mainly bounded by flooding and/or abandonment surfaces, as well as autocyclic bed-set boundaries (Fig. 7).
Modern and Pleistocene analogs (including submarine fan complexes from the Kutai basin, Indonesia; offshore Nigeria; and the Brazos-Trinity slope system, offshore Texas) provide estimates of the size, or scale, of depositional elements (Fig. 8). The first two examples are nearly the same size as we estimate for the Pony-Knotty Head depocenter. The third, which is half the size, is scaled up to match. Calculated averages suggest that Pony-Knotty Head lobes may have been roughly 4.6 miles wide and 3.9 miles long and channels could have been 1-6 miles long and 0.5-1 mile wide.
Supporting stratigraphic data and interpretation
In general the overall log patterns and core data are interpreted as a 2 m.y. record (Serravallian) of fan outbuilding and retreat (Fig. 2). Furthermore, the patterns indicate that the Pony-Knotty Head depocenter became increasingly more confined and channelized throughout the Serravallian. Salient characteristics of each of the three reservoir units interpreted from core and log data are listed below.
The amalgamated sands are blocky with a sharp base in core and log.
The sands have low gamma ray signatures on logs.
We interpret this pattern to indicate deposition of Ser 1 reservoir sands in sand-rich amalgamated lobes and marginal-distal lobe elements, characteristic of deposition on a largely unconfined basin floor.
Ser 1 reservoirs are easily correlated with only minor changes in log/rock character between wells.
The middle reservoir (Ser 2) contains amalgamated lobes and marginal-distal lobe elements plus channelized lobe, distributary channel-axis, and channel margin elements.
Compared to the lower (Ser 1) reservoirs, the middle reservoirs (Ser 2) are more variable (clean to heterolithic), more difficult to correlate, and show greater variability in log character (blocky and fining-upward, steeper dip-meter).
We interpret this pattern to indicate greater lateral variability in depositional processes and elements (i.e. some channelization) compared with the Ser 1 reservoirs.
Upper reservoirs (Ser 3) show the greatest variability in sand content, bed thickness, and log patterns (fining-upward, coarsening upward, blocky, serrated).
Ser 3 sands are the most difficult of the three to correlate between wells, both on logs and in core.
The upper reservoir (Ser 3) contains all elements of the two deeper and older reservoirs as well as levee-overbank elements.
Sedimentologic and stratigraphic features indicate a progressive increase in lateral variation of depositional setting and processes between the deposition of the oldest reservoir unit (Ser 1) and deposition of the youngest reservoir unit (Ser 3). This is consistent with channelization and confined flows as well as deposition from unconfined flows and outbuilding. As a whole, the Ser 3 reservoirs probably have been deposited as lobes, channels, and overbank elements.
Proportions of depositional elements within each major reservoir interval are graphed in Figure 9. Figure 10 shows the inferred spatial representation of the lower, middle, and upper reservoir intervals with respect to the depositional elements in the schematic depositional model. The stratigraphic distribution of elements and their inferred positions within the depositional model indicate that there was an overall basinward shift of facies tracts through time and that the depocenter became increasingly confined.
The overall basinward shift, however, was punctuated by several higher frequency cycles of fan outbuilding and retreat, driven by local to regional changes in accommodation and sediment supply. As a result, the middle and upper reservoir intervals (Ser 2 and Ser 3) contained most of the depositional elements shown in Figure 10, but in varying proportions.
Secondary Scale Compartmentalization: Pancakes or Pie Slices?
In addition to the large scale, or primary, compartments between reservoir units, fluid pressure data from initial exploration and appraisal drilling clearly indicate additional, or secondary compartments, expressed as pressure cells that are separated by tens of psi’s of hydrocarbon and brine excess pressure within each reservoir unit. Understanding the nature and geometry of compartments at this scale, within reservoir units, is critically important for field development.
We examine two separate, though not mutually exclusive, hypotheses for the origin of observed reservoir compartments: (1) Compartments are formed and separated by faults that impede or prevent hydrocarbon fluid flow; and (2) Compartments are formed and separated by clay-rich stratigraphic elements that impede or prevent hydrocarbon flow. The two hypotheses imply very different geometries for the pressure cells, or reservoir compartments, and the boundaries between them. Stratigraphic barriers should form reservoir compartments having low dip boundaries, like a plate of overlapping pancakes. In contrast, fault barriers should form reservoir compartments with high dip boundaries, like pie slices.
In order for faults to form reservoir compartments by separating and isolating hydrocarbon accumulations, they must have either: (1) sufficient throw to juxtapose a sandstone reservoir against a sealing shale across the entire length of the compartment boundary; or (2) sufficiently low permeability in the fault gouge to form a membrane seal across the entire length of the compartment boundary (Knipe, 1997). In either case, the sealing fault must extend fully across the accumulation, and the sealing mechanism must be fully developed along that entire length. To examine those parameters and evaluate the likelihood of reservoir segmentation by faults, we both created fault plane juxtaposition diagrams and calculated shale gouge ratios (SGR’s) on cross-fault reservoir juxtapositions.
Our analyses of this early stage data set indicate that reservoir compartments in the Pony-Knotty Head Field are primarily caused by stratigraphic barriers, not by faults. Well data indicate that boundaries between reservoir compartments identified by fluid pressures have low dips, characteristic of stratigraphic elements. They appear to be stacked like pancakes, not cut by faults like pie slices. None of the faults detectible from 3D seismic data have sufficient throw along their lengths to form compartment boundaries. In fact, all detectible faults tip out within the field limits. We are able to detect and interpret faults having throws down to 50m using our 3D data set, so it is unlikely that seismically invisible faults are able to segment reservoirs over the required length of the entire field.
Although we see no direct evidence in our current data, faulting may in fact contribute to compartmentalization where narrow lobe-confined compartments cross high throw portions of faults. However, it appears from our data that within this field faults juxtapose different sandstones lobes within reservoir units to allow cross lobe flow and enlarge rather than decrease compartment size.
Data and methods
MDT pressure data were analyzed using the excess pressure method. Values of excess pressure were estimated and used to interpret potential pressure compartments and pressure compartment boundaries within and between reservoir units.
Seismic and Vshale interpretation for fault analyses
Our fault analyses were based on horizons tied-to-wells and fault interpretations of 3D seismic data. Iterative reinterpretation of the seismic data was prompted by analyses of fault geometry, intersections and branch lines, fault throw distributions, and fault linkages. We used resulting fault and horizon interpretations, with petrophysically interpreted Vshale curves from seven (7) wells, to evaluate cross-fault juxtaposition and along fault SGR. The Vshale curve from a single well, Pony 1 BP1, was used to construct fault triangle diagrams.
Reservoir structure maps
We evaluated likelihood of reservoir compartmentalization by faults using seismic based, tied-to-wells, maps of reservoir structure; loci of well cuts in reservoir intervals; excess pressures at those cut points; and the positions of cut points relative to mapped faults. These were all coupled with fault juxtaposition and SGR analyses (Yielding, 2002) to interpret the likelihood of fault seal.
We coupled stratigraphic interpretation of turbidite depositional elements based on Pony-Knotty Head logs and core with our excess pressure analysies to interpret potential stratigraphic compartments and boundaries between them.
Geochemical data were considered for this study but found to be inconclusive with respect to compartmentalization of individual reservoir units. Nevertheless, oils were segregated geochemically between three broad reservoir packages (Ser1, Ser2, and Ser 3), with the earliest generated oil in the shallowest package (Ser 3) and the latest generated oil in the deepest reservoir package (Ser 1). This distribution fits a migration model of the oldest oil having the most time to migrate into the reservoir furthest from the source rock.
Fault analysis methods
The triangle diagram in Figure 11 is a simple tool to show the effect of varying fault throw on cross fault juxtaposition and SGR (shale gouge ratio) for the stratigraphic column of interest. SGR is an estimate of the shale fraction within the fault zone which, in a sand-shale system, largely controls the cross-fault permeability of juxtaposed reservoirs. Higher values of SGR indicate higher clay or shale fractions, and greater seal potential. SGR is calculated at each point on the fault as a fraction equal to the fraction of shale in the section that has slipped past that point (Knipe, 1997).
We calculated juxtaposition and SGR on all four seismically mapped fault surfaces using Badley’s TrapTester software. All calculations were done after geometrically constrained reinterpretation of fault planes and fault throws.
We interpreted the traces of footwall cut-offs and hanging-wall cut-offs of primary horizons on all faults where primary horizons abut fault surfaces.
We calculated footwall cut-offs and hanging-wall cut-offs of the tops and bases of individual reservoir packages based on their vertical distances above, below, or between primary cutoffs depending on the geometry of the primary cutoffs. Distances, or section thicknesses used, were calculated based on section thicknesses in wells and interpolated by relative distances to the wells.
Vshale distributions were calculated on both the footwall and the hanging-wall sides of fault surfaces using horizon cut-offs and Vshale curves.
We filtered the Vshale distribution on the fault surface using a cutoff (Vsh >30%) to show reservoir sand distribution on each side of the fault. We overlaid Vshale plots from both the footwall and hanging-wall sides of fault surfaces to show sand-on-sand cross fault juxtaposition (Fig. 13).
We calculated SGR over entire fault surfaces and displayed it in areas of sand-on-sand juxtaposition (Fig. 14).
Figure 11 is a standard triangle plot output from Badley’s TrapTester. The vertical axis is depth (ft) and the horizontal axis is fault throw (m). A Vshale log, from the Pony 1 BP1 well, is used to represent both footwall and hanging-wall stratigraphy. In the plot, footwall stratigraphy has a constant depth, but hangingwall stratigraphy is depth shifted to represent fault throw. The fault throw and hanging-wall depth shift are both zero at the left end of the plot and increase to the right. Horizontal lines represent the tops and bases of reservoir intervals on footwall cut-offs. Dipping lines represent the tops and bases of reservoir intervals in hanging-wall cut-offs. Cross fault juxtaposition is indicated by crossing lines. Calculated shale gouge ratio (SGR) is displayed by color, with values indicated by the color bar.
Faults in the Pony-Knotty Head structure have throw distributions ranging from 0 (no throw) at fault tip(s) to a maximum throw somewhere on the interior of the fault plane. Variations in throw, both horizontally and vertically, have led to variations in sealing properties of faults along their surfaces. Because each fault has a range of throw values over its surface, there is a corresponding range of cross-fault juxtapositions and SGR values
For any value of throw greater than zero, the footwall (FW) cut-off of a reservoir will have different juxtaposition and SGR value than the hanging-wall (HW) cut-off of the same reservoir. At zero throw, the faults don’t actually exist and cannot seal. We estimate that faults with more than 50m throw are clear and well interpreted on our 3D seismic data. Because seismic resolution is limited, we use triangle diagrams to qualitatively understand cross fault juxtaposition of reservoirs and the potential for fault gauge seals as a function of throw below and near the limits of resolution.
20m throw on triangle plot
With only 20m throw, which may not be evident on seismic data, the Ser 1 reservoirs are fully juxtaposed against shale (both HW and FW) and will likely be sealed, though it is important to recall that this applies only to the part of the fault with 20m throw. If a fault’s maximum throw is 20m, or even 50m, it is unlikely that the fault will span the entire hydrocarbon-bearing reservoir: leaks are likely near and around fault tips. Similarly, the upper sand of the Ser 2 interval in the footwall block is juxtaposed against shale in the hanging-wall block. Deeper sands of the Ser 2 are either self-juxtaposed or juxtaposed against one another that have low (<20%) values of SGR. In contrast, although the Ser3 is self-juxtaposed, it has SGR greater than 50% and is likely to seal. The same 20m throw juxtaposes various middle Ser 2 sands with one another at values of SGR so low, near 15% or 20%, they are unlikely to seal. The lower Ser 3 sand is self-juxtaposed and has SGR values between 15% and 45%, unlikely to seal at the low end. In contrast the upper Ser 3 sand in the (FW) is juxtaposed against a Ser 1 sand (HW) and has a potentially sealing SGR greater than 45%.
40m throw on triangle plot
At 40m throw the upper Ser 3 sand (FW) and some of the Ser 2 sands (FW and HW) are juxtaposed against shale. SGR on the Ser 3 drops to near 40%. The upper Ser 2 (HW) sand is juxtaposed against the mid-Ser 2 sands (FW) with SGR near 40%. The SGR on most mid-Ser 2 juxtapositions are between 15% and 20%, likely to leak.
60m throw on triangle plot
At 60m throw, many Ser 2 and Ser 3 sands are fully juxtaposed against shale. However, the lower sands of the Ser 2 (FW) are juxtaposed against the upper sands of the Ser 2 (HW) having SGR between 15% and 20%, very likely to leak.
80m throw on triangle plot
By 80m of fault throw, the lower-most Ser 2 sand (FW) is juxtaposed against the upper-most Ser 2 sand (HW), which has a SGR between 20% and 30% and is unlikely to seal. All other reservoir sands are juxtaposed against shale.
Fault plane juxtapositions and SGR
Four large faults cut the Pony/Knotty head structure near the crest. All strike roughly northeast-southwest and dip to the southeast. All mapped faults, at all mapped levels, tip out (either at one end or at both ends) both within the closing structure and above the oil-water contact. It is highly unlikely that the faults seal near or at the fault tips. Similarly, it is implausible that faults seal beyond fault tips where the faults do not exist.
Of course, smaller unmapped faults exist, but since they are smaller than can be interpreted with the current data they are also unlikely to have sufficient length and displacement to provide barriers that cross and compartmentalize the field on their own. It is, however, possible that small displacement faults could segment small, forth-order, turbidite lobes (Fig. 7), if the lobes themselves form small compartments.
Faults, as interpreted from seismic data, are neither long enough nor have sufficient throw distribution to separate observed pressure compartments. Additionally, fluid potential analyses indicate more variation vertically within individual well than horizontally between wells. That pattern strongly suggests at least a component of compartmentalization by stratigraphic barriers. The results of this study indicate that observed pressure compartments result from an interaction between stratigraphic barriers and fault offset, reservoir compartmentalization largely is due to startigraphic variation, and faults primarily allow cross-fault communication between reservoir units.
Juxtaposition plots for four mapped faults were built by extracting cut-off lines onto fault surfaces and using those cut-offs to guide both hanging-wall and footwall projections of Vshale, interpolated from well logs, onto each fault surface (Fig. 13). SGR values, filtered to show only on sand-sand juxtapositions, clearly indicate lack of seal near at and beyond fault tips (Fig. 14). Portions of some faults are likely to provide local seals due to juxtaposition of reservoir sands against shale but the sealing portions do not span the field within the hydrocarbon column. These faults are therefore highly unlikely to segment the field into pressure compartments unless they cross and combine with stratigraphic seals.
Stratigraphic pressure cells and reservoir compartmentalization
Fault analyses indicate that mapped faults, on their own, are unlikely to segment the Pony-Knotty Head field to the degree dictated by the fluid pressure data. Compartmentalization in this field is primarily due to stratigraphic boundaries and only secondarily due to combinations of stratigraphic boundaries associated with large throw portions of faults. Depositional models derived from log and core data show a northern source of sediment within the Pony-Knotty Head depocenter. In some reservoirs, we suggest that north-south to north-northeast/south-southwest elongated turbidite lobes incorporate debrites, which may parallel or cross the depositional axis. Shale breaks between lobes and low permeability debrites provide stratigraphic boundaries sufficient for the observed segmentation. Figures 15–18 are maps of potential pressure cells formed by turbidite lobes and crossing faults. Data and analyses indicate that, at the reservoir scale, faulting and juxtaposition in the Pony Field are likely to: (1) permit cross-fault, cross-strata flow near fault tips where throw is less than sand package thickness; and (2) be flow barriers where throws are larger than sand package thicknesses. Because all observed faults lose throw and tip out within the field, it is unlikely that faults alone compartmentalize the field into the observed pressure cells.
Excess pressure values (representing fluid potentials in oil) were extracted as single values relative to an arbitrary, but consistent, datum and plotted at appropriate cut points on structure maps. The resulting maps illustrated the spatial distribution and values of oil-phase fluid pressure data at each reservoir level.
Two features of fluid pressure distribution are apparent from the maps. (1) Some wells, at some reservoir intervals, recorded fluid pressure values that indicate vertical stacking of multiple pressure cells. This pattern indicates compartmentalization over tens or hundreds of feet in the vertical direction. (2) In contrast, multiple wells, separated by thousands of feet on the maps, share very similar values of fluid pressure. This pattern allows the possibility of laterally extensive compartments. These observations indicate that, within individual reservoir units, there is more variation in fluid pressure within wells than between wells. At least some component, and perhaps all, of compartmentalization must result from low dip stratigraphic pressure and flow barriers.
On the structure maps in Figures 15–18, colors describe interpreted fluid flow characteristics of fault surfaces based on juxtaposition and SGR analyses. A red line indicates that the fault is most likely sealed on the side shown. A green line marks a fault that is most likely self-leaking: it allows fluid flow across the fault within the mapped sand. A blue line indicates that the fault is likely to allow cross-fault flow to a reservoir of a different age. The direction of potential cross-fault flow is indicated by an arrow showing the direction of juxtaposition. Yellow lines indicate cross fault juxtaposition of reservoir sands having SGR values that could either leak or seal. Excess pressure values are posted on the maps at appropriate cut points. These data are coupled with approximations of turbidite lobe size and shape to sketch potential pressure cells (reservoir compartments) on the maps. Lobe outlines are drawn to honor the pressure data, take advantage of fault seals where available and required, and multiple pressure values in a single sand unit are taken to indicate proximity to a lobe edge. These plots (as well as Figs. 19 and 20) suggest that pressure cells may cross from one sand into another. If so, pressure equalization may occur by vertical communication, cross-fault and cross strata communication, or a combination of both.
Estimates of excess pressure values were correlated with MDT sample points and reservoir packages on the logs. They were then posted on well log sections at the appropriate reservoir package so possible pressure cells could be interpreted from the combination. The logs were ordered from east to west in an attempt to capture variations in stratigraphy resulting from a northern source of turbidite fan systems. Interpreted depositional elements were used as guides to identify reservoir compartments and compartment boundaries (Figs. 19 and 20).
Summary and Conclusions
We explore the nature of pressure cells, both at the field scale and at the reservoir scale, using initial well data tied to seismic interpretation of the Pony-Knotty Head Field. We propose that, at both scales, compartments result largely from stratigraphic architecture, and even though that stratigraphic architecture is influenced by structural and tectonic processes, structural features play only supporting or indirect roles in reservoir segmentation.
At the field scale, three large-scale pressure cells: (1) show hundreds of psi’s difference in excess pressure; (2) show geochemical differences related to different hydrocarbon maturity levels; and (3) correspond with three main reservoir packages. In broad terms, the deepest and oldest reservoir package has the lowest excess fluid pressure of the three, the youngest and shallowest has the highest excess pressure, and the middle reservoir package has intermediate excess pressure. These pressure cells are clearly separated by low permeability shale sections that lie stratigraphically between reservoir units. However, the observed pattern of excess pressure with depth is opposite to that expected from purely depth or temperature dependent processes. We propose that this pattern of upward increasing reservoir overpressure results from structural/tectonic influences on turbidite sandstone architecture which lead to
higher three dimensional sandstone connectivity deeper in the Serravalian. The deeper, more connected sandstones are more likely to extend over structural highs where they may have bled-off higher fractions of excess pressures. In contrast, the same data imply that the shallower reservoirs may be less homogeneous and more highly compartmentalized by stratigraphic barriers, limiting hydraulic connections to shallower pressure relief points.
This conclusion is supported by regional and local reconstructions of the structural and tectonic evolution which indicate that the reservoirs were deposited during an interval of accelerating compression related to shortening just landward of the downdip limit of the underlying salt detachment. This has been accompanied by accelerating growth of salt-cored highs and folds flanking the depocenter. Additionally, core and log data indicate upward increases in both confinement and variability, reflected by a transition from primarily marginal to distal lobe elements in the deepest reservoir units (Ser 1), through those elements plus channelized lobe, distributary channel-axis, and channel margin elements in the middle reservoir units (Ser 2), to all-of-the-above plus levee-overbank elements in the upper reservoir units (Ser 3).
At the reservoir scale, compartments are defined by excess pressure differences of tens of psi’s. Measured pressures in wells indicate that pressure cell transitions occur over shorter distances vertically than they do horizontally, strongly suggesting a near horizontal orientation of pressure cell or compartment boundaries. Fault seal analyses demonstrate that none of the faults mapped in the field are long enough or have sufficient seal capacity to account for all of the observed segmentation. Plotting excess pressure data on maps and well sections indicate that faults likely compartmentalize only the smallest isolated lobe elements. Instead, stratigraphic architecture is far more likely to provide the observed, pancakelike, pattern of pressure cells at the reservoir scale, just as it does at the field scale.
The authors thank two anonymous GCSSEPM reviewers for their contributions to this manuscript. Anonymous reviews from the Knotty Head partners also improved the manuscript. We must make clear that all of the information in this paper is in the public domain, either already presented at conferences or made public by governing agencies. Several people within Hess Corporation contributed to this work.
Among them are Hartley Clay, Henry Zollinger, Gunardi Sulisto, Ryan Mann, Ryan Murphy, Jie Huang, and Rick Beaubouef. Despite this help, the interpretations, ideas, and conclusions as well as any logical, scientific, or factual errors are solely those of the authors and not necessarily of the mentioned contributors or of the Pony-Knotty Head partners.
Figures & Tables
New Understanding of the Petroleum Systems of Continental Margins of the World
- Atlantic Ocean
- chemically precipitated rocks
- clastic sediments
- deep-water environment
- geophysical methods
- geophysical profiles
- geophysical surveys
- Green Canyon
- Gulf of Mexico
- mass movements
- middle Miocene
- North Atlantic
- oil and gas fields
- overbank sediments
- reservoir rocks
- salt domes
- sedimentary rocks
- seismic methods
- seismic profiles
- stratigraphic traps
- structural traps
- turtle structures
- Pony-Knotty Head Field