Abstract We cannot hope to predict Mesozoic depositional processes and sediment properties well enough to plan effective regional exploration strategies without considering the big picture of Gulf of Mexico deposystem evolution. The two critical big picture elements are the kinematics and timing of the Yucatan Block's detachment and separation from North America and the various major expansions and contractions and the ultimate disappearance of the Western Interior Seaway. Although a number of authors, including this one, have speculated on the timing of separation of Yucatan from North America ( Fillon, 2007a ), no definitive evidence exists: i.e. , drilled samples of the ocean crust and the sediments directly overlying it. Without that unambiguous information we must infer the paleogeographic evolution of the early Gulf of Mexico Basin from deposystem architecture by asking questions such as when do Gulf of Mexico deposystems transition from architectures consistent with deposition in a youthful blockfaulted basin underlain by actively attenuating continental crust to deposition in a mature basin having stable margins surrounding a central region underlain by subsiding ocean crust. An understanding of the paleogeography and paleoceanography of the Gulf of Mexico Basin derived from deposystem architecture can help provide answers to crustal kinematic questions and to more exploration focused questions such as: where, and in section of what age should we look to find facies similar to the organic rich, generative Haynesville Shale facies of eastern Texas and western Louisiana. Although we all know something about the Western Interior Seaway, most of us working on the Mesozoic of the Gulf of Mexico Basin have not spent much time considering what effects it might have had on the prospectivity of Gulf of Mexico deposystems. Through much of Albian and Late Cretaceous time the Western Interior Seaway connected the Gulf of Mexico Basin with the Arctic Ocean Basin. The effects of the establishment and intermittent blocking of this major seaway connecting arctic and tropical water masses on global paleoceanography, on global paleoenvironments, and locally on onshore and offshore Gulf Basin deposystems cannot be ignored in our quest to understand the Mesozoic of the Gulf Rim. This paper is a “big picture” review of Gulf of Mexico Basin deposystem evolution within the Late Jurassic (Oxfordian)–Late Cretaceous (Maastrichtian) interval. Seventeen Mesozoic chronosequences are defined therein based on chronostratigraphic data garnered from over 130,000 industry well and pseudowell penetrations of Mesozoic section in the Gulf of Mexico Basin region. Examination of the collected data suggests that grouping the seventeen Gulf of Mexico Mesozoic chronosequences into seven super-chronose-quences optimally distinguishes key phases of deposystem and basin evolution. The oldest super-chronosequence defined in this study, dubbed “MG,” encompasses ca. 16.45 Ma of Norphlet through lowermost Cotton Valley Late Jurassic deposition. Sediment distribution and accumulation rates within the MG interval clearly define the rectilinear configuration of the earliest Gulf of Mexico Basin. This early basin geometry is consistent with fault controlled attenuation and foundering of North American continental crust, associated flooding, and rapid depositional infill concurrent with the earliest detachment of the Mayan (Yucatan) crustal block from North America. The Yucatan block, although showing an affinity with South American (Amazonian) terranes ( Martens, 2009 ), was left attached to the North American plate when North America began pulling away from Gondwana during the initial breakup of Pangea ( Fillon, 2007a ). The next younger super-chronosequence, “MF,” contains a. ca. 13.47 Ma record of Cotton Valley, Bossier, Knowles limestone., Late Tithonian through mid Hauterivian, deposition. The “MF” interval reflects the same rectilinear outline as the “MG,” but is marked by decreased accumulation rates, suggesting that the fault bounded crustal attenuation, rapid sediment infill phase had markedly slowed. The ca. 9.4 Ma of Hosston, Sligo, Sunniland limestone, James limestone, mid-Hauterivian through Early Aptian section contained within the succeeding “ME” super-chronosequence records modification of the early rectilinear basin outline by a temporary reactivation of attenuation and foundering in the western portion of the Gulf of Mexico Basin. “ME” sediment distribution patterns also indicate development of a depositional continental margin and accumulation of true continental margin type deltaic and reef systems. These observations suggest that during this interval a deep continental basin, probably floored by ocean crust, was beginning to form outboard of the attenuated continental crust. Sediment distribution and accumulation rates within the ca. 23.5 Ma Rodessa through lower Washita, Early Aptian through Early Cenomanian “MD” super-chronosequence reflect growth of the Wisconsin interior seaway and a stable phase of relatively low accumulation rates throughout the entire Gulf of Mexico Basin deposystem. During this interval, deposition was very likely influenced by a vigorous tidal and thermohaline current circulation driven by strong temperature contrasts within the Gulf of Mexico–Wisconsin interior seaway–Arctic Ocean connection. The next younger super-chronosequence, “MC,” contains a ca. ca. 16.0 Ma record of Dantzler, Washita, Lower Pine Key, Eutaw, Woodbine, Eagle Ford, Austin, and Early Cenomanian through Late Santonian (Late Cretaceous) deposition. During this phase, there is a marked reduction of accumulation rates in the north-western portion of the basin, attributable perhaps to expansion of the Western interior seaway and continued subsidence of the old Gulf of Mexico Basin margin. Associated small, perhaps tidal submarine delta-like depopods developed, perhaps in response to the regional Western interior seaway transgression ( Blakey, 2014 ). These delta-like depocenters appear to define a new basin margin presaging the modern curved shape of western Gulf of Mexico so familiar to us today. Here also we see the first unambiguous evidence of abyssal deposition in the deepest portion of the Gulf of Mexico Basin underlain by ocean crust. The succeeding ca. 12.82 Ma interval of Late Santonian through Early Maastrichtian upper Pine Key, upper Selma, upper Austin, Taylor, Olmos, Saratoga, and low accumulation rate mainly chalk and marl deposition contained within the “MB” super-chronosequence provides evidence of transgressive onlap associated with an expanding and deepening interior seaway during “MB” time. “MB” onlap has the effect of temporarily reemphasizing structural trends inherited from crustal attenuation that took place during “ME” time. Finally, the ca. 5.4 Ma long terminal Mesozoic “MA” super-chronosequence consists of Maastrichtian, Navarro equivalent, low accumulation rate marls deposited along the basin margin. These low accumulation rate basin rim sediments and low accumulation rate slope sediments are punctuated by high accumulation rate canyon fill and lobe-shaped slope depopods which are probably attributable to sediment reworking, transport and deposition by transitional Cretaceous-Paleogene (K/P) interval mega-tsunami backwash flows immediately following the Chicxulub impact. Higher accumulation rates in the deeper parts of the basin underlain by ocean crust are also consistent with high volume backwash flows.

Abstract Major normal and growth faults are known to extend from sea floor through to base of the sediment wedge, their origins generally occurring along the then shelf break/uppermost slope and with overall less contemporary tectonics farther landward. The loci of extensional tectonics proceeds basinward, as the entire sediment wedge migrates offshore. Lesser sediment depocenters are successively incorporated due to sea level oscillations. The wedge-transiting faults appear to terminate often into plastic salt accumulations. Semiplastic unconsolidated clays, whose deposition are dominated by electromagnetic forces (ionic bonding), can create breakage/weakness zones along which extruded fluids from dewatering can migrate. Thus, the sea floor expression of significant faults can range from well-defined fault breaks to varying concentrations/domains of clay-sized particles. Granting continental margin extension from rifting while a new ocean basin deepens, normal faults may occur within subsiding crust. Given synchroneity of extension and subsidence in sediments and crust, breakage zones in both might coincide. Upper crust is brittle fracturing. Lower crust temperatures and pressures suggest semi-plasticity with shear dislocations between separate masses. Lying between the upper and lower crust is transitional crust, possibly associated with fluid-injection along brittle fracture zones. As measured by earthquake seismology, crust maintains constant densities of 3.3 and 2.7 for oceanic and continental crust, respectively, and a transition zone between; such density transition indicates the Airy-Pratt controversy is unresolved. The mantle, being plastic and heterogeneous, contains convection cells having lateral extents ranging from 10-100 km to basin-spanning. Fluids from ocean and mantle could find avenues to transit from one to another.

Shelf-margin deltas and linked downslope depositional systems are in most cases fed by alluvial valleys that serve to deliver sediment eroded from the hinterland. Accordingly, alluvial valleys provide the link between processes that control sediment flux to the continental margin and processes that control dispersal into the basin. Current research shows the volume of sediment delivered to the margin will reflect hinterland drainage areas and large-scale relief. Superimposed on this background rate will be an unsteadiness that reflects climate change in hinterland source regions, but the rates and directions of change in sediment supply will vary regionally. Alluvial valleys modulate unsteadiness in sediment supply through changes in sediment storage. However, regional variability in the rates and directions of change in sediment supply insures that responses to climate change are regionally circumscribed, and alluvial valley systems in different regions may respond in opposite ways to the same global climate change. Sea-level change has little effect on the total volume of sediment delivered to the margin, but instead forces channel extension and shortening, which plays a major role in determining the proximal to distal location of the river mouth point source through which sediment is dispersed to the shelf and beyond. Moreover, the widely used concept of incision and complete sediment bypass within incised valley systems during periods of relative sea-level fall should be abandoned. Instead, falling stage to lowstand fluvial deposition is actually common in well-studied Quaternary analog systems, and falling stage sand bodies may comprise a significant proportion of reservoir-quality sands within many incised valley fill depositional sequences. Models for falling stage and lowstand systems tracts should therefore incorporate significant fluvial channel belt deposits that are likely connected to, and feeding, the offlapping shore faces, shelf-margin deltas, and linked downslope systems.

Abstract Along the shelf margin of the northern Gulf of Mexico, numerous late Quaternary deltaic systems occur where ancestral rivers encountered the shelf-slope break. These shelf-margin deltas are products of deposition during glacioeustatic fluctuations resulting from expansion and contraction of continental ice sheets. Lowered sea level shifts paralic environments seaward and creates widespread subaerial unconformities, well-defined drainage networks (incised valleys), and deltaic systems that prograded to the shelf margin. Shelf margin deltas are primary mechanisms for shelf margin and upper slope progradation, and serve as important conduits of sediment to deeper water environments.

Abstract An ultra-high resolution, short-offset 3D seismic survey (EBHR3D) has been used to study the sedimentary fill of an intra-slope basin in the East Breaks area of the Gulf of Mexico. The site chosen for the seismic program is the fourth and southernmost basin (Basin 4) in the Brazos-Trinity Slope System. The Brazos-Trinity Slope System is a set of latest Pleistocene salt-withdrawal basins that are connected by channels in the upper to middle portion of the Texas continental slope. They are filled with sediment delivered to the slope by the ancestral Brazos and Trinity rivers and associated shelf edge deltas. Together, the linked shelf and slope depositional systems form a late Pleistocene lowstand systems tract. The seismic survey has been designed to target a large submarine fan at the top of the basin-filling succession (the Upper Fan), but imaging of the entire 250 m (maximum) of basin fill is excellent. The results are providing detailed information regarding deep water deposition far surpassing what is possible from outcrop or conventional subsurface studies. The data provide unprecedented images of the three-dimensional geometry and internal architecture of these deepwater deposits. The fill of Basin 4 records a stratigraphic evolution that includes a “ponded” fill stage followed by a “perched” fill stage. Contrasting deposit geometry and stacking patterns occur during these two stages of evolution. The perched fill of the basin contains the Upper Fan, which is located in the shallowest portion of the subsurface beneath an extensive Holocene drape. The Upper Fan represents the terminal, distributive complex of the lowstand system tract. It is a basinward-tapering wedge of sediment that contains both channel-form and sheet-like depositional elements. The prominent stratigraphic features interpreted from the Upper Fan are: (1) off-lapping, clinoform reflection patterns; (2) distributary channel systems linked to channel mouth lobes; (3) down-fan progression in architecture from channel-form elements to more sheet-like elements; and (4) down- and across-fan decrease in sand percent and/or grain size inferred from seismic attributes. In these and other ways, the stratigraphy of the Upper Fan is similar to that commonly observed for modern and ancient river deltas.

Abstract During the previous glacial-eustatic fall, the ancestral Brazos and western Louisiana rivers, which flowed across low gradient coastal plains and shelves, constructed large fluvial-dominated deltas that extend to the shelf margin. These rivers shifted to new locations prior to the lowstand, resulting in shelf-margin deltas that have no associated down-dip lowstand deltas or fans. The Trinity and Colorado rivers remained fixed in their locations throughout the eustatic fall and lowstand, resulting in linked valley/delta/fan complexes. Re-incision of lowstand valleys by these rivers over several eustatic cycles resulted in significant sediment bypass to the slope. Factors that influenced the response of rivers to falling sea level include long-term sediment supply, diapiric controls on channel location, and the physiography of the shelf over which the rivers flowed.

Abstract Lithologic, biostratigraphic and isotopic data from cuttings provides calibration of three sigmoidal, clinoform packages separated by regionally continuous, parallel seismic facies. This depositional geometry is interpreted as a shelf-margin, deltaic wedge deposited within the East Breaks 160-161 minibasin. A chaotic seismic facies package extending over more than 84 mi 2 (218 km 2 ) occurs between two clinothem packages of Ericson Zone Y (71 ka BP to 12 ka BP). The chaotic package is interpreted to be a mass-transport complex that failed during the accelerated rate of sea level fall during late Oxygen Isotope Stage 3 (approximately 30 to 20 Ka). Clinothem foresets, bottom sets and the mass-transport complex are predominantly clay with minor siltstone. Sands are restricted to proximal topsets and fluvial channels. The mass-transport complex consists of three subfacies: rotated-block, hummocky-mounded, and disrupted. Distribution and volume of these facies suggests that only the rotated-block subfacies has been significantly transported, and that the hummocky-mounded and disrupted subfacies result from different degrees of disruption of ponded, clay-prone layered sediments by compression and dewatering triggered by the submarine slide of rotated blocks.

Abstract Linked shelf edge deltas and slope channel systems are observed in the eastern Gulf of Mexico. The slope channels are characterized by deep incision into the substrate and moderate sinuosity nearly to the shelf-slope break. Channelized flows were not fully confined as evidenced by well-developed levees up to 90 m thick. This sinuosity suggests that turbulent flow within the channel was likely nearly from the uppermost slope. With apparent turbulence characterizing these channels nearly to the shelf-slope break, the dominant mode of sediment delivery to the slope and basin beyond probably was in the form of density underflow ( i.e. , hyperpycnal flow) rather than shelf edge slump and/or slide progressively transformed into turbidity flow. The stages of evolution of these slope channels are (1) clustering of small slope gullies on the slope at the initiation of lowstand deposition, (2) dominance of one of these slope gullies and formation of one significant channel, formation of a frontal splay fed by the dominant channel, (3) abandonment of frontal splay deposition in favor of leveed channel deposition across the entire slope, and (4) entrenchment of the leveed channel into the earlier deposited leveed channel and frontal splay.

Unravelling end-Cretaceous paleobathymetric dip-profiles along strike in the northern Gulf of Mexico continental margin (nGoM) is important because it provides us with the initial framework in which to assess the evolution of Cenozoic clastic systems and burial history. Light can be shed on the problem by integrating seismic and potential field data with analytical basin modelling techniques and palinspastic-kinematic paleo-geographic analysis. Progressive palinspastic reconstruction of the Gulf is critical to setting up the appropriate lithospheric models to assess subsidence history. Kinematically, the opening of the GoM involved an early rift stage of northwest–southeast stretching (with minor counterclockwise [CCW] rotation) between North America and Yucatán (Triassic–Early Oxfordian), followed by a drift stage of seafloor spreading between the opposing rifted margins that entailed significant CCW rotation of Yucatán Block. This Stage 2 rotation was accommodated by transform motion of Yucatan/Chiapas Massif along the foot of the very narrow eastern Mexican margin, but transform motion stopped once the Central Gulf Spreading Ridge passed any point along this margin. Thus, the crustal boundary along eastern Mexico is ultimately an igneous, constructional contact that is overlain by the entire stratigraphy visible on seismic, and no transform faults are to be expected ( Pindell, 1985 ). The rift stage was asymmetric ( Pindell et al. , 1986 ; Marton and Buffler, 1994) such that Yucatán collapsed off the mainly northwest-vergent Alleghenian Orogen of the southern USA. The nGoM margin (foot-wall) underwent large and rapid tectonic subsidence during the rift stage (because the crust was highly stretched), but little thermal subsidence during the drift stage and thereafter (because the lithosphere was NOT stretched much). In contrast, the Yucatán hanging wall underwent little syn-rift subsidence (because the crust was not stretched much), followed by considerable postrift (Late Jurassic and younger) thermal subsidence (because the lithosphere was stretched). During the rift stage, thick red beds, possibly with lacustrine or even marine intervals, effectively buried basement in most nGoM areas and, toward the end of the rift stage (Callovian–Early Oxfordian), gave way to salt deposition across much of the half-open Gulf basin. Original salt thickness is generally considered to exceed what could have been achieved by thermal subsidence alone (~2km) during Callovian–Early Oxfordian time (presumed agespan of salt deposition); thus, salt accumulation was coeval with Stage 1 tectonic subsidence (syn-rift) and/or involved the marine inundation of pre-existing, isolated, sub-sea level accommodation space, which, by Oxfordian time, was almost definitely filled to sea level with salt. Together, red beds and salt probably are 5–10km thick beneath much of the nGoM rifted margin. Oxfordian onset of CCW rotational seafloor spreading in the Gulf split the pre-existing red-bed/salt basin into the Louann and Campeche halves. Backstripping shows that ocean crust was generated near its typical 2.6km depth below sea level, and not at an Icelandic-setting near sea level. The continent-ocean boundary typically is marked by a large step up from continental to oceanic crust ( i.e. , the rifted continental crust was already buried by red beds and salt far thicker than 2.6km ocean-generation depth, so the basement step to ocean crust is UP). As spreading ensued, a central, widening “chasm” was produced that was floored by oceanic crust and that, once Smackover open marine conditions were established, received no new depositional salt. To our knowledge, truly autochthonous salt cannot be shown to overlie definite oceanic crust; thus, initial spreading was effectively coeval with the onset of Smackover open marine conditions, and there may have been a causal relationship between the onset of spreading and the breaking of the evaporitic sill, wherever that was (Florida Straits or Veracruz Basin are equally viable guesses). As seen south of the Middle Grounds margin, the immediately adjacent shoulders of these salt walls halo-kinetically collapsed into the widening chasm, but, given the enormous width of the nGoM rifted margin, an important question is to assess how far north into the salt basin such early collapse occurred. The Red Sea analogue shows that the salt could have supported shelf platforms at least into the Cretaceous; we typically observe minor (<20km) extrusion of salt across the step up onto oceanic crust, but locally, such as at Sigsbee Escarpment, salt may have collapsed much farther (100km) onto the ocean crust, possibly as early as Late Jurassic–Cretaceous times. In such places, the term “parautochthonous salt” applies, because the salt still underlies most stratigraphy although it acquires a tapered-wedge cross-sectional geometry as it collapses. Because the nGoM margin was the footwall during Jurassic asymmetric rifting, thermal subsidence had far less influence on paleobathymetry than is commonly believed. Thus, determination of paleobathymetry can be roughly gauged by structural analysis of halokinesis. Thus, for large areas of the nGoM margin, we propose (1) that a relatively shallow, “supra-salt platform” persisted until the Late Paleocene onset of the well-known Wilcox growth faulting, and (2) that the Upper Jurassic–Cretaceous supra-platform section remained shallow, and was never deeply buried until halokinetic collapse began. This contrasts sharply with the Campeche Salt margin of Mexico, which was drowned to truly basinal depths in the Late Jurassic-Early Cretaceous due to far higher rates of post-rift thermal subsidence and weak clastic sediment supply. Thus, in the north but not in the south, we envision a very broad, relatively shallow supra-salt platform with a thinner-than-often-assumed Upper Jurassic-Cretaceous section that extended well beyond much of today's coastline. In this case, the true continental slope and rise would have been located much farther out than the Stuart City carbonate trend (which is often inferred as the paleo-shelf edge). This platform may have been stepped due to early halokinesis, particularly at Sigsbee, and along its outer reaches probably sloped or ramped down to the area of oceanic crust. Given this scenario for the paleobathymetry, it should not be surprising to find early Paleogene sands at the foot of that platform slope ( e.g. , Perdido area). The sands could have been transported there from (1) the north or northwest by shelf bypass across the suprasalt platform, or (2) the west, out of a proto-Rio Grande river system, or both. By the end of the Paleocene, salt collapse in updip areas of the supra-salt platform began due to differential burial by prograding clastics, producing syn-depositional counter-regional faults and basin-facing half-grabens at the Wilcox and younger fault trends, which controlled the [new, syntectonic] position of the paleo-shelf edge. Such collapse fed downslope shortening behind (landward of) the Paleogene sands at the foot of the true continental slope. We infer detachment on salt, such that the Mesozoic marine supra-salt section was cut both updip and downdip by at least some of the faults. Apparent rafting of the Mesozoic shelf section at the landward limit of the Wilcox trend ( e.g. , Anderson and Fiduk, 2003 , and also observed in NE Mexico on seismic by the authors) demonstrates that the salt (and inferred Upper Jurassic–Cretaceous shallow shelf) was mobile during end-K/early T time. The concept of the Mesozoic supra-salt platform in the nGoM: (1) requires significant changes to commonly-accepted Late Jurassic through Paleocene paleobathymetric and paleogeographic maps of nGoM , and therefore of reservoir and source rock distribution; (2) indicates the need to develop maturation models for the inner shelf areas that do not assume a pre-existing deep-water setting outboard of the Sligo/Stuart City “reef trends” prior to Tertiary clastic deposition; and (3) provides a new paleogeographic context for assessments and models of Cenozoic deltaic and progradational depositional systems along the northern Gulf of Mexico.

Abstract The upper Miocene (late middle to early late Miocene) depositional episode (UM depisode) records a long-lived family of sediment dispersal systems that persisted for nearly 6 Ma with little modification. In the central Gulf of Mexico basin, this depisode records extensive margin offlap, primarily centered on the paleo-Tennessee River and Mississippi River dispersal axes, that began immediately following the Textularia W/Textularia stapperi flooding and is terminated by a regional flooding event associated with the Robulus E biostratigraphic top. Thickest sediments are deposited in the paleo-Tennessee River delta beneath modern southeast Louisiana, where three major depocenters are recognized. These depocenters have migrated in both strike and dip directions, and margin progradation is very prominent. The composite fluvial-dominated paleo-Tennessee and Mississippi delta system rapidly built beyond the subjacent middle Miocene shelf margin to construct a sandy delta-fed apron. Margin outbuilding was locally and briefly interrupted by hyper-subsidence due to salt withdrawal and consequent slope mass wasting. Sediments also continuously bypassed into the Mississippi Canyon, Atwater Valley and Green Canyon OCS areas throughout the entire upper Miocene, forming two secondary depocenters composing the McAVLU submarine fan system at the base of the paleo-continental slope. A broad, but relatively thin, sandy strandplain and clastic shelf succession, supplied by reworking of the deltaic deposits, extended eastward and westward from the delta system. Abundant strike-reworked sediment locally prograded the strand plain to the shelf edge, and slope offlap exceeds 30 mi (50 km). The presence of extremely large volumes of high-quality shelf margin delta and deep-water fan sandstone reservoirs results in the great productivity of the central Gulf of Mexico upper Miocene, and upper Miocene production is dominated by a major deltaic oil and gas trend straddling the southeast Louisiana coast.

Deltas that cross the shelf, either by prograding at a high stand or due to sea level drop during a low stand, produce distinctive depocenters which are important exploration targets at the shelf edge. The deltas become unstable as they try to prograde into deep water and deposit units that are geologically unique and very different from their counterparts that prograde across a stable shelf. These unstable deltas produce a variety of contemporaneous (early) structures including growth and glide plane faults numerous associated faults, diapirs, and gravity slides. The growth faults result in a greatly expanded reservoir section. These downdip delta systems are uncoupled from their updip feeder systems by these major growth faults and are typically encased in highstand deep water shales thus becoming excellent exploration targets: an over-pressured section, early structures, expanded reservoirs, and good seals. Virtually all of the largest plays (the multiple TCF type) for the onshore and shelf parts of the Gulf of Mexico during the last 30 years have been in these shelf-margin delta systems. Examples from the Tuscaloosa, Wilcox, Yegua, Vicksburg, Frio, and offshore Miocene are examples. These plays are not unique to the Gulf Basin but will occur in any basin where deltas reach the shelf margin and prograde into deep water. The successes here in the Gulf Basin are simply the beginning of an ongoing worldwide exploration effort in these types of deposits. They provide both useful analogues and important exploration guidelines.

Abstract Shelf-edge deltas are the main driver for the delivery of sediment to the deep water lowstand systems tracts. However, the mere presence of deltas at the shelf margin does not guarantee accumulation of deep-water sands. The two main reasons for this are: (1) deltas that develop at the shelf edge during relative sea-level fall generally need to be significantly incised by their own distributaries for sand delivery to be focused down to a basin-floor fan system, and (2) deltas that develop when sea level is rising (late lowstand) tend to be inefficient sand-delivery systems, and disperse sand mainly onto the slope as sheet-like turbidite lobes, with few or no basin-floor fans. Thus, given the presence of deltas at the shelf-edge, both the likely magnitude and direction of sea-level change at the shelf edge needs to be estimated, before significant time-equivalent, deep-water sand can be predicted on the basin floor. Shelf-edge deltas are generally thicker, significantly more unstable, and markedly more turbiditeprone than inner/or mid-shelf deltas. These major differences are due to longer run-out slopes (greater water depths), steeper mud-prone slopes, and greater accommodation at the shelf margin compared to deltas in more proximal shelf settings. There are four main types of shelf-edge deltas that have been documented from a database developed mainly from the Eocene shelf margin on Spitsbergen and the Miocene shelf margin of the Carpathian Foredeep: Type A deltas develop on the outer shelf/shelf-margin transition but without significant progradation beyond the shelf edge onto the slope. These deltas usually form during the falling stage of a fall-to-rise cycle on the shelf. Type B deltas develop at the shelf margin but are significantly cannibalized by fluvial-feeder erosion. Such deltas also form during falling stage, but base level falls below the shelf edge. The deltas are fairly sharp based on the outer shelf, are sand prone, and are deeply eroded by their own river distributaries. Because of the fluvial incision, only remnants of these deltas are preserved. However, their main significance and legacy is their time-equivalent, downslope suite of deep water, lowstand deposits including basin-floor fans. Type C deltas develop at the shelf edge, produce significant basinward growth of the shelf margin but rarely link down to basin-floor fans. They form during a late, rising stage of the fall-to-rise cycle, as they overlie earlier cannibalized deltas and older basin-floor fans of the same sequence. They are many tens of meters thick and consist of stacked, well-developed upward-coarsening and thickening units.. Type D deltas are progradational to aggradational delta complexes at the shelf margin, without underlying shelf-edge erosion, and only rare, linked basin-floor fans. Type A and C deltas simply amalgamate during a fall-to-rise cycle to become a single, thick (many tens of meters) deltaic wedge that is perched at the shelf margin and drapes far out onto the slope.

Sandy shelf-margin clinoforms in the Eocene strata of the Central Tertiary Basin of Spitsbergen are usually generated by river-dominated shelf deltas, or by wave-dominated shorelines, though these two regimes can also be strike-equivalent to each other. Clinoforms occur in series or sets that show both sub-horizontal and rising trajectories of shelf-edge accretion. Clinothems involved in the former style of margin growth, however, tend to be dominated by delta deposits. Shelf-edge deltas of such clinoforms are commonly severely eroded by their own distributary channels, and this is especially noticeable at (though not restricted to) shelf-edge locations. Fluvially incised shelf edges are commonly linked directly across the shelf break, to turbidite-filled channels, gullies and small canyons on the slope. Examples of this type of shelf-edge situation are present on Brogniartfjellet in Van Keulenfjorden, where the outer-shelf segment of the clinothem contains shelf-edge deltaic units that are 20-30m thick deposited during falling base level and lowstands. The deltas have been cut by deep erosive channels (up to 12 m) paved by shale rip-up conglomerates. The channel infill is dominated by up to 3 m-thick, flat and low-angle laminated, medium-grained sandstone bedsets deposited from upper-flow-regime conditions in riverine and shallow sand flats. Multiple phases of erosion can be demonstrated, separated by phases of minor re-establishment of delta-front facies. At peak regression of the delta system, still during falling relative sea level, the channels have reached the shelf break and allowed the river system to feed sediment directly into slope channels that were turbidity-current conduits to the basin floor. These are incised more than 25m deep on the upper slope, appear to have originated from fluvial input and retrogressive slumping on the slope, and link back up to the shelf-edge incisions. The infill of the slope conduits strongly suggests repeated phases of erosion/bypass that alternated with phases of low-efficiency, hyperpycnal-flow deposition. The apparent off-lapping architecture within the slope conduits strongly suggests oblique or downslope accretion of infill during continued relative fall (forced regressive and lowstand conditions) of sea level, and probably during basin-floor growth of the fan. In the latest stage of the lowstand, the shelf-edge deltas have re-established themselves onto the shelf, aggrading and prograding onto the underlying canyonized succession, thus forming a lowstand prograding wedge. Minor fluvial incision occurs, but overall the system is less sand prone. During the subsequent transgression of the shelf, when sea rose back up to and above the shelf edge, the slope is blanketed by mud, there is tidal re-working and infilling of the older shelf-edge channels and a transgressive barrier/lagoon or estuary system migrated landwards.

Africa's major rivers are associated with some of the largest deep-water petroleum systems in the world. The most successful of these in terms of hydrocarbon discoveries, the Congo, Niger and Nile, are currently sites of intense exploration activity, while others (such as the Kwanza, Zambezi, and Rufiji-Ruvuma) have yet to yield a significant number of finds and remain under explored. The theme of this paper is to review examples of these systems, attempt to show why this has occurred and where some of the future offshore hydrocarbon provinces of Africa will be. Seismic data will be used to review the deep water plays and prospectivity associated with each major river system. The deep-water systems reviewed include the Congo, Niger, Nile, Kwanza, Zambezi, and Rufiji-Ruvuma. The more mature deep-water hydrocarbon provinces will be compared with the less explored areas. Suggestions as to the reasons for limited exploration of some areas will be made, concluding that visible structuring is as least as important as the perceived presence of source rock in the initial decision to explore in a basin. Visible structuring in the form of salt or mud diapirs, anticlines, etc., undoubtedly makes exploration “easier” in choosing locations and “selling” prospects in the early stages of basin exploration. Seeking stratigraphic traps when there are few or no structures is very difficult in an unexplored basin. The future exploration potential of the deep water provinces will be reviewed with particular reference to the currently less explored or less mature areas to show their untested prospectivity. As exploration has moved into deeper water in known hydrocarbon provinces such as Niger and Congo deltas, then the targets have been the deeper water analogies of existing plays; i.e., Niger delta mud supported anticlines and Congo fan channelized sands. In areas in which there are few or no existing working plays, extension by analogy into deep water breaks down or does not exist. So often in these areas new plays and source rock postulations have to be developed. Examples of new untested plays that do not exist in shallower water are shown from Tanzania Rufiji-Ruvuma and Mozambique deep water systems where there has been no drilling. Many of these involve structuring that does not exist in the shallower explored areas and are beginning to attract interest. In areas such the Kwanza Basin where there is a very structured sedimentary section the (correct) perception that the play type is not the same as farther north in the Congo Fan has caused the area to be downgraded, despite the fact that hydrocarbons have been found here. This downgrading is partially due to recent wells targeting the same play as farther north and failing to find oil. In this case, the primary reservoir targets are not highly visible Tertiary channel sands but Cretaceous sands having a more subtle seismic expression, a play that has proven to work by the Semba-1. The potential of the ultra deep water (i.e., water depths >3km) and abyssal areas beyond the Congo and Niger fans will be examined to demonstrate that as yet untested plays and prospectivity exist in these future exploration provinces.

Abstract Pleistocene shelf-margin deltas (SMDs) of the northern Gulf of Mexico (Mobile, Mississippi, Brazos-Trinity, Colorado, Rio Grande) and the corresponding shelf-slope transition illustrate some of the conceptual and procedural issues pertaining to sequence stratigraphy of continental margins. As one approaches the shelf margin from the landward side, it is customary to pick the sequence boundary (SB) at the erosional base of channelized fluvial deposits, typically cut into marine clinoforms. Near the shelf margin, the sequence boundary could be picked at the erosional base of a submarine canyon cut into a SMD, or alternatively at the base of the SMD, especially if no submarine canyon is present. Thus, the SMD can be placed either below the SB (making the SMD part of the HST) or above the SB (making the SMD part of the LST). In practice, the latter is seldom done, because there is rarely a distinctive surface or break in stratal geometry to uniquely mark the change from shelf-phase delta to SMD. Therefore, SMDs are usually considered part of the HST. In some cases the SMD is characterized by submarine landslide deposits within the clinoforms, resulting in hummocky or chaotic clinoforms, which can grade downdip into massively chaotic, sand-bearing deposits on the upper slope. Even in these cases, the change from shelf-phase delta to SMD is typically gradational, without a single distinctive surface to uniquely define the SB or to place the SMD in the LST. In general, there is no consistent rule as to where the SB should occur relative to a SMD, or where the SMD should fit into the systems tract classification. As one approaches the shelf margin from the basinward side, where mini-basins are present on the upper slope, the SB is typically picked at a major onlap surface, which in late Pleistocene deposits can often be correlated across saddles between mini-basins with little ambiguity. In the Brazos-Trinity deposystem, the SB defined by this onlap surface is clearly different from (and stratigraphically below) the SB defined by the erosional base of fluvial deposits landward of the shelf margin. In the Mississippi deposystem, SBs defined by such onlap surfaces are also clearly different from SBs defined by erosional bases of submarine canyons. Onlap surfaces (and immediately underlying MFS shales) are useful for correlation along strike, especially on the upper slope. In contrast, submarine canyon surfaces are useful for long-distance correlation in the dip direction from shelf to basin plain, but are of very limited extent in the strike direction. The basin-floor fan (BFF) phase of slope deposition typically occurs just above the onlap surface, whereas the slope fan (SF) phase occurs at and above the submarine canyons. A composite framework of onlap surfaces and submarine canyons is useful for establishing temporal relationships within the Mississippi depositional province, although this framework does not fit readily within standard systems tract nomenclature. In concept, sequence boundaries are isochronous surfaces which separate deposits that are less closely genetically related while grouping deposits that are more closely related. Two difficulties are recognized with this concept. First, sequence boundaries picked at erosional surfaces are subject to regional diachroneity, such that some fluvial deposits above the SB may be coeval with some marine deltaic deposits farther downdip below the same SB. Secondly, the SB typically groups slope deposits with immediately younger transgressive deposits while separating them from immediately older deltaic deposits. However, in map view, Pleistocene deposystems of the northern Gulf of Mexico consistently show a close paleogeographic relationship between slope systems and immediately older deltaic systems. Conversely, the paleogeographic relationship between slope systems and immediately younger transgressive and highstand systems is typically much more distant. From studying a variety of Quaternary deposystems associated with Gulf Coast rivers, we recognize a composite succession, although not all phases will necessarily be present in any one deposystem: Incised valleys are filled by a combination of estuarine transgression, fluvial aggradation, and deltaic progradation. Once the incised valley is filled, the alluvial plain continues to aggrade without the lateral confinement of valley walls. This aggradation may or may not be accompanied by deltaic progradation. The delta progrades across the shelf, probably punctuated by minor transgressions and lobe switching. The net progradation is typically forced by sea-level fall, but may also occur purely by sedimentary progradation, as in the Holocene Mississippi delta. Forced regression of the delta is typically accompanied by valley incision farther updip. As the delta approaches the shelf margin, the deltaic depocenter becomes thicker and smaller in areal extent, while the prodelta becomes steeper and increasingly prone to slope failure. Slope failures may be manifested in a variety of ways, such as a single slide complex which is healed by subsequent clinoform progradation, or as repeated slides during progradation, resulting in chaotic clinoforms. Also during this phase, turbidity currents may be generated at or near river mouths, which generate sinuous slope channels without significant incision of the shelf margin. Alternatively, the SMD can remain gravitationally stable, with minimal generation of sediment gravity flows. The SMD is incised by a submarine canyon, typically connected to an incised valley. After a phase of sediment bypass to the slope and basin plain, the canyon is typically filled or healed by clinoform progradation. Regional transgression resets the paleogeography, and the next depositional succession is likely to be offset along strike from the previous, due to large-scale lobe switching. Overall, this depositional succession is controlled by eustacy. However, sea level and transgressive-regressive cycles are not necessarily in phase, and these phase relationships may vary from one one river to the next, and from one cycle to the next. In addition, the stratigraphic expression of a given eustatic cycle can be present in one locality and absent or cryptic in another. Therefore, inference of eustacy from a local stratigraphic record, or from a single dip section or corridor through one or more SMDs is likely to yield a sea-level history that is incomplete or otherwise inaccurate. From an operational standpoint, we prefer a descriptive classification of major surfaces. The relationships of various kinds of stratigraphic surfaces define the stratigraphic framework. For sequence boundaries we use those surfaces that are most robust for regional correlation.