Stratigraphy of Linked Intraslope Basins: Brazos–Trinity System Western Gulf of Mexico
- File PDFChapter PDF
Bradford E. Prather, Carlos Pirmez, Charles D. Winker, 2012. "Stratigraphy of Linked Intraslope Basins: Brazos–Trinity System Western Gulf of Mexico", Application of the Principles of Seismic Geomorphology to Continental Slope and Base-of-Slope Systems: Case Studies from SeaFloor and Near-Sea Floor Analogues, Bradford E. Prather, Mark E. Deptuck, David Mohrig, Berend Van Hoorn, Russell B. Wynn
Download citation file:
The Brazos–Trinity depositional system consists of four linked late Pleistocene intraslope basins (I–IV) located on the upper slope, offshore Texas, U.S.A. Conceptual understanding of the fill history in these basin include “fill and spill” models where basins fill sequentially in the seaward direction, to models which invoke coeval basin filling with the coarse fraction retained preferentially in the up-dip basins and synchronous early bypass of the fine fraction to down-dip basins. Integration of recent coring results with nearly complete 3D seismic coverage has improved age dating and reconstruction of infill history. Initially sediment gravity flows bypassed the upper basins, as they had not yet formed, depositing a basal sandy unit in Basin IV. Higher net/gross sands in Basin II deposited from mixed flows, with mud suspended high within the flows exiting Basin II through a tributary-like flow-gathering zone near the basin exit point. The muddy parts of these flows preferentially accumulated in the lower part of Basin IV, which was a three-dimensionally closed basin with deep ponded accommodation. In contrast, the upper fill in Basin IV comprises a submarine apron that is sourced by a continuous channel system that extends directly from a shelf-margin delta located in Basin I. Within this apron, the observed seaward tapering is controlled by lower-efficiency sandy sediment gravity flows of relatively small volume with respect to basin size. These observations allow us to distinguish perched aprons from ponded aprons, with direct implications for reservoir continuity. We further recognize that low-relief ponded aprons have lower ratios of sand net/gross than either high-relief ponded aprons or perched aprons.
The oil and gas industry has numerous discoveries from the late Tertiary Central Gulf of Mexico (GoM) slope apron (cf. Galloway, 1983) accounting for more than 3.2 billion barrels of oil and 8.1 Tcf gas. Many reservoirs from Tertiary slope systems, in general and in the GoM, in particular, are resolved as single wavelets on conventional 6–40 Hz seismic. Therefore, reservoir-architecture detail has to be inferred from borehole data in context with conceptual geologic models developed from understanding of subregional geologic setting, integrated with outcrop and production analogs. Adequately accounting for subseismic detail within static models therefore requires conceptual geological models that link slope depositional processes with reservoir architecture and are inherently uncertain.
Industry uses a combination of outcrop and near-seafloor analog studies to understand reservoir architecture in the context of geologic setting. Early work on continental slopes was primarily two dimensional—most outcrops are 2D with rare marginally 3D exposures depending on local topography, and shallow analogs in the Rudder/Magellan (Winker, 1993; Winker, 1996) and Einstein/Fuji (Hackbarth and Shew, 1994) areas were studied with high-resolution 2D seismic profiles, calibrated with logs and cores. For decades study of modern or near-modern seafloor systems focused on the basin floor (e.g., Normark, 1970; Piper, 1970; Coleman and Bouma, 1984; Flood and Damuth, 1987). With the advent of significant hydrocarbon discoveries in the deep waters of the Campos Basin and the GoM followed by west Africa and northwest Borneo, the search for modern and near-modern analogs for reservoirs at depth moved from the basin floor to continental slopes, where a proliferation of high-quality marine 3D seismic volumes exists. These data provide unique opportunities to integrate seismic stratigraphy to sea-floor gradient, accommodation, and in some cases lithologic calibration, in order to formulate depositional models useful for understanding controls on both presence and architecture of reservoirs at depth. This phase of work focused on both submarine valleys and their associated facies (e.g., Hackbarth and Shew, 1994; Deptuck et al., 2003; Posamentier, 2003; Deptuck et al., 2007; Heiniö and Davies, 2007), and the submarine aprons of the Brazos–Trinity system, offshore Texas, U.S.A. (e.g., Suter and Berryhill, 1985; Gardiner, 1986; Satterfield, 1988; Winker, 1996; Beaubouef and Friedmann, 2000).
Although the resolution of near-seafloor analogs imaged with conventional 3D seismic is of lower resolution (50–80 Hz) than these 2D HiRes surveys or outcrops, they provide three-dimensional information typically lacking from 2D seismic and outcrops. Conventional near-seafloor seismic typically images units and surfaces that are at or below the tuning thickness of their more deeply buried counterparts. These features locally relate to episodes of starvation, bypass, and/or erosion that control both reservoir bed length and connectivity. Outcrops show us that these surfaces are too subtle to be easily detected with wireline tools and too small to be confidently mapped with conventional 6–40 Hz seismic. Near-seafloor analogs provide dimensional data such as channel width, thickness, sinuosity, and areal extent of slumps useful in reservoir models (Prather et al., 2000; Steffens et al., 2004; Moscardelli et al., 2006; Wood and Mize-Spansky, 2009).
An additional advantage of studying analog systems from the late Quaternary is that the factors that control the sediment delivery and stratigraphic evolution of the margin are relatively well understood from independent lines of evidence. Key factors include glacioeustasy (Labeyrie et al., 1987; Shackleton, 1987), climate change and the resulting hydrology and sediment yield of rivers entering the Gulf Coast basin (Toomey, 1993; Blum and Price, 1994; Blum et al., 1994; Blum and Valastro, 1994; Aslan et al., 2006), and the morphology and oceanography of the receiving basin.
We focus this study on the stratigraphic evolution of basins II, III, and IV using a combination of conventional 3D and high-resolution 2D seismic profiles (Fig. 1), two shallow boreholes in Basin II, three IODP boreholes in Basin IV (IODP Expedition 308; Flemings et al., 2006), and published descriptions of giant piston cores taken from the seabed in Basin IV (Fig. 2); Mallarino et al., 2006). These data provide an unprecedented opportunity for the study of reservoir architectures typically resolved as single wavelets in deeply buried hydrocarbon-bearing systems imaged with conventional seismic.
The stratigraphic framework defined in this study, combined with absolute age dating of key surfaces (Pirmez et al., this volume), allows us to test multiple working hypotheses for the temporal and spatial stratigraphic evolution of deposits in a series of linked intraslope basins. The objectives of this paper are to: (1) define the seismic and lithofacies patterns in a well-imaged and calibrated linked intraslope basin system, (2) establish the subregional context for the wells that calibrate Basins II and IV, (3) better understand the reservoir architecture on a ponded above-grade slope system (sensuPrather, 2003), and (4) calibrate depo-sitional models for sand distribution and architecture in the Brazos–Trinity system, including ponded and perched submarine aprons (defined in later section).
Late Quaternary deposits of the Texas continental margin have been extensively studied by Berryhill et al. (1986), DuBar et al. (1991), Anderson et al. (2004), and Aslan et al. (2006). The Texas continental margin is an archetype for ponded above-grade slopes globally (sensuPrather, 2000). Ponded above-grade slopes are a class of continental slope that exhibits well-developed ponded accommodation as well as healed-slope and slope accommodation (Prather, 2003). Above-grade slopes in the classification scheme of Prather (2003) refers to a downslope seafloor profile that is elevated above the level of a theoretical concave-upward smoothed graded reference profile. Ponded accommodation lies within three-dimensionally closed topographic lows (doubly plunging synclines) on continental slopes. Healed-slope accommodation is the space across a step and ramp on the slope below a three-dimensional convex hull fit to the rugose seafloor topography (Prather, 2003; Steffens et al., 2003). Healed-slope accommodation is created by local subsidence and is the space left after filling of ponded accommodation.
Some authors (Pyles, 2008; Hubbard et al., 2010; Kane et al., 2010; Pyles et al., 2011) refer to above-grade slopes as out-of-grade. We feel that above-grade slopes should not be confused with the out-of-grade slopes as originally described by Hedberg (1970) (i.e., slopes where sediments bypass oversteepened continental margins into a base-of-slope position). Although local gradients on above-grade slopes are steep, particularly where ramps enter ponded intraslope basins, the overall gradient is lower than graded reference profiles and much lower than out-of-grade slopes.
Steffens et al. (2003) find that intraslope basins in the GoM exhibit abundant ponded accommodation throughout the central portion of the slope, with the largest areas located on the upper slope and becoming progressively smaller down slope. The intraslope basins are typically circular to elliptical in plan view with diameters up to 25–30 km. Ponded accommodation constitutes 55% of the total accommodation in the central slope of the GoM. Steffens et al. (2003) note that healed-slope accommodation is more pervasive in offshore Texas than in other parts of the GoM slope.
The Brazos–Trinity system (Fig. 3) is a point-sourced, linked series of intraslope basins, representing the latest phase of deposition within a large line-sourced constructional slope apron (the Tertiary central slope apron of Galloway, 1998). The Tertiary Central slope apron is built by frequently shifting sediment entry and exit points on a slope with complex “ponded” topography.
The Brazos–Trinity slope system is sourced by a lowstand confluence of the Brazos and Trinity Rivers to form a single shelf-margin delta. Confluence of these rivers is probably the consequence of a westward switch of the Mississippi River during a preceding lowstand (probably Stage 6; Fig. 3). This older Mississippi depocenter occupied much of the southwestern Louisiana shelf and extended into the Texas shelf (Berryhill et al., 1986; Coleman and Roberts, 1988; Winker, 1999); during the last falling stage the Calcasieu, Sabine-Neches, and Trinity were all deflected westward along the inner shelf to form a single river that ultimately joined the Brazos at lowstand. All of these rivers drain areas of exclusively sedimentary rocks (silici-clastics and carbonates) of Cenozoic, Cretaceous, and Permo-Triassic age (Fig. 3); the sediment load consists of polycyclic, fine-grained sand, silt, and clay. The Colorado River drainage fed the Colorado shelf-edge delta well to the west of the study area, and did not link directly with the Brazos shelf valley (Anderson et al., 2004). Rivers from the western Louisiana area built shelf-edge deltas east of the study area, and did not contribute to the BT system during the last sea-level cycle (Wellner et al., 2004; Anderson, 2005).
The receiving basin is microtidal; wave energy of the eastern Texas coast is intermediate between the low-energy, river-dominated coast of Mississippi and the moderately high-energy coast of southern Texas. It is not clear how much different these conditions were during lowstand conditions when the present-day shelf was emergent, and climate and oceanography were altered by lower sea-surface temperatures and influx of glacial meltwater from the Mississippi.
Configuration of Intraslope Basins
Basin I is the most proximal basin in the system. The center of the Basin I is located 17 km from the present-day shelf–slope break (Fig. 1). The upper reaches of the basin are occupied by a shelf-margin delta consisting of stacked delta-front clinoforms described by Suter and Berryhill (1985), interfingering with thick mass-transport deposits (MTDs) and upper-slope submarine apron deposits (Fig. 4). These deltas were likely fed from the ancestral Brazos–Trinity Rivers and represent the staging area for sediments entering the Brazos–Trinity intraslope system (Suter and Berryhill, 1985; Abdulah, 1995; Anderson et al., 1996; Winker, 1996). Growth faults and rotational slides are common along the shelf–slope break in this area. The rotational slides link downdip to mass-transport deposits characterized by chaotic seismic character (Fig. 4). The MTDs overlie slope aprons that onlap a thick interval of possible unconfined (hemipelagic) slope deposits (Winker, 1996). A surface channel network forming a tributary gather zone is observed at the southern exit point of the Basin I and links with the surface channel network in Basins II–IV (Winker, 1996; Beaubouef and Friedmann, 2000; Pirmez et al., 2000). Several erosional events identified at deeper levels in the onlap fill of Basin I suggests that the present surface configuration of channels may have existed in the past and sourced older deposits in the downdip basins (Winker, 1996).
Basin II, located 20 km downslope from Basin I (Gardiner, 1986; Satterfield, 1988; Satterfield and Behrens, 1990; Winker, 1996; Beaubouef et al., 1998; Badalini et al., 2000; Beaubouef and Friedmann, 2000; Beaubouef and Abreu, 2006), is located in water depths ranging between 600 and 993 m, asymmetric along strike with the eastern margin tilted by a major fault (Winker, 1996).
Basin III, located east of Basin II, is a small graben formed in response to local collapse above a salt-cored ridge that originally separated Basin II from Basin IV (Fig. 1). Basin III, located in 892 to 1267 m of water, captured the drainage from Basin II during formation of the eastern channel that feeds Basin IV (Winker, 1996; Badalini et al., 2000; Beaubouef and Friedmann, 2000).
Basin IV is the most distal of the four linked intraslope basins making up the Brazos–Trinity system (Gardiner, 1986; Satterfield, 1988; Satterfield and Behrens, 1990; Winker, 1996; Beaubouef et al., 1998; Badalini et al., 2000; Beaubouef and Friedmann, 2000; Beaubouef and Abreu, 2006). Basin IV has a water depth ranging between 1102 and 1479 m. The basin is underfilled, with at least 300 m of residual ponded accommodation left unfilled following the latest episode of deposition. Oversteepening of the basin margins (especially the eastern flank) appears to have initiated three locally sourced mass-transport deposits (Winker, 1996; Beaubouef and Abreu, 2006) whose rugose draped relief is still evident at the seafloor in the east-central part of the basin. Circular planform patterns corresponding to “wipe-out” zones in seismic sections suggest that mud volcanoes may also contribute to the rugose seafloor character in parts of Basin IV.
Seismic and Borehole Data
The seismic base data for this study comprises a dense (500 m spacing) grid of ultra-high-resolution 2D seismic profiles acquired by Shell in the 1990s together with continuous coverage with conventional 3D seismic data extending from the shelf to water depths in excess of 1500 m (Western-Geco surveys; Fig. 2). We focus this study on a combination of both the conventional 3D seismic, high-resolution 2D seismic, as well as abstracts and papers of near-seafloor studies using high-resolution 3D (20–750 Hz bandwidth, 200 Hz peak frequency; Beaubouef et al., 2003a; Beaubouef et al., 2003b; Beaubouef et al., 2003c; Beaubouef and Abreu, 2006). The high-resolution 2D seismic has a 100–600 Hz bandwidth, with peak frequency at ∼ 300 Hz yielding a vertical resolution of 1 m (Fig. 5). The conventional 3D seismic are binned with a spacing of 25 m × 12.5 m, with peak frequency of ∼ 60 Hz at -10 dB (Fig. 5).
The 3D seismic volume was processed to approximate a zero-phase wavelet. In addition to conventional reflection-coefficient three-dimensional seismic volume, seismic texture, which uses a combination semblance and amplitude, trace shape of the seabed reflector, and -90° phase-rotated volumes, were used for mapping purposes. The phase-rotated volume approximates trace integrated seismic data where the zero crossings of seismic traces correspond to stratal surfaces. Individual wavelets on these data represent the impedance of rocks averaged over approximately 15–30 m. This averaging allows identification of stratigraphic packages and their bounding surfaces, as well as acoustic impedance (this differs from typical reflection-coefficient data, where a trough or peak is associated with stratal surfaces of sufficient impedance contrast to produce reflections).
Two Shell boreholes with spot cores and continuous well logs in Basin II combined with well logs, continuous cores (Flemings et al., 2006), and giant piston cores in Basin IV (Mallarino et al., 2006) provide lithologic and stratigraphic calibration. Our dataset is further complemented by published results based on the study of a high-resolution 3D seismic survey in Basin IV (Beaubouef and Abreu, 2006).
Earlier work on the Brazos–Trinity system focused on description of seismic facies and stratigraphic evolution, leading eventually to a depositional concept model referred to as “fill and spill” (Winker, 1996). Suter and Berryhill (1985) first describe the “interdiapiric basins” of the Brazos–Trinity system, downdip of coeval shelf-margin deltas in offshore Texas. They publish details of the four intraslope basins and the connecting channel system. Winker (1996) describe the linked basins of the Brazos–Trinity system as structurally ponded fans within three salt-withdrawal intraslope basins (I, II, IV) and a graben (III). He notes that each basin filled with onlapping deposits consisting of alternating bedded and nonbedded units, and proposes that Basins I–III are filled to their topographic spill points in a progressive basinward direction (the fill-and-spill model).
Badalini et al. (2000) suggest that the various basins fill coevally, with the coarse fraction (sand) retained preferentially in the up-dip basins with the fine fraction (mud) spilling over the inboard basins sill, depositing sediment in downdip basins. Once channels connected the basins, the sand fraction was able to bypass updip areas, with part of the muddy fraction accumulating laterally on levees (Winker, 1996; Badalini et al., 2000; Beaubouef and Friedmann, 2000). Beaubouef and Friedmann, (2000) propose that the basins filled between isotope Stages 5 and 2 over a time span of 100 ky, but they presented no supporting age data. Since this time interval is represented by a “stepped” progression of higher-frequency sea-level rises and falls superimposed on a longer-term fall culminating in the Stage 2 lowstand, they conclude that the entire system represents a fourth-order sequence.
Mallarino et al. (2006), using long piston cores in Basin IV, demonstrate that: (1) high-resolution lithostratigraphy and chro-nostratigraphy can be established by coring the basin margins, (2) hemipelagic mud can be distinguished from gravity-deposited mud, (3) these cores can be tied to seismic units mapped on a very high-resolution 2D seismic grid in the basin, and (4) the timing of gravity-induced versus hemipelagic sediment accumulation can be linked to an independently established sea-level curve (Lambeck and Chappell, 2001) for the last glacial–interglacial cycle. Mallarino et al. (2006) conclude that the results of their study do not provide for any direct constraints for the correlation of these deposits with those of other linked basins in the system. Furthermore, they state that “High-resolution chronostratigraphic data from the other basins will be required to accurately reconstruct the timing of deposition within and between all four basins.”
Beaubouef and Abreu (2006) conclude that grain-size partitioning and/or flow stripping in upslope basins may have significantly affected the nature of turbidity currents reaching Basin IV. Alternatively, changes in depositional planform in Basin IV may have resulted from shallowing of the basin floor below the elevation of the inlet channel. Finally, changes in the nature of turbidity currents entering the basin through time may reflect an allogenic control related to evolution of the sediment-supply “conveyor belt” during the late Pleistocene sea-level fall.
Abreu et al. (2006) suggest that the fundamental depositional element in the Basin IV upper fan is a “scour lobe”, which is strongly erosional and confined updip, and lobate and somewhat erosional or weakly confined downdip. Abreu et al. (2006) believe that the scour-lobe model has strong implications for reservoir characterization of deepwater weakly confined and distributary systems. This model redefines mapping of environment of deposition in these systems, with implications for prediction of lateral and vertical connectivity as well as distribution of physical properties. Importantly, the scour lobe seems to be the preponderant reservoir architecture in producing fields with distributive distribution planform of reservoir (Abreu et al., 2006).
The stratigraphic framework used in this study builds on the work of Winker (1996), Badalini et al. (2000), Beaubouef and Friedmann (2000), and Mallarino et al. (2006), by incorporating the results from three IODP Expedition 308 wells described by Flemings et al. (2006) and Shell cores from Basin II, to date and correlate seismic horizons (Pirmez et al., this volume). Analysis of stratigraphic sequences evident on seismic line 3020, which passes through the IODP 308 wells, provides criteria for identification of key boundaries and corresponding seismic-horizon nomenclature as used in this study (Fig. 6). Line 3020 was chosen for this purpose because it images a complete record of slope stratigraphy calibrated by the IODP wells above hemipe-lagic deposits that form the bases for all basins in the Brazos– Trinity system.
Series boundaries are high-amplitude reflectors that converge into slope drapes on the flanks of the basin and against which overlying reflectors converge or base lap (Fig. 6). Series boundaries form the mapping horizons that we have correlated across the study area and into the various boreholes. The series boundaries identified from the high-res 2D are used to guide mapping of the conventional 3D survey, thereby extending the mapped horizons throughout the study areas.
Four seismic horizons, 10, 20, 30, and 80 (water-bottom reflector), correlate with a high degree of confidence among the Brazos– Trinity basins (Fig. 7). Horizons 10, 20, and 80 correlate with a high degree of confidence throughout the 3D seismic volume. Seismic horizon 30 occurs above an interval of thin-bedded silt-and sand-bearing clay, with local foraminifer sands, capped by 2.5 cm of foraminifer-bearing clay with volcanic glass shards cored by IODP at sites U1319 and U1320 (Flemings et al., 2006), Shell Ru# 1, and Marion Dufresne cores 37 and 40–44 (Mallarino et al., 2006). This interval provides an independent means of correlating a fourth high-confidence horizon (30) among the Brazos– Trinity basins.
Seismic Series 20 through 70 all converge into slope drapes at basin flanks (Fig. 8). Because of the lower resolution of the conventional 3D seismic data, correlation of individual seismic horizons through the slope drape from one basin to another cannot be done with confidence. Although the high-resolution 2D has sufficient bandwidth to resolve beds of about 1 m in thickness (Fig. 9), it does not have the areal coverage to allow physical correlation of the seismic horizons between basins (Fig. 2), but it is sufficient to allow “jump” correlation of marker beds between basins (Fig. 10). Radiocarbon dates measured from woody fragments recovered from IODP and Shell cores (Pirmez et al., this volume) provide an additional independent means of correlation between the basins for horizons 40, 50/60, and 70. Since dating sediments using 14C in this way provides only minimum absolute ages, correlation of these horizons, although better constrained, is still uncertain.
Within the basins all of the seismic horizons can be correlated with a high degree of confidence on the 2D grid of high-resolution seismic (Fig. 11, 12). Because of the lower resolution of the conventional seismic data, correlating individual seismic horizons within the basins can be done with more confidence than correlation between basins, but some uncertainties remain. To achieve the most accurate maps, seismic horizons correlated from the ultra-high resolution-2D grid were used to refine conventional 3D seismic correlations between basins.
Series 10 reflectors, located between the horizons 10 and 20, form a wedge that thickens from about 75 m in the middle of Basin IV to the north, towards the shelf edge (Fig. 8). Local thickness variations subordinate to the general updip thickening trend correspond to structural features such as faults, anticlines and synclines, and erosion at the bases of transparent-chaotic units and channels (Fig. 8). The reflectors are highly continuous and parallel, with reflectivity generally decreasing progressively upwards through the series (Fig. 11, 12).
All of the IODP and Shell coreholes penetrate Series 10. The U1319, U1320, and U1321 coreholes cut ∼ 75 m of a greenish gray and reddish brown clay (Flemings et al., 2006), and Ru#1 and #2 coreholes cut 30 m and 24 m of this interval respectively, but core was recovered only from Ru #1 (Fig. 7). The lithofacies is layered with distinct greenish gray and reddish brown clay beds ranging from ∼ 10 cm to several decimeters in thickness, and dark gray to black, locally pyritic layers, ranging from millimeters to centimeters in thickness. Burrows commonly disrupt the layering, imparting an overall mottled appearance to the unit. Smear slides show that sediments throughout the unit have a minor component of bioclastic origin, including CaCO3 fragments and foraminifers. Bioclastic sediments are locally enriched within burrow fills and thin laminae. These clays have a low total organic carbon (average 0.5%), but are rich in CaCO3 (average 23%), most of which is associated with finegrained detrital carbonate and dolomite. There is an ∼ 2-m-thick microfossil-bearing clay with intense bioturbation capping the unconfined deposits at site U1319. At site U1320 the same interval occurs within a zone of poor core recovery (Flemings et al., 2006).
Flemings et al. (2006) interpret this facies to represent deposition from muddy plumes or from a nepheloid layer. The 3D seismic reveals that thickness of this interval increases towards the shelf, and that the unit locally thins over the salt domes separating basins, suggesting that the unit formed by gravity-flow deposits. These flows were probably of very low density, akin to a nepheloid layer or weak gravity currents.
Although the Series 20 reflectors, located between horizons 20 and 30, correlate across all Brazos–Trinity basins with a high degree of confidence (Fig. 8), their character varies significantly from Basin II to Basin IV. Series 20 reflectors in Basin II form a very thin (∼ 3 m) set of high-amplitude continuous reflectors that drape the basin, except where truncated beneath channels at the proximal and distal ends of the basin (Fig. 11).
Series 20 in Basin IV consists of a similar set of highly continuous reflectors, except that those in Basin IV are thicker, 36 m at IODP site U1320, thinning to 15 m on the basin margins at sites U1321 and U1319 (Fig. 13A). The upper part of this unit drapes the basinward extension of the western channel in Basin IV (Fig. 14). Maximum Series 20 thickness is offset 2.4 km to the south-southeast relative to the present-day basin center (Fig. 15A). Horizon 20 also truncates beneath a transparent-chaotic facies, forming a cuspate thin localized in the eastern part of the basin (Fig. 13A).
IODP, Shell, and some of the Marion Dufresne holes cut Series 20 (Fig. 13A). IOPD logged an ∼ 8 m fine to very fine sand at site U1320 (Flemings et al., 2006). IODP U1319 and U1320, Shell Ru# 1 and Marion Dufresne 37 and 40-44 (Mallarino et al., 2006) cored thin-bedded silt- and sand-bearing clay, with local foraminifer sands, capped by several meters of bioturbated foraminifer-bearing clay containing a 2.5-cm-thick layer of volcanic glass shards. Smear slides show a clear dispersed carbonate component in the clay (Flemings et al., 2006). Point counts and petro-graphic observations indicate that the sands cored by U1320 are a fine- to very fine-grained lithic arkose (Fig. 15). This sand has lower quartz content than other sands sampled as part of this study. The heavy-mineral indexes from Series 20 sand (Fig. 16) are similar in composition to sands encountered today in Galveston Bay (Van Andel and Poole 1960), suggesting that during Series 20 deposition significant sand came from the Trinity River. Flemings et al. (2006) believe that thin-bedded silt- and sand-bearing clay records the first pulse of turbidity-current input into Basin IV.
Series 30 Basin II
Reflective continuous convergent thinning and baselapping seismic facies characterize Series 30 in Basin II (Fig. 11). Finer subdivisions of Series 30 are possible, especially in the lower part of the sequence, where a continuous seismic reflector separates a distinctive baselapping sequence from a convergent thinning sequence of reflectors (Fig. 11). For the sake of simplicity we lump these two intervals together because we see little lithologic difference in the Ru#1 cores that cut both sequences (Fig. 10). Series 30 in Basin II forms an isochore thick of about 40 m at the basin center, thinning to < 10 m on the basin margins, except for narrow thick extending to the northeast, tracking a channel incised into the underlying Series 10 (Figs. 10, 17A). Erosional truncation creates a toplap configuration as Series 30 near the distal edge of the basin where the western channel cuts down the slope towards Basin IV (Fig. 11). Reconstruction of the hemipelagic deposits down and across the western channel axis suggests that as much as 40 m of slope mud may have been removed during downcutting and formation of the channel.
Shell Ru#1 cut several cores through a low net/gross interval of interbedded sands and muds in Series 30 (Fig. 10). Sands in Series 30 are sublitharenites (Fig. 15) with heavy-mineral indexes (Fig. 16) indicating a stronger Brazos River affinity than sands from Series 20 in Basin IV, similar to the Brazos Province as mapped on the modern shelf by Van Andel and Poole (1960).
Reflective convergent baselapping facies that lap laterally onto transparent-chaotic seismic facies characterize Series 30 in Basin IV (Fig. 12). Series 30 thins towards the northern, western, and southern margins of Basin IV by a combination of thinning and baselap, similar to that observed in Basin II. This pattern of thinning is disrupted to the east, where the continuous reflectors of the converging unit onlap the transparent-chaotic facies. This transparent-chaotic facies truncates all of Series 20 and some of the Series 10 reflectors as described above. High-resolution 2D lines show that the transparent-chaotic facies truncates one to two reflectors near the base of Series 30 before aggrading sufficiently to onlap reflectors of Series 30 (Fig. 12). The isochore planform associated with the convergent facies show a remnant isochore thick located at the center of the basin outside of the part of the Series 30 associated with the transparent-chaotic facies (Fig. 13B).
The interval corresponding to transparent-chaotic facies consists of clay (∼ 90%) with only two sand beds, one at the base and one in the middle portion of the subunit. The key distinguishing characteristics of this unit are mud clasts of various colors, folding, steep erosional contacts, and steep apparent bedding dips (Flemings et al., 2006). “Chevron” or “wood-grain” patterns are indicative of coring-induced biscuiting of folded or steeply dipping sediment, as confirmed with borehole images (Flemings et al., 2006).
Average sand content of the Series 30 unit is ∼ 40% at sites U1320 and U1321, with only minor amounts of sand at the basin-margin site U1319. Cores recovered from the convergent baselapping facies have an average of 68% clay and 32% sand, with the sand arranged in centimeter to decimeter sand packages. The sands are very fine to fine grained, with normally graded tops and locally abundant plant fragments. Sands in Series 30 are litharenitic (Fig. 15) with heavy-mineral indexes (Fig. 16) indicating a Brazos River affinity, similar to the Series 30 sands in Basin II.
Low-amplitude transparent-chaotic and reflective moderately continuous seismic facies characterize Series 40 of Basin II (Fig. 11). The transparent-chaotic facies onlaps horizon 40 and is onlapped by reflective, moderately continuous seismic facies. Reflectors in the moderately continuous seismic facies truncate beneath transparent-chaotic facies in the overlying Series 50 and 60. Surfaces of erosion truncate Series 40 reflectors near the exit point of the basin (Fig. 11). These surfaces converge into the head of the western channel outboard of the saddle that separates Basin II from the slope that continues down to Basin IV. Isochore maps of Series 40 have a “distributary” planform which reaches its maximum thickness in the proximal part of the basin, offset updip from the basin center (Fig. 17B). A distinctive, “tributary” thickness planform is evident in the distal end of the basin where the truncation surfaces converge and join with the western channel (Fig. 17B).
Spot cores cut from the chaotic facies in the Shell Ru# 2 show it to be composed in part of contorted mud with thin sands. Shell Ru#1 cuts an ∼ 80% n/g sand interval in the reflective seismic facies. Three spot cores cut from Series 40 recovered subarkosic and sublitharenite sands (Fig. 17) with Brazos River heavy-mineral indexes similar to the sands in Series 30 (Fig. 18). The sand-prone reflective seismic facies forms a circular unit in the middle of the basin.
Reflective continuous convergent thinning and baselapping seismic facies characterize Series 40 in Basin IV (Fig. 12). Series 40 thins dramatically from 40 m in the center of Basin IV to its flanks, where it converges into the slope drape penetrated at the U1321 and U1319 sites (Fig. 10). Series 40 reflectors onlap to the north against topography created by the underlying Series 30 transparent-chaotic facies (Fig. 12). Excluding its thinning edges, Series 40 has a tabular isochore planform with its maximum thickness located in the center of the basin (Fig. 13C).
Series 40 is approximately 75% sand in the central part of the basin where penetrated by the U1320 site. The sand members correspond to reflective, highly continuous, parallel seismic fa-cies that extend to the limits of the basin before abruptly onlap-ping the basin flanks. The sand is very fine to fine grained with normally graded tops and abundant plant-debris sand and arranged into two members near the base of the series and is a mud cap. Sands in Series 40 are lithic arkoses (Fig. 15) with heavy-mineral indexes (Fig. 16) indicating a Brazos River affinity. The lower clay is greenish gray, and intensely bioturbated, producing black mottling, with rare thin silt laminae disrupted by burrows.
Series 50 and 60
Series 50 and 60, although resolvable on the high-resolution 2D seismic, are lumped together for this study because separately they are too thin (∼ 10 m maximum) to differentiate and map on the conventional 3D seismic. Series 50/60 in Basin II consist of a basal low-amplitude transparent-chaotic seismic facies similar to the lower of the two facies that make up the underlying Series 40 perched apron (Fig. 11). Reflectors in Series 50/60 truncate below erosional surfaces at the exit point of the basin, forming a “tributary” isochore planform, similar to the underlying Series 40 perched apron (Fig. 17C). 50/60 isochore maps have a “distributary” planform with a distinctive bifurcated pattern in the middle of the basin (Fig. 17C). There are no cores cut from this unit in Basin II, but logs show that it is muddy (Fig. 10).
Low-amplitude parallel seismic facies separated from transparent-chaotic unit by the subtle reflections of horizon 61 characterize Series 50/60 in Basin IV (Fig. 12). Series 50/60 have a tabular isochore planform except where it thins by onlap against the basin margin, similar to the underlying Series 40 (Fig. 13C). The transparent-chaotic seismic facies in the lower unit thickens into the western channel (Fig. 18A), whereas the upper unit of low-amplitude parallel seismic facies thins across the area where the lower unit thickens (Fig. 18B). IODP site U1320 cored a 17 m thickbedded sand interval overlying a 5-m-thick muddy interval with thin beds, corresponding to the upper low-amplitude parallel seismic facies and to the lower transparent facies, respectively. U1321, logged a 25 m sand of nearly 100% n/g corresponding to the low-amplitude parallel seismic facies. Horizon 61 separates these sands from a transparent-chaotic seismic unit of unknown composition to the north (Fig. 6). Horizon 61 also separates regions of variable trace shapes on the conventional 3D seismic where the interval resolves as a single seismic wavelet (Fig. 19).
At site U1320 an interval of contorted mud and thin beds of silt and sand correlates below the 25 m sand correlates with horizon 61 (Fig. 6). An ∼ 10-m-thick, organic-rich, homogeneous dark green to black clay with a sharp base and top caps Series 50/60. Clay at the top of the 25 m sand is 10.16 m thick and lacks any silt or sand laminae. Freshly split cores show ephemeral black mottling that disappeared a few hours after splitting (Flemings et al., 2006). Rare millimeter-scale silt-filled burrows and small wood fragments occur throughout the unit. Smear slides contain 5–10% dispersed organic matter, volcanic glass, and CaCO3 fragments (Flemings et al. 2006). Organic-rich black clay was also cut in Marion Dufresne cores 37 and 40–42
Isochore planform show that Series 70 in Basin II is very thin, ∼ 20 m, where penetrated by the Shell coreholes (Fig. 10). Series 70 reflectors converge on the flank of the basin by baselap. Series 70 is thickest along levees that flank the eastern channel, which connects Basins I to IV, through II and III (Fig. 17D). A seafloor image shows three channels near the Basin II exit point (Fig. 20). Channel 1 is linked to a small knickpoint on the footwall of the Basin III bounding fault. This channel extends updip to the Basin II entry point, where it is cut by the eastern channel (channel 3 in Fig. 20), forming a hanging valley. A second younger channel (channel 2 in Fig. 20) also leaves a hanging valley on the western flank of the eastern channel (channel 3 in Fig. 20) in Basin II. Hanging valleys along the eastern channel (Fig. 20) demonstrate clearly that channels 2 and 1 precede incision of the eastern channel into Basin II. Channels 2 and 1 do not crosscut each other, and thus we cannot estimate their relative timing.
The prominent through-going eastern channel (channel 3 in Fig. 20) has well-developed levees where it crosses Basin II (Fig. 17D). Both Shell coreholes in Basin II penetrate regions lateral to the leveed channel. Ru#1, the well closest to the channel, penetrates thin-bedded turbidites that are part of an ∼ 20 m fining-upward succession. Ru#2 is the farthest from the leveed channel. It cuts a thick ∼ 8 m sand confined to a shallow channel and is overlain by mud with a few thin sand members.
Series 70 in Basin III consists of a basal chaotic facies overlain by reflective continuous seismic facies of a leveed channel (Fig. 21). The Basin III seafloor image shows that Channel 1 links Basin II to a buried leveed channel in Basin III where it enters through the footwall knickpoint on the fault that bounds Basin III to the northwest. The prominent through-going eastern channel (channel 3 in Fig. 20) also has well-developed levees where it crosses Basins III (Fig. 20), as it has in Basin II.
Series 70 in Basin IV forms a prominent wedge-shaped unit that tapers into the basin center. The isochore planform of Series 70 is not concordant with the structural confines of the basin and thickens towards the eastern-channel entry point that links to Basin II through Basin III and buries the western channel (Fig. 13E). The Series 70 wedge in Basin IV tapers over ∼ 6 km into a more tabular interval in distal parts of the basin (Fig. 13E). The Series 70 apron is extensively channelized throughout except for smaller-scale distributary lobate forms located at the ends of channels and is capped by a similar network of channels and lobes (Beaubouef et al., 2003b, their Fig. 7).
At IODP site U1320 Series 70 is about 25 m thick with thick and medium beds of fine and very fine sand (Flemings et al., 2006). Series 70 thins to about 18 m at site U1321 and has a similar log character but slightly lower sand content than at site U1320 (Fig. 10). The correlative unit at site U1319 is only ∼ 3 m thick and is mostly mud with some thin beds of silt and sand. Sand beds in the Series 70 apron are organized in bed packets ranging in thickness between 2 and 8 m, capped by intervals of mud with thin beds. Mallarino et al. (2006) report that the Marion Dufresne cores cut medium to fine, well-sorted quartz-rich sand and gray mud. IODP cores cut from this series recovered lithic–arkosic sands (Fig. 15) with Brazos River heavy-mineral indexes similar to the sands in the younger than Series 20 (Fig. 16).
In the tapered submarine apron of the Series 70 apron there are at least seven depositional subunits that are seismically mappable, consisting of smaller-scale channel–lobe complexes (Figs. 6, 19). These channel–lobe complexes backstep progressively during aggradation of the unit and are present at the seabed (Fig. 13F). Seismic profiles show that the channelized part of the lobe complex consists of discontinuous reflections and mounded external form. The lobes are thicker and more mounded nearer the basin entry point and show clear compen-sational offset stacking down dip in the basin center (Beaubouef et al. 2003b, their Fig. 9).
Foraminifer-bearing greenish gray clay caps Series 70 in the U1319 and U1320 coreholes. This unit is characterized by relatively high values of spectrophotometric lightness. Flemings et al. (2006) interpret this facies to be the Holocene drape, as did Badalini et al (2000) and Beaubouef and Friedmann (2000).
The Brazos–Trinity system is an archetype locality for understanding depositional processes on ponded above-grade slopes (sensuPrather, 2000) and provides the only known example where detailed three-dimensional reservoir architecture can be studied in a cascading series of linked intraslope basins where both the timing of deposits in adjacent basins and their subre-gional geologic setting are well constrained.
In this section we use the terms “perched apron” and “ponded apron” with a further subdivision of the latter into two varieties: “high-relief” and “low-relief” ponded aprons, to describe most of the deposits in this collection of intraslope basins (Fig. 22). This is a modified version of the terminology of Beaubouef and Friedmann (2000) by use of the terms high-relief and low-relief ponded aprons for deposits in ponded accommodation, and perched apron for deposits in healed-slope accommodation (Fig. 22).
Gorsline and Emery (1959), Galloway and Hobday (1996), and Galloway (1998) introduce the term “apron” for all line-sourced slope systems. In geomorphology, an apron is any laterally extensive deposit lying at the base or in front of its source. On continental slopes sediment may be fed relatively uniformly over the shelf margin to form a line-sourced slope apron. The depositional product is a strike-elongate prism of slope sediment. Two variations of line sourcing occur. In constructional aprons (see Figure 8C from Galloway, 1998), sediment spills onto the slope relatively uniformly along the shelf margin or is locally funneled through notches or erosional channels that shift frequently, so that over geologic time a line source is effectively created.
Slope aprons can be very large. The entire slope accumulation in the late Tertiary Central GoM is considered a slope apron by Galloway et al., (2000). The evolution of slope topography in the Central Gulf Apron, however, creates a complex temporal and spatial distribution of both ponded and healed-slope deposits (Prather et al., 1998; Booth et al., 2000; Prather, 2000; e.g., Booth et al., 2003), similar to most other ponded and stepped above-grade slope systems, whether in the GoM, Nigeria, NW Borneo, or Lower Congo basins. We therefore subdivide the slope-apron gross depositional environment as used by Galloway (1989) and Galloway (1998) into smaller-scale stratigraphic units, which can be described as seismic loop sets. We use bounding-surface type, external geometry of surface-bounded seismic facies, event geometry internal to bounding surfaces, seismic reflectivity, and event continuity to define a seismic loop set, or seismic facies unit.
Submarine aprons are associated with convergent-baselapping seismic facies (cf. Prather et al., 1998). Convergent-baselapping seismic facies display external convergence of bounding surfaces and well-defined internal baselap. Internal reflection continuity ranges from extremely continuous to moderately discontinuous (Prather et al., 1998). Submarine aprons in our scheme include fan aprons (O’Byrne et al., 1999), ponded-basin and healed-slope deposits (Prather et al., 1998; Prather et al., this volume), perched-slope fills (Beaubouef and Friedmann, 2000), healed-slope aprons (Booth et al., 2002), frontal splays (Posamentier and Kolla, 2003), and transient fans (Adeogba et al., 2005).
Perched aprons (Fig. 22) represent deposits that accumulate in healed-slope accommodation (Prather, 2000). Such deposits have reflectors that converge by baselap and thinning, isochore planform with maximum thickness offset updip from the basin center, and presence of a bypass channel, knickpoint(s), and/or a “gather zone” at the basin exit point (Fig. 22). In contrast, ponded aprons (Fig. 23) represent deposits that accumulated in ponded accommodation (Prather, 2000), confined by elevated basin margins on all sides. We further recognize two types of ponded aprons: (1) “low-relief” ponded aprons and (2) “high-relief” ponded aprons (Fig. 13). Low-relief ponded aprons have reflectors that converge by baselap and thinning, and tabular isochore planform with maximum thickness located in the center of the basin (Fig. 17A). High-relief ponded aprons have reflectors that converge by baselap and thinning, and isochore planform with maximum thickness offset toward the entry-point channel, but unlike perched aprons they have no evidence of bypass channels, knickpoint(s), or a “gather zone” since there is no basin exit point (Fig. 13E).
Gather zones are surfaces of erosion with tributary planform located at the downdip edge of a perched apron (Fig. 17B). Gather zones form as turbidity flows converged across the apron top into the gather zone where they erode into the apron. Erosion takes place here as turbidity flows thicken and accelerate into the gather zone. A knickpoint forms where turbidity flows enter the top of a ramp that separates steps on a slope profile. A bypass channel forms on the ramp and conveys turbidity flows farther down the slope profile. A leveed bypass channel forms where the knickpoint migrates updip across the top of the apron and into the entry-point channel that feeds the apron, thereby capturing and bypassing all subsequent turbidity flows across the apron.
We are able to build a high-resolution correlation framework using carbon-isotope, oxygen-isotope, tephrostratigraphy, and heavy-mineral stratigraphy and more conventional seismic stratigraphy to differentiate among different seismic-facies-based interpretations of the Brazos–Trinity basin fill architecture of previous authors. It now seems that during the stepwise sea-level fall between MIS 5e and 2, Basin IV received up to 175 m of sediment-gravity-flow deposits, comprising turbidites, slumps, and debris flows (Pirmez et al., this volume).
We interpret this interval to represent hemipelagites that are part of an unconfined slope deposit resulting from deposition of distal turbidity currents overspilling from basins next to Basin IV (laterally and/or updip), possibly with a significant contribution of sedimentation from surface plumes of coastal rivers in the form of mudbelts (Fig. 22). Subtle changes in thickness around present-day anticlines and synclines (Fig. 23A) suggest that sediment gravity flows contribute to unconfined slope deposition, and the slope, although unconfined, probably had at least a stepped profile (Fig. 24A). Mud with composition rich in CaCO3 in the form of fine-grained detrital carbonate and dolomite and devoid of sand suggests a provenance from the Mississippi River. It is unclear whether these unconfined slope deposits contain a record of mudbelt deposition on the upper slope, spillover from Basins II and I updip, or whether they represent distal turbidity currents from adjacent basins. Based on the mud composition, rich in fine detrital dolomite, the shelf-edge deltas of the Brazos–Trinity drainage system were probably located to the west of Basin I, with the Mississippi river debouching to the east of Basin I on the upper Texas shelf.
This episode of deposition continued through the formation of a condensed interval represented by horizon 20 (MIS 5e after Pirmez et al., this volume). During this time ponded accommodation formed in Basin IV and a step formed where Basin II would eventually form (see next section).
We classify Series 20 in Basin IV as a low-relief ponded apron because the high-amplitude continuous reflectors converge by baselap and thinning (Fig. 22), and produce a tabular isochore planform with maximum thickness coincident with the center of the basin and coeval Basin II deposit as a slope drape (Fig. 23B). The oldest turbidite deposits in the linked basins occur in Basin IV above horizon 20 (MIS 5d-c after Pirmez et al., this volume). Low quartz content in these lithic arkoses suggests an affinity with sands from the Mississippi province (Hsu, 1960). Seismic line 3008 demonstrates that erosion and sediment entry occurred through the western channel during Series 20 deposition, inasmuch as the erosional surface is draped by horizon 20 (Fig. 14). The absence of any noticeable erosion or onlapping deposits in Series 20 of Basin II indicates that turbidity currents must have bypassed the Basin II area without leaving a significant trace of their passage (Fig. 23B), suggesting that Basin II formed later (Fig. 24B), probably during deposition of the Series 20 condensed interval, which contains the Y-8 ash bed (MIS 5a to ∼ MIS 3 after Pirmez et al., this volume).
There is 360 m of structural relief from the center of Basin IV to the spill point located in a saddle along the eastern basin-bounding salt ridge. The slope drape associated with the sands in Series 20 extends 7 km west and 14 km north of the U1321 location, but the sand is not found in any of the Marion Dufresne cores (Mallarino et al., 2006, their Fig. 7). The extent of the Series 20 drape, the presence of sand in IODP coreholes, and the expectation that sand can be suspended no higher than about 60 m in a typical muddy turbidity current (Alessando Cantelli, personal communication, 2011) imply the basin was much shallower during deposition of the lower part of Series 20 and that the current configuration of Series 20 results from hundreds of meters of subsidence following deposition (Fig. 24B). Furthermore, assuming that the Series 20 sand in Basin IV is a sheet, its absence in the piston cores (Mallarino et al., 2006, their Fig. 7) suggests the basin has also tilted to the northeast producing a 2.4 km isochore offset relative to the present-day basin center during this phase of subsidence (Fig. 13A). These relationships indicate that salt withdrawal plays a significant role in the evolution of the system. It is possible that this tilting is coincident with emplacement of mass-transport deposits sourced from the eastern flank of the basin.
Onlap geometries around the basin and an isochore thick coincident in the deepest part of Basin II suggest gravity-flow deposition in ponded accommodation (Beaubouef and Friedmann, 2000). We classify Series 30 in both basins II and IV as low-relief ponded aprons because (1) the high-amplitude continuous reflectors converge by baselap and thinning, and (2) it has tabular isochore planform with maximum thickness located in the center of the basin (Fig. 22). In Basin IV the apron consists of a basal mass-transport deposit (MTD) overlain by mud with thin turbidites. The narrow isochore thick in Basin II that tracks the channel incised into underlying hemipelagites of Series 10 suggests a sediment entry point to the northeast with a sediment exit point to the southwest through the western channel (Fig. 23C).
The flat-lying, onlapping units in Basin IV are locally eroded and interbedded with wedges of transparent seismic units interpreted as mass-transport complexes (Beaubouef et al., 2003b). Unlike the ponded apron in Basin II, these units show no evidence of erosion associated with basin entry or exit channels. Beaubouef et al. (2003b) describe this unit as part of a sequence of confined sheet complexes that are restricted and roughly concordant with the structural confines to the deepest part of the basin. Uniform seismic amplitude and high reflector continuity suggest inter-bedded sheets of sand and mud (Beaubouef et al. 2003b). Beaubouef et al. (2003b) observe no channels in these units, and stacking patterns appear aggradational.
The absence of entry channels makes it difficult to know how turbidity flows entered Basin IV (Fig. 24C), but it seems likely that they spilled out Basin II, entering Basin IV through the western channel. It is also possible that this unit was sourced entirely from within the basin as a submarine slide coming off the eastern flank of the basin, where there are some prominent slide scars (Fig. 23C). The turbidites in this scenario would be gravity flows associated with the disintegration and transformation of nonturbulent slope failures into turbulent flows. The presence of sands in the convergent baselapping unit, however, requires the presence of sand in the areas from which the MTD are sourced for this scenario to be correct
We classify Series 40 in Basin II as a perched apron because the reflectors converge by baselap and thinning, isochore maximum thickness is offset toward the entry-point channel, and there is a “gather zone” at the basin exit point (Fig. 22). Contorted mud in the low-amplitude transparent-chaotic facies suggests the presence of a MTD with overlying turbidites represented by the reflective moderately continuous seismic facies. The same interval in Basin IV is a low-relief ponded apron. In Basin IV the high-amplitude continuous reflectors suggest that each sand member may be sourced from large single sediment gravity flows, large enough to flow across the entire basin.
The tributary planform at the downdip edge of the perched apron suggest that turbidity flows converged across the apron top into a “gather zone” where they eroded into the apron toe to form a series of knickpoints at the edge of the step (Fig. 23D). Acceleration of these flows as they exit the basin along the western channel caused headward migration of the knickpoints truncating apron deposits and forming the tributary planform of the “gather zone” (Fig. 24D).
High-resolution 2D seismic lines in Basin IV show Series 40 to have a laterally extensive pattern, with no apparent channels. Beaubouef and Abreu (2006) report no channels in this unit, and stacking patterns appear largely aggradational on their 3D seismic data. Thickening into the western channel suggests that sediment flows that bypassed Basin II entered Basin IV through the western channel (Fig. 23D). Alternatively, more subtle thickening towards the eastern channel suggests either sediment entry exclusively or some additional contribution to Series 40 from the eastern channel (Fig. 23D). The geometries might also represent simple side lap on the edge of Basin IV combined with thinning over basin-flank slumps.
Similar to the underlying perched apron in Basin II, we classify Series 50/60 in Basin II as a perched apron (Fig. 22). The bifurcating isochore planform in the middle of the apron suggest deposition around topography on an underlying Series 40 MTD located outboard of basin entry point (Fig. 17C). Similarly to the underlying perched apron in Basin II, this interval also has the distinctive tributary planform in the distal part of the basin of a perched apron (Fig. 23E). Reflectors in the perched apron are truncated beneath a gather zone at the head the western channel (Fig. 11). Ru#2 cores show that the chaotic seismic facies consists only of contorted mud (Fig. 11). Ru#1 penetration has little sand. Its thickness, the presence of a gather zone, and little to no sand suggest that sediment gravity flows largely bypassed Basin II during this episode of deposition, exiting the basin through the western channel (Fig. 24E).
Isochore planform of Series 60 in Basin IV suggests that sediments of this interval are derived from turbidity currents entering the basin through the western channel (Fig. 23E). In Basin IV we classify Series 50/60 below horizon 61 as a high-relief ponded apron because reflectors converge by baselap and thinning, isochore maximum thickness offsets toward the entry point channel, and there is no evidence of bypass channel, knickpoint(s), or a “gather zone” (Fig. 22). Transparent-chaotic seismic character suggests that the interval below horizon 61 is lithologically homogeneous, either mostly sand or mostly mud. If this apron is sandy the sands must pinch out into mud before reaching the U1320 site. Alternatively this unit could be a muddy MTD also carried into the basin through the western channel, although its isochore planform suggests that this is unlikely, and the expected impedance contrast in the two units should produce a higher-amplitude horizon-61 reflector.
The isochore planform of the interval above horizon 61 containing the 25 m sand indicates that it was likely derived from turbidity currents passing through the eastern channel, skirting around underlying high-relief ponded apron — thickening against the “backstop” at the end of the basin (Fig. 18B). Previous basin models derived from seismic facies analyses interpreted this transparent unit as muddy mass-transport deposits (Winker, 1996; Badalini et al., 2000; Beaubouef and Friedmann, 2000). Badalini et al. (2000) believe that the transparent seismic facies is correlative across basins II, III, and IV, possibly recording an isochronous event related to catastrophic failure of the delta front in Basin I, and this could still be the case. If the Series-61 represents the first deposits laid down from turbidity currents entering Basin IV through the eastern channel, then it may be related to development of the leveed bypass channels I and II in Basin III (Fig. 21). The wide lateral extent, tabular geometry (Fig. 23E), and homogeneous nature of the organic-rich black mud that caps the 25 m sand suggests deposition from a muddy suspension, possibly from a large turbidity current trapped or ponded in the basin (cf. Lamb et al., 2005).
Like the immediately underlying units, we classify Series 70 in basins II and III as perched aprons because reflectors converge by baselap and thinning, isochore maximum thickness offsets toward the entry-point channel, and a bypass channel and knickpoint exist at the basin exit point (Fig. 13).
The history of bypass in Basin II is linked to the evolution of the small fault-bounded Basin III, which became active before or during deposition of Series 70. Hanging valleys along the eastern channel in Basin II suggest the existence of two early bypass channels (bypass channels I and II; Fig. 21). The sequence of deposition interpreted from reconstructing the history of bypass from Basin II and Basin III suggests that this high-relief ponded apron in Basin IV is sourced directly from the lowstand delta front in Basin I (Winker, 1996) through the eastern channel (Fig. 23F). Winker (1996) shows that the eastern channel formed from a knickpoint that retreated from Basin IV, across Basins III and II.
In Basin IV we classify Series 70 as a high-relief ponded apron because reflectors converge by baselap and thinning, and isochore maximum thickness offsets toward the entry-point channel, but there is no evidence of bypass channel(s), knickpoint(s), or a “gather zone” (Fig. 13). Beaubouef et al. (2003b) suggest that the geometry of the high-relief ponded apron in Basin IV may be related to progressive shallowing of the basin floor and associated loss of confinement. The updip and eastern shift of the depocenter associated with a prominent thick near the eastern channel suggests that the eastern channel was the exclusive conduit for sediment flows passing downdip across Basin III from Basin II during this phase of deposition (Fig. 23F).
Figure 3 in Beaubouef et al. (2003c) shows high-resolution 3D images of channelized lobes from within the Series 70 apron. Abreu et al., (2006) call these channelized lobes “scour lobes”, and suggest that they are the fundamental depositional element in the Series 70 apron. The proximal equivalents of these lobe units are highly discontinuous and eroded lobes near the basin entry point. Updip, the surface at the base of the “scour lobe” has a deep and narrow scour. Downdip, the scour broadens and shallows. Abreu et al. (2006) estimate that the scour lobes are 1 to 5 kilometers long, hundreds of meters wide, and up to 25 meters deep in the scoured portion, and 1 to 3 kilometers wide and about 10 meters deep in the lobate portion. Farther upstream of the proximal part of the scour lobe, the depositional surface is conformable (a surface of bypass). The updip extent of the scours is interpreted to result from knickpoint migration in the upstream direction (headward erosion). Deposition evolves from prograding and then backstepping packages probably associated with flow avulsion, when a portion of the flow is diverted towards another location at the margin of the apron (region of maximum gradient), resulting in progressive abandonment of the previous sediment fairway (Fig. 24F).
Understanding near-sea-floor analogs can have important implications for predicting presence and connectivity of sand in analogous subsurface reservoirs. To describe the depositional history of these basins in order to relate them to the subsurface we introduce the terms ponded submarine apron for deposits in ponded accommodation and perched submarine apron for deposits in healed-slope accommodation. We further subdivide ponded aprons into “high-relief” and “low-relief” varieties. Submarine aprons in the Brazos–Trinity basins are composed of amalgamated channels in proximal parts of the aprons and compensationally stacked lobes in distal parts, with scattered mass-transport deposits. Through-going leveed bypass channel(s) or submarine valley formed from an upslope-migrating knickpoint further differentiate perched aprons from ponded aprons.
Sand content in this collection of basin appears to have a strong dependence on internal controls such as the relative relief of basins along the flow pathway. Well penetrations of the low-relief ponded aprons and observations from high-resolution 3D seismic by Beaubouef and Abreu (2006) suggest that sands are associated with lower net/gross sheet complexes. The low aspect ratio of these low-relief ponded aprons suggests deposition from muddy highly efficient flows (?) capable of ponding within the basin. These muddy flows could originate directly off the fronts of lowstand deltas feeding the apron system or represent finegrained flow-stripped deposits from flows whose coarse load was captured in up-dip accommodation.
High-relief ponded aprons in contrast tend to have higher sand content, particularly in the proximal portions, where intense channelization and scouring probably imparts a relatively high connectivity to the various sand units. Since there are no significant through-going bypass channels, abandonment of high-relief ponded aprons has to result from either relative sea-level rise or avulsion of the shelf-to-slope delivery system. Progressive abandonment of the apron can result in deposition a series of small compensating and either prograding or retrograding lobes across the top of the apron as smaller flows with limited run-out distance enter the basin. Like low-relief aprons, high-relief aprons reflect net accumulation outboard of bypass channels and/or valleys. Since the high-relief apron lies entirely in ponded accommodation, the observed seaward tapering has to be controlled by limited-volume, lower-efficiency sandy sediment gravity flows with run-out lengths shorter than (or confined by) the dip length of the basin. Relatively high net/gross distributary-channel complexes should be expected near the basin entry point with smaller-scale channels and distributary lobate forms distally within the basin.
The through-going leveed bypass channel(s) or submarine valley that characterize perched aprons have the potential, if mud filled, to compartmentalize reservoirs. Perched aprons consist in part of progradation sets of distributary channel and lobe complexes (also see Prather et al., this volume). Distributary-channel complexes, high sand content, and backfilling characteristics suggest that the fill in perched aprons may be very similar to that in high-relief ponded aprons. The fine-scale architecture of perched aprons may differ significantly in the propensity for the preservation of channel-base drapes, since perched aprons, by virtue of being perched above either filled ponded basins or across stepped profiles, may comprise higher proportions of bypassing sediment gravity flows. This will add risk for further stratigraphic compartmentalization in addition to knickpoint erosion.
Sedimentation in the Brazos–Trinity Basin IV is the result of a complex interaction between fluvial–deltaic dynamics, sea-level changes, and interactions between turbidity currents and submarine topography. Our high-resolution correlation framework using 14C and 18O stratigraphy supported by sand provenance studies shows that the bulk of the basin-fill sequences co-evolve, at least within the time resolution provided by 14C age dates, in the fashion described by Badalini et al. (2000), rather than pure “fill and spill” models of Winker (1996) and Beaubouef and Friedmann (2000). The exception to this occurs where the updip basins are clearly in bypass mode, such as during deposition of Series 20 and 70. Although we reached the conclusion of co-evolution in the basin, our data show that this process occurred in one-tenth the time proposed in any of the previous models (Pirmez et al., this volume). The high-frequency stratigraphic transitions as observed from Series 30 to 80 within a short time of only 12 ky (MIS 2 after Pirmez et al., this volume) indicate that autogenic processes, such as build-and-collapse cycles in the shelf-margin delta, climate variation, etc., should be considered in addition to short-term allogenic processes such high-frequency sea-level forcing as a control shaping the architecture of slope deposits. Additional studies of the updip shelf-margin systems are needed to help further understand the timing and nature of progradation and sediment transport through the shelf-margin delta in Basin I.
The Series 70 apron in Basin IV is illustrative of high-relief ponded aprons. The combination of high-resolution 2D cross sections and the high-resolution 3D horizon slice from this apron represents the best image of what would normally be subseismic stratigraphic architecture on more conventional seismic data. These images reinforce the idea that GoM single-seismic-cycle reservoirs should be dominated by amalgamated channels near intraslope-basin entry points and that these channels should grade downdip into lobes. Series 20 and 30 in basins II and IV contains examples of low-relief ponded aprons. Basin II has examples of perched aprons with thickest deposits shifted toward an entry-point break in slope in the absence of three-dimensional basin confinement, and both early-stage bypass represented by the knickpoint front, gather zone, and knickpoints in the Series 40–60 perched aprons as well as late-stage bypass represented by the leveed-channel connected Series 70 perched aprons in Basin II to the high-relief ponded apron in Basin IV. The abandonment facies assemblage typical of some perched aprons (see Prather et al., this volume) is the only component of typical perched-apron depositional processes and stratigraphic architecture not represented well.
We gratefully acknowledge Shell International E&P for permission to publish this paper. This research used samples and data from Expedition 308 provided by the Integrated Ocean Drilling Program (IODP). We also acknowledge the helpful contributions from our Shell colleagues. Special thanks go to Ciaran O’Byrne and Zoltan Sylvester, who helped us with maps and cross sections, but we thank them especially for the spirited exchange of ideas. We thank Maria Banks for contributing her analysis and interpretation of sand composition and provenance, and Andy Morton for heavy-mineral analysis. The quality of the manuscript was improved by reviews from Mark Deptuck, Dave Jennette, Gianluca Badalini, Mike Mahaffie, and Tom Wilson.