Abstract Direct simple shear experiments on mud samples from 0 to 15 mbsf (metres below seafloor) in the Ursa Basin (northern Gulf of Mexico) document that stress level impacts shear strength and pore pressure during failure. As burial depth increased (from 7.35 to 13.28 mbsf), cohesion decreased (from 12.3 to 6.5 kPa) and the internal friction angle increased (from 18° to 21°). For a specimen from 11.75 mbsf, an increase in maximum consolidation stress (from 45 to 179 kPa) resulted in an increase in the shear-induced pore pressure (from 29 to 150 kPa); however, the normalized peak shear stress decreased (from 0.37 to 0.25). Our results document that consolidation at shallow depths induces a positive feedback on pore-pressure genesis. For resedimented samples, which lack a stress history, cohesion was 3.6 kPa and the internal friction angle was 24°. As the maximum consolidation stress increased (from 40 to 254 kPa) on resedimented samples, the shear-induced pore pressure increased (from 22 to 203 kPa), whereas the normalized peak shear stress decreased (from 0.32 to 0.25). Our experiments showed that resedimented samples have similar strength and failure behaviour to intact samples. By constraining pore pressure, strength and initial stress state, we gain a better insight into slope-failure dynamics. Therefore, our experiments provide constraints on strength and shear-induced pore pressure at the onset of shallow failure that could be included in slope-failure and hazard models.

Abstract: We have examined the Neogene stratigraphic successions recovered from six wells located along a present-day middle neritic (current depositional depth 92 m) to upper bathyal depth (current depositional depth 482 m) transect oblique to the shelf/slope margin in the northern Gulf of Mexico (GOM) using calcareous plankton biostratigraphy. The latter were used to conduct stratigraphic interpretation of the sections and to determine their completeness. We establish that all sections vary considerably in thickness and completeness, depending on depth of deposition, as estimated from benthic foraminiferal analysis, which shows that depositional depth at the six sites changed little through the Neogene. The shallowest section (~90-m estimated depositional depth through the Neogene) is the thinnest with the most complete Upper Miocene–Pleistocene record, whereas the deepest section (~600– 800-m estimated depositional depth) is the thickest but also contains the least complete Pliocene–Pleistocene record. The Upper Miocene to Pleistocene sediments deposited between ~200- and 500-m estimated depositional depth exhibit a characteristic allostratigraphic architecture, with sedimentary units bounded by unconformities associated with 1- to 2-Myr hiatuses that vary little along the transect. We integrate the stratigraphic architecture along our local transect in the regional Cenozoic depositional framework in the GOM of Galloway and coauthors and establish that the allostratigraphic units (AUs) correspond well with several of the genetic and seismic sequences delineated. We interpret the depth-related increase in thickness of the Upper Miocene–Pleistocene AUs in light of the sedimentary processes discussed by these authors. However, our interpretation differs considerably from theirs based on our documentation of temporally incomplete sections in the wells. The sedimentary pattern in Well-3 (~200-m estimated depositional depth) is quite different from that at nearby Well-1 (~100–600-m estimated depositional depth), although very similar to the wells further west, even though the distance between Well-3 and Well-6 is about four times that between Well-3 and Well-1. We note also that the stratigraphic pattern in Well-1 changed ~8 Ma, from highly discontinuous before to remarkably continuous after. We have found no clear evidence that glacio-eustasy shaped the Neogene stratigraphic record in the study area. Therefore, we question whether glacio-eustasy was the primary forcing mechanism on stratigraphic architecture in the GOM beyond the shallow part of shelves and propose that salt tectonics may have been a more prominent factor in controlling accommodation. An allostratigraphic architecture was described earlier from the De Soto Canyon northeast of the GOM transect, where the AUs and their boundaries were shown to match, respectively, the seismic sequences and surfaces on the nearby Florida margin. We therefore consider the AUs along the GOM transect as corresponding as well to seismic sequences and therefore to parts of depositional sequences. Based on this, we review notable difficulties in characterizing seismic features (sequences and surfaces) in concrete stratigraphic records and recommend a greater awareness of the temporal significance of unconformities, many of which are associated with multimillion-year hiatuses.

Abstract Stacked mass transport deposits interbedded with turbidite sandstone reservoirs were characterized in detail from image logs at Thunder Horse, a conventional asset located in the Boarshead mini-basin, lower Mississippi Canyon, Gulf of Mexico. Image logs visualize the borehole wall, allowing bedding boundary information and some sedimentary fabric to be identified. From bedding boundaries, dip azimuths and magnitudes can be calculated, and depositional processes can be inferred from patterns in the dips. Dips in the primary sandstone reservoirs show little variation (< 20°) in structural tilt, indicating continuous deposition by high density turbidity currents prior to major structural deformation and are thus the best proxy for regional structural dip. In contrast, dips in 80% of all mudrocks in this canyon show widely varying dip magnitude and azimuth over small and large scales and are interpreted as having been deposited by slumps, slides, and folds, building up to form mass transport complexes. Only very rarely do we observe mudrock dips that are conformable to the regional dip and strike. Unconformable bedding contacts at the tops of the sandstone reservoirs in northern Thunder Horse suggest scouring and erosion by overlying mass transport deposits, whereas basal sandstone contacts are in conformance with structural dip, suggesting amalgamation. On seismic data, albeit low-resolution due to the overhanging salt canopy, the mass transport complexes appear to build up into a gradual mound. Core was acquired in the upper 60 m of the mass transport complex, revealing convoluted and folded bedding capped by highly bioturbated marl, which is thought to represent a Glossifungites hiatal surface resulting from the mass transport complexes building up a topographic high, forcing subsequent debris and turbidite flows to travel around rather than over that location. The topographic high ponded the overlying reservoir behind it, preferentially thickening the sandstone reservoir behind the mound. Thus, mass transport complexes were found to both erode and augment sandstone reservoirs. Biostratigraphic dating of the two reservoirs brackets their deposition as occurring in <400,000 years during the Serravalian, Middle Miocene, probably as a result of the Harang shelf failure. As a case study this type of integrated, high resolution data (core, image logs, seismic) has wide applicability to other deepwater, subsalt reservoirs, as improved structural and depositional interpretations inform reservoir performance and impact future well plans. Further, recognition of mass transport complexes as agents of reservoir compartmentalization and sandstone ponding allow for more accurate reserve estimation.