The north-central Gulf of Mexico area received rapid deposition of a basin-floor fan system consisting of interbedded muds, silts, and sandy turbidite deposits during the Pleistocene. Overpressure occurs at shallow depths when burial rates exceed the dewatering rates of sediment pore fluids. Two stratigraphic sequences in the region contain significant overpressure with elevated shallow-water flow risk within these units. We have used publicly available seismic and well data to identify the geomorphology and overpressure variation of these units. The previously described “Blue Unit” and its lateral extent, thickness, depth below sea level (BSL), and overpressure gradient have been revised. The Blue Unit extends from the northern portion of the Mississippi Canyon (MC) protraction area to as far south as the Atwater Valley (AT) protraction area. For the first time, the Green Unit’s lateral extent, thickness, depth BSL, and pore pressure are defined. The “Green Unit” was found to extend further south than the Blue Unit into the AT protraction area and further east in the Desoto Canyon protraction area. The tops of both units are highly incised by postdepositional erosional systems, whereas the base of each unit is well preserved. The top of the Blue Unit below the mud line (BML) varies from <70 m (<230 ft) in the north to as deep as 701 m (2300 ft) in the south, whereas the top of the Green Unit is as shallow as 300 m (985 ft) in the north to 901 m (2956 ft) in the south. Overpressure in the MC area has been reported just BML. The pore pressure gradient ranges from 0.47 to 0.52 psi/ft at the base of the Blue Unit and increases to 0.60 psi/ft within the Green Unit.


Recent advances in data acquisition and exploration for hydrocarbons in deepwater frontier areas have not only helped understand the depositional styles of the shelf-slope-basin-floor sedimentary deposits (Posamentier and Kolla, 2003; Sawyer et al., 2007; Kneller et al., 2016), but have also helped predict and mitigate the geohazards associated with such deposits (Ostermeier et al., 2002; Winker and Stancliffe, 2007a). Deepwater exploration in the Gulf of Mexico (GOM) provides an opportunity to understand the interaction among sediment supply, structural dynamics, sea-level changes, salt tectonics, and the resultant overpressure. Many wells drilled through these abnormally overpressured sand bodies have experienced shallow-water flow (SWF) events resulting in well abandonments. Winker and Stancliffe (2007b) and Eaton (1999) provide examples of overpressured permeable rocks that caused SWF events, and several wells were abandoned or lost in the Ursa field. Alberty et al. (1997) identify overpressured sandstone within 2000 ft below the mud line (BML) as one of the four causes for the SWF events in the deepwater (>1000 ft water depth) wells in the GOM. Overpressure has been observed within the top 200 m (656 ft) BML in the Ursa field of the Mississippi Canyon (MC) area (Winker and Stancliffe, 2007a; Flemings et al., 2008). Overpressure exists in the very rapidly deposited low-permeability mudstone and is a precursor for the large regional submarine landslides (Flemings et al., 2008). Although high overpressure over longer time periods drives failure on low-angle slopes, earthquakes may play a critical role in initiating slope failure in sediments weakened by overpressure (Sawyer et al., 2009, 2014; Stigall and Dugan, 2010).

During the Pleistocene, the Mississippi River system and the rivers draining through the southern Appalachian Mountains deposited several intraslope fan sequences containing layered and chaotic seismic facies (Dixon and Weimer, 1998; Winker and Booth, 2000). These intraslope fan units are separated by thick mud/shale deposits and have been interpreted to be separate lowstand hemicycles (Winker and Booth, 2000). The chaotic seismic facies are found to be sand-rich and are prone to SWF occurrences (Ostermeier et al., 2002; Winker and Stancliffe, 2007a; Ahmed, 2015). One such intraslope fan sequence has previously been named the Blue Unit (Sawyer et al., 2007; Stigall and Dugan, 2010), where numerous SWF events have originated (Figure 1). Some of these SWF events have compromised exploration and development well integrity (Ostermeier et al., 2002; Winker and Stancliffe, 2007b).

This study further investigates the lateral extent of the Blue Unit and reports on a second, deeper intraslope fan sequence, to be designated the Green Unit. The Green Unit has been delineated as the top and base of a sand-dominated channel complex with deposits from various river systems including the Mississippi River and the southern Appalachian rivers. The Green Unit bears the same overpressure characteristics and associated SWF hazards as the Blue Unit. Mapping these units will help to systematically identify SWF prone sand bodies as well as any overpressured zones. The SWF events (Ahmed, 2015) have generally sourced from the sand bodies from these two units in the MC area (Figure 2). Rapid burial might have inhibited dewatering of the sediment resulting in abnormal overpressure of the stratigraphic sequences. The SWF results during underbalanced drilling (when the mud-weight pressure is less than the formation pressure) through loose sand bodies. It has often been observed that overpressured mudrocks do not cause SWF. The results presented are based on publicly available 2D/3D seismic, well-log, mud-weight, and other pressure-related data to evaluate these two intraslope fan sequences in the north-central GOM (Figure 1) where most of the current exploration activities are ongoing.

Geologic setting

The study area is located on the northeastern GOM continental slope and covers parts of the MC, De Soto Canyon (DC), and Atwater Valley (AT) protraction areas from 27°39′43″N, 89°48′58″W to 28°48′41″N, 87°03′16″W in variable water depths ranging from approximately 520 m (1700 ft) in the north to approximately 2350 m (7700 ft) in the south (Figure 1). Winker and Booth (2000) describe the Pleistocene Mississippi River drainage system as depositing its sediment load by way of turbidites, debris flows, slumps, and channel-over bank deposits onto the continental shelf, continental shelf-margin, and along what is currently the abyssal plain during glacial lowstand sea level.

Diegel et al. (1995) and Prather et al. (1998) study the interplay among active salt movements, rapid sea-level-driven sedimentation, and gravity-slope failures. Sediment filled minibasins and salt-cored ridges dominate the morphology of the northern GOM. The continental shelf and continental shelf-margin sediments have been described by Coleman and Roberts (1988) and McFarlan and LeRoy (1988). Minibasin stratigraphy commonly displays cyclical sections of chaotic sediments overlying laminated sediments representing eustatic sea-level fluctuations (Rowan, 1995). Large volumes of sediment were deposited during sea-level lowstands in contrast to sea-level highstands when deposition was significantly reduced, resulting in alternating expanded and condensed sequences as often indicated by seismic facies. Condensed sections are thin, typically fine-grained sediments, deposited over wide areas. Expanded sections are comparatively thick sequences, rapidly deposited and dominantly composed of coarse-grained sediments. The expanded sections may contain sands with trapped pore fluids resulting in abnormally overpressured conditions. During glacial lowstand sea level, the Blue Unit and the Green Unit interpreted in this study were deposited on the continental shelf break, the continental slope, and part of what is now the abyssal plain (Figure 1) and are separated by a condensed section.

The marine isotope stage (MIS) 5, which contains the extinction events of the planktonic foraminifera Globorotalia flexuosa (70 k.y.), and the calcareous nannofossil Pontosphaera 1 (approximately 70 k.y.) separate the two units (Styzen, 1996; Winker and Booth, 2000). This datum is identified regionally in the GOM (Joyce et al., 1990; Martin et al., 1990). The Blue Unit was deposited during MIS 2–4 in response to the late Wisconsinan North American continental glaciation (Winker and Booth, 2000; Winker and Shipp, 2002). The Green Unit is a pre-MIS 5 event and correlates to the 1.3 Ma sequence of Gonzalez et al. (2004).

Available data processing

Publicly available seismic data (Figure 1) along with proprietary information were used to identify and map the Blue and the Green units (BOEM, 2016). The 2D seismic data include 2025 lines with a coverage of 6270 mi2, whereas the 3D seismic data have coverage of 7020 mi2. The seismic data processing includes wavelet rotation to zero-phase American standard (Brown, 2011) and amplitude balancing across multiple data sets. These public data were used to map the extents of these units and in correlating several smaller proprietary data sets.

The top surfaces of the Blue Unit and the Green Unit were mapped as the interpreted uppermost sand within the unit, or as a continuous reflector associated with the top of the sand (Figures 3 and 4). The lower surfaces of the units were mapped as the base of the deepest interpreted sand or as a continuous reflector associated with the base of the sand units. The Blue Unit is confined within the seismic data; however, the Green Unit described here may extend further to the south and east beyond the extents of the available seismic data. The surfaces of both units were picked initially on every 25th inline and crossline on the 3D seismic data. The inlines and crosslines are spaced at 82.02 ft intervals. An autopicking horizon hunt was conducted using the initial seed lines to map a complete surface. The final surfaces were further smoothed and interpolated to remove any mispicks and outliers. Nominal offset faults were interpreted with a grid of 25 × 25 lines; however, where the faults were interpreted with significant offset through more than two prominent reflectors, the interpretation density was increased to 5 × 5 line spacing. The impact of the faults is more evident in areas close to salt movement. Structural depths were determined by adding water depths to the sediment thickness BML. The bathymetry time horizon was converted to depth by using the seventh-order polynomial of Advocate and Hood (1993). Subsurface horizons were converted from time to depth using a constant velocity of 5000 ft/s. This value was chosen as a consistent conversion ignoring local variations within the vertical limit of investigation. Stratigraphic variations within the study area result in significant velocity differences (up to hundreds of feet) compared to a constant value of 5000 ft/s. Isopach maps were constructed by computing the difference between the upper and lower structure maps of each unit. The 3D models were generated to interpret the lateral and temporal variations in the geometries of each unit.

The regional pressure gradient maps of Burke et al. (2013) and Morris et al. (2015) were modified with the Berger Geosciences LLC., in-house well data sets and BOEM SWF events catalog (Ahmed, 2015). Mud-weight and SWF event data from 220 wells drilled in the MC and AT protraction area since 2014 support the regional pore pressure gradient maps of Burke et al. (2013) and Morris et al. (2015). The mud-weight and SWF event data were first corrected for water depth at each location to obtain a gradient referenced to the mud line. The mud weights used to contain a SWF event are a pseudomeasurement of the pore pressure and were treated separately from the weights that were used during normal drilling activities. Typically, drilling mud weights are chosen to exceed predicted pore pressure, anticipated over a drilling interval, to provide a safety margin. The resultant data sets from SWF events and drilling mud weights were compared to get a general trend and remove any outliers that represented significant overbalance. The final data were compared with the 0.47–0.52 psi/ft pore pressure gradient map between 3500 and 4500 ft BML of Morris et al. (2015) and the 0.60 psi/ft map of Burke et al. (2013). A close correlation among these three data sets was observed, and the results were used to analyze the pore pressure variations within the Blue Unit and the Green Unit.

Results and discussion

The Pleistocene to Late Pleistocene Mississippi River system deposited several intraslope fan sequences containing layered and chaotic seismic facies (Winker and Booth, 2000). These intraslope fan units are separated by layered mud/shale deposits and have been interpreted as separate lowstand hemicycles. The Blue Unit (Figure 5) and the Green Unit (Figure 6), the focus of this study, are two distinct examples of such intraslope fan units.

Blue Unit

The lateral extent of the Blue Unit, as mapped in this study, extends farther to the north, east, and south (Figures 3 and 5) compared to the Blue Unit extents reported by Winker and Booth (2000). This spatial extent adjustment of the Blue Unit is attributable to the accumulation of a significant volume of additional 2D/3D seismic data since the original Winker and Booth study.

The seismic character of the top of the Blue Unit varies considerably across the area due to varying degrees of channel incision and erosion, which may have dewatered and/or reduced or completely removed the permeable units. Gaps in the surface have been observed where syndepositional to postdepositional channels have cut through the entire thickness of the unit, or where the Blue Unit has pinched out against the salt bodies. Prevalent channel incision has been observed in the western half of the study area, which is proximal to the present-day MC (Figure 3a).

The Blue Unit is well developed and is generally thicker near the MC but thins away and eventually tapers out toward the east (Figure 3b and 3d). Local variations in thickness are associated with varying degrees of channel incision, landslide erosion of the unit’s upper surface, and pinch out against salt bodies. Laterally, the Blue Unit extends from the northern portion of the MC protraction area to as far south as the AT protraction area.

The top of this unit represents a gently dipping surface that deepens toward the southeast (Figure 5). The average depth to the top of the unit BML varies from <70 m (<230 ft) in the north to as deep as 701 m (2300 ft) in the southeast. The average thickness of the unit varies from approximately 60 m (approximately 197 ft) to 475 m (approximately 1558 ft) (Figure 5d). Overpressure begins at or near the mud line at the drilling locations in the Mars-Ursa basin of the MC area (Ostermeier et al., 2001; Flemings et al., 2008; Stigall and Dugan, 2010; Flemings et al., 2012). The depth (3500–4500 ft BML) of the 0.47–0.52 psi/ft gradient range modified after Morris et al. (2015) coincides with the mapped base of the Blue Unit (Figures 5 and 7), whereas the depth of the regional 0.60 psi/ft pressure gradient (modified after Burke et al., 2013) falls below the base of the Blue Unit throughout the study area (Figure 8a). The documented SWF events in the GOM (Ostermeier et al., 2001; Ahmed, 2015) are due to the underbalanced drilling of the unconsolidated and permeable sand bodies (with formation pressure greater than the drilling fluid pressure). The overpressure does exist very close to the mud line in the study area but will not result in an SWF event at a drilling location when drilling with mud weights greater than the prevailing formation pressure or when there are no permeable rocks. Underbalanced drilling through overpressured mudrocks will not result in an SWF event but may cause other borehole stability issues. All of the SWF events in the study area deepen from west to east and north to south following the top of the Blue Unit. This implies that SWF events are mostly associated with the sand bodies within the Blue Unit and any underbalanced drilling of these sand bodies will result in further SWF events. Although the onset of overpressure (a pressure gradient greater than hydrostatic gradient of 0.465 psi/ft) is found to be at or very near to the mud line above the top of the Blue Unit, it is important to map any sand bodies within the Blue Unit at a drilling location to mitigate against potential SWF.

Green Unit

The characterization and extents of the Green Unit have not been published previously. The green unit has been delineated as the top and base of a sand-dominated channel complex with possible various coalescing river systems including the Mississippi River and the Appalachian rivers (in the northern extents). Detailed mapping of channels and channel-levee deposits from each river system is beyond the scope of this study. The lateral extent, thickness, depth below sea level (BSL), and pore pressure have been mapped for this study. The MIS 5 condensed section separates the Blue Unit from the underlying Green Unit. This unit was found to extend further south than the Blue Unit in the AT protraction area and further east in the DC area (Figures 4 and 6).

The top of the Green Unit is less corrugated compared to the top of the Blue Unit and was found to have a continuous seismic reflector to trace across the study area. However, gaps in the top surface have been observed where the unit pinches out against salt bodies (Figure 4). Salt intrusion and withdrawal (diapirs and minibasins) have resulted in localized highs and lows in the upper surface of the Green Unit (Figure 6b and 6c). Channel cuts within the Green Unit are prevalent in the western half of the study area, which is closest to the MC depocenter.

The Green Unit is generally uniform along the interpreted extent, with the greatest thickness variation occurring in the west due to salt withdrawal (Figure 6d). The unit thins over the top of localized salt uplifts and thickens within the adjacent basins. Additional variations in thickness of this unit arise from channeling, most commonly observed in the west near the MC river system. The Green Unit was found to be more extensive and well developed than the Blue Unit. Laterally, the Green Unit extends further south beyond the limit of the Blue Unit into the AT protraction area and further to the east in the DC protraction area where lower deposition rates have occurred and the Green Unit is shallower. The top of the Green Unit is as shallow as 450 m (1475 ft) in the west–northwest to approximately 875 m (2875 ft) in the southeast. The average thickness of the unit varies between approximately 75 m (approximately 246 ft) and 640 m (approximately 2100 ft) in the study area (Figure 6d).

A comparison of the regional 0.6 psi/ft pressure gradient contour line in the GOM and the top of the Green Unit based on the seismic character suggests that the pressure gradient reaches 0.60 psi/ft prior to the depth of the top of the Green Unit (Figure 8b). This is true primarily for the western half of the mapped Green Unit. On the eastern side, the 0.60 psi/ft gradient is below the Green Unit. The deepening of the 0.6 psi/ft gradient is potentially related to greater water depth and lower deposition rates, which is supported by the SWF events reported in the area. The modified pressure gradient of 0.47–0.52 psi/ft of Morris et al. (2015) exists in the depth range of 3500–4500 ft BML, whereas the 0.6 psi/ft pressure gradient of Burke et al., 2013 is below the base of the Green Unit. Therefore, the pore pressure gradient within the Green Unit ranges from 0.50 to 0.60 psi/ft.


Detailed mapping using the significant volume of additional 2D/3D seismic, well-log, mud-weight, and other pressure data indicates that the extents of the Blue Unit are larger (by several square miles) than previously studied. The occurrence of SWFs deeper than the Blue Unit led to the investigation of the possibility of another overpressured sand-dominated sequence termed the Green Unit. The Green Unit is an extensive overpressured sand body that covers most of the MC area and portions of the AT and DC protraction areas.

The onset of overpressure (deviation from hydrostatic pressure gradient — 0.46 psi/ft) has been observed to be associated with the top of the Blue Unit. The overpressure gradient rapidly increases from 0.50 to 0.60 psi/ft once below the top of the Green Unit. The overpressure analysis further confirms the rapid deposition of the fan-slope system of the Blue Unit and the Green Unit, hindering dewatering and enabling subsequent overpressure generation. The information presented here is based on a regional model. Sand bodies within the Blue Unit and the Green Unit should be assessed locally within a prospect area to mitigate the risk of an SWF event. Planned accordingly, drilling operations will minimize the potential influx (SWF) due to overpressure within these sand units. In addition, real-time monitoring during drilling operations can result in early detection of the potential onset of a potential water flow.


BOEM is highly appreciated for providing the public 2D/3D seismic data used in this study. W. Berger II’s review highly improved the preliminary draft. D. Wedding’s effort in gathering the seismic data into one concise project is highly appreciated. A. Hewitt and K. Bates helped in picking the tops of the units. Reviews from C. Scherschel and an anonymous reviewer helped improve the manuscript significantly.

Data and materials availability

Data associated with this research are available and can be obtained by contacting the corresponding author.

William J. Berger received a B.S. in marine sciences from Texas A&M at Galveston and completed the 49th session of the Owner/President Management Program of Harvard Business School. He established Berger Geosciences in 2007 and has been working as the CEO/president of the company. With more than 25 years of experience in several deepwater basins around the world, site investigation projects include areas relating to archaeology, biology, geology, and oceanography. In 2010, he was involved with operations while drilling the riserless section of the second relief well in response to the Macondo project. This project is in an area of prominent overpressured sand units. Other interesting projects include providing support for an ROV survey within a brine pool in the deepwater GOM, and also a project relating to seafloor subsidence within a development area.

Shams Ul-Hadi received an M.S. in geology from the University of Peshawar, Pakistan, and a Ph.D. in geology from the University of Houston, Texas. He has been working with Berger Geosciences since 2012 as a geoscientist with a special focus on pore pressure and fracture gradient prediction in the GOM. He worked as an exploration geologist with the Oil and Gas Development Company Limited of Pakistan from 2001 until 2008. He started his professional career as a lecturer in geology at the University of Peshawar in 1999. His research interests include the tectonics of continental plate boundaries.

James Keenan received a B.S. in geology from the University of Houston. He joined Berger Geosciences in 2011 and holds the position of manager of the geohazards services group. The group is responsible for site characterization and evaluating potential drilling locations and field development. Before joining Berger Geosciences, he worked at Geosciences Earth and Marine Services. He has 20 years of experience in the oil and gas industry and is a member of AAPG and HGS.

Zachary Metz received a B.S. in geology from the University of Houston and has worked in the oil and gas industry for 15 years. He joined Berger Geosciences as a geoscientist in 2010 and currently works as chief operating officer of the company. He has experience around the world with high-resolution geophysical survey planning, acquisition QA/QC, geotechnical coring planning, QA/QC, and testing, geologic, biologic, and archaeologic visual survey planning, acquisition, and reporting, infrastructure planning and field development, seismic interpretation for geohazards, shallow hazard monitoring while drilling, integrated geoscience/geotechnical studies, cartography, GIS analysis, field remediation and decommissioning, well-log interpretation, drilling mechanics interpretation, and predrill and real-time geopressure monitoring, interpretation, and reporting. He has participated in 150+ different projects and has been a client representative in 50 different onshore/offshore projects including the monitoring of 30 different wells monitoring 70 sections for PPFG. In this role, he is responsible for business development and planning, staffing, training, employee performance, employee career development, project management, scheduling, and the quality of all deliverables in the organization. Before joining Berger, he worked at Geoscience Earth and Marine Services (2005–2010) as a student assistant, cartographer, and client representative. He has interest in shallow overpressured sand-prone units in the Gulf of Mexico and in overpressure prediction in carbonates and along lateral zones in unconventionals.

Thien Nguyen received a degree in earth science from the University of Houston. He is a geoscientist with Berger Geosciences LLC in a role to assist oil and gas companies in identifying offshore shallow geohazards. Prior to joining Berger in 2015, he worked with Fugro as a project geoscientist. He also worked with Geoscience Earth and Marine Services as a support service supervisor and geoscientist.

Freely available online through the SEG open-access option.