Keywords

Introduction

Pleistocene sedimentation in the Gulf of Mexico from the Mississippi River is characterized by rapid sedimentation upon a mobile salt substrate (Worrall and Snelson 1989). Offshore Texas and Western Louisiana, individual slope minibasins are surrounded by elevated salt highs (Pratson and Ryan 1994) producing a remarkable hummocky topography. This morphology is obscured in front of the Mississippi delta, where sedimentation has been very rapid, exceeding 25 m ky−1 (Expedition 308 Scientists 2005).

Ursa Basin (~150 km due south of New Orleans, Louisiana, USA) lies in ~1,000 m of water (Fig. 1). From bottom to top, the Pleistocene to Holocene sedimentary sequence in Ursa Basin consists of: (1) the lower Mississippi Canyon Blue Unit, a late Pleistocene, sand-dominated, “ponded fan” (Winker and Booth 2000; Sawyer et al., 2007), (2) a mud dominated channel-levee assemblage with dramatic along-strike variations in thickness and (3) a mud drape deposited during the last ~20 ky (Fig. 2; Flemings et al. 2006). The mudstone package, belonging to the eastern margin of the larger channel-levee system of the Mississippi Canyon, has Mass Transport Deposits (MTDs) that record failure of the margin during the Pleistocene (Figs. 1 and 2; Flemings et al., 2006; Sawyer et al., 2007). Holocene failures have also been mapped all along the Gulf of Mexico, with the largest ones originating from Ursa Basin (Fig. 1; see also McAdoo et al. 2000). Ursa Basin is of economic interest because of its prolific oilfields, which are currently being exploited in the nearby tension leg platforms of Mars and Ursa. Geohazard characterization is therefore essential for safe offshore activities. For the purpose of this paper, we use data from two boreholes (Figs. 1–3) acquired during IODP expedition 308 (Expedition 308 Scientists 2005, Flemings et al. 2006), including Logging While Drilling (LWD), measurements of physical properties and geotechnical analysis performed on whole round samples.

Fig. 1
figure 1_15

Detailed swath bathymetry (US national archive for multibeam bathymetric data) shaded relief of the Mars Ridge (eastern levee of the Mississippi Canyon) showing location of Ursa drill sites U1322, U1324 and seismic line shown in Fig. 2. Profuse evidence of mass-wasting processes is evident. Inset shows location

Full size image
Fig. 2
figure 2_15

Top. Seismic cross-section (for location see Fig. 1). Bottom. Interpreted cross-section. The sand-prone Blue Unit has been incised by a channel-levee complex and then overlain by a thick and heavily slumped hemipelagic mudstone wedge that thickens to the west (left). The Blue Unit sands are correlated to a distinct seismic facies. The thickness of the hemipelagic mudstone above the Blue Unit does not change significantly in the north-south direction. Major lithostratigraphic units identified during IODP Exp. 308 are labeled for correlation with Figs. 3–5. Seismic section reproduced with permission of Shell Exploration and Production Company (After Flemings et al. 2006)

Full size image
Fig. 3
figure 3_15

Stripe of seismic profile with major reflectors labeled and lithologic log (MTDs in red, non-failed sediments in green) for drill sites U1324 and U1322 (see also Expedition 308 Scientists 2005). Logs of physical properties correspond from left to right to: (1) Median grain size (red line) and grain size fraction abundance; (2) Undrained shear strength as determined from shipboard motorized vane tests and trends for Cu•(γ'z)−1 ratios of 0.1, 0.3 and 0.5; (3) Plastic limit (red), liquid limit (cyan) and shipboard measured water content (blue) and (4) Overpressure determined from Skempton's (1957) eq. (blue line), porosity vs. stress relationships (green line with dots), preconsolidation pressures measured in oedometer tests, DVTPP (blue inverted triangles and T2P (green inverted triangles) piezometers and 1-D modeling results (red line)

Full size image

The aim of this paper is to document how pressure, stress, and geology couple to control fluid migration on Ursa Basin, illuminate the controls on slope stability and understand the timing of sedimentation and slumping.

Methods

To accomplish the objectives stated above a laboratory testing program was established, which included grain size analysis using a Coulter LS100 laser diffractom-eter and Atterberg Limits determined using a British Fall cone device (Feng et al. 2001), isotropically consolidated undrained (CIU) triaxial tests, and uniaxial incremental loading tests (see also Urgeles et al. 2007). The data resulting from these tests are here discussed together with IODP Expedition 308 Shipboard data (Expedition 308 Scientists 2005; Flemings et al. 2006), including visual core descriptions, moisture and density data, vane shear and pocket penetrometer strength determinations and in-situ pore pressure measurements using the DVTPP and T2P piezoprobes (see Flemings et al. 2008; Long et al. 2008 for further details on these measurements). In this paper, pore pressure is most often described in terms of overpressure (λ), defined as:

$$\lambda = {{\left({P - P_h} \right)}/{\left({\sigma _v - P_h} \right)}}$$
((1))

where P is pore pressure P h is hydrostatic pressure and σv is lithostatic or total stress. A value of 0 implies hydrostatic conditions, a value of 1 means pore pressure equals the lithostatic stress.

2-D margin simulations along the transect of Fig. 2 were carried using the Finite Element Software “BASIN” (Bitzer et al. 1996, 1999), which allows for strati-graphic, tectonic, hydrodynamic and thermal evolution to be modeled. In “BASIN” compaction and fluid flow are coupled through the consolidation equation and the nonlinear form of the equation of state for porosity, allowing non-equilibrium compaction and overpressuring to be calculated.

Results

In Ursa basin, Site U1324 (on the upslope part) was drilled to more than 600 mbsf and site U1322 (downslope) to about 250 mbsf (Fig. 2). The sediments drilled at these sites included several MTDs (5 at Site U1324 and 9 at U1322). On seismic reflection profiles the thicker MTDs are characterized by discontinuous and/or transparent to low amplitude reflections. MTDs show increased resistivity and density in logs, presumably due to transport-induced compaction, while folds, some with half-wavelengths of a meter or more, are apparent both in cores and logs. Despite differences in visual aspect and physical properties the sediment composition shows no major differences between MTDs and non-failed deposits. Sediment grain size analyses indicate that at both sites the sediments are made of ~30% clay and ~70% silt, the mean grain size is about 4 microns, and these values remain fairly constant with depth (Fig. 3). The sediment plastic limit is at about 35%, while the liquid limit decreases from 70% to 55% in the upper 10 m of sediment column and then becomes relatively constant around the latter value (Fig. 3). The sediment water content moves from values at or higher than the liquid limit for the upper 20 mbsf and then gradually decreases to values close to the plastic limit at 100– 125 mbsf. From that depth downhole the water content sticks to the plastic limit (IL =0%). On the Casagrande plot, samples from both sites plot on a line parallel and above the A-line identifying the sediment as clays of high plasticity.

The vane shear and pocket penetrometer data show similar values of undrained shear strength at equivalent depth for both drill sites, ranging from a few kPa near the seafloor to about 250 kPa at about 600 mbsf (Fig. 3). CIU triaxial tests (see Urgeles et al. 2007) were carried out on samples obtained from MTDs and non-failed deposits at Sites U1324 and U1322. Prior to shearing the samples were isotropically consolidated. Some of the samples were brought into the normally consolidated state others remained in the overconsolidated state prior to shearing. However, on the stress path plot the whole set of tests show relatively consistent results with a friction angle of 28° and little cohesion of around 7 kPa.

Using consolidation theory, overpressure was estimated from pre-consolidation pressures determined from incremental loading consolidation tests, and measurements of pore pressure using the DVTPP and T2P pressure probes (Fig. 3; see also Flemings et al. 2006, 2008; Long et al. 2008). Despite significant scatter, results appear to show overpressures in the range of 0.6 to 0.8, suggesting that non-equilibrium consolidation occurs in Ursa Basin (Fig. 3).

The consolidation tests results also provide parameters that can be used in basin and interstitial fluid flow modeling. These parameters include the initial porosity, hydraulic conductivity and specific storage (see Table 1 and Fig. 4). Consolidation tests were performed in sediments of the Southwest Pass Canyon Formation and hemipelagic sediments above (corresponding to Lithostratigraphic Units I and II in the cores; see Fig. 2 for equivalence). Unfortunately, no samples could be retrieved from the Blue Unit and central sandier part of the Ursa Canyon, as drilling within these formations implied a high risk for shallow water flow sands. Parameters for Sand and Silt which, according to industry well-log data (Sawyer et al. 2007) are a significant constituent of the Blue Unit and core of the Ursa Canyon are taken from the literature (Table 1; Reed et al. 2002).

Fig. 4
figure 4_15

Permeability and specific storage vs. effective stress derived from incremental loading consolidation experiments used to derive the set of initial parameters for 2-D basin modeling within “BASIN”. For location of experiments see Fig. 3

Full size image
Table 1 Parameters for basin and fluid flow modeling
Full size table

Discussion

As shown above, overpressure estimates and measurements performed in various ways, indicate that an overpressure of 0.6–0.8 is present in Ursa Basin (Fig. 3; see Flemings et al. 2006; Long et al. 2008). Despite the significant scatter observed in these estimates and measurements, all data suggest that similar overpressure exists at Sites U1324 and U1322. 1-D modeling results however indicate that given the sedimentation rates at both sites, which are much higher at Site U1324 than at U1322, there should be a significantly higher overpressure at Site U1324 (Fig. 3). Mean sedimentation rates at Site U1324 are 9.6 m/ky with peaks exceeding 25 m/ky, while at site U1322 mean sedimentation rates are 3.5 m/ky with peaks at 16 m/ky (Expedition 308 Scientists 2005). According to 1-D modeling results, these sedimentation rates imply that overpressure should be around 0.8 at Site U1324 and only around 0.2 at U1322 (Fig. 3). Using the 2-D modeling software “BASIN” the margin stratigraphic evolution and the resulting interstitial fluid flow pattern and overpressure generation can be better understood (Fig. 5). For this experiment initial and boundary conditions are: (1) no flow occurs at the basement and model sides, (2) initial conditions are hydrostatic, (3) no subsidence occurs and (4) the initial (decompacted) thickness of the various formations can be estimated from van Hinte's (1978) equation:

$$T_0 = \left({1 - \Phi _N} \right)T_N /\left({1 - \Phi _0} \right)$$
((2))
Fig. 5
figure 5_15

Margin stratigraphic and hydrodynamic modeling with “BASIN” at final simulated present-day conditions. (a) Margin stratigraphy (red MTDs) according to seismic profile depicted in Fig. 2. (b) Fractional porosity. (c) Log hydraulic conductivity (m/s). (d) Excess pore pressure (kPa). (e) Overpressure (λ)

Full size image

where φ N ., T N are the present-day porosity and thickness (assuming ν p : 1,600 m/s; Flemings et al. 2005) and φ0, T 0 are the original porosity and thickness. MTDs removed/added little overburden because the failed masses did not evacuate the failing zone and therefore have little to no effect in both 1-D and 2-D simulations.

The margin stratigraphic evolution is modeled since deposition of the Blue Unit, roughly 100 ka ago, along the seismic line shown in Fig. 2. Margin modeling is performed using a set of initial sediment types and their corresponding hydraulic conductivity and specific storage shown in Table 1. Each stratigraphic unit is made of a mixture of the different sediment types (Table 1), which result in averaged initial physical properties (see Bitzer et al. 1999 for further details). The “BASIN” simulations show that, as it should be expected, porosity and hydraulic conductivity decrease with time and depth along the section (Fig. 5b, c). However, high hydraulic conductivities ~10−85 m/s remain in the sandier Blue Unit and in the core of the Ursa Canyon after the whole sedimentary package has been deposited. The hydraulic conductivities on the muddy sediment wedge above the Blue Unit are between 1 to almost 3 orders of magnitude lower (Fig. 5c). The modeled porosity and hydraulic conductivities result in fluid flow from West to East along the Blue Unit and Ursa Canyon and a vertically upward migration on the overlying muddy formations. The model also shows that pore pressures above hydrostatic started to develop at ~53 ka with onset of the Southwest Pass Canyon sedimentation, which deposited more clayey material (Fig. 5d). Due to the higher sedimentation rates excess pore pressure develops further near Site U1324, where the overburden is thicker. The more permeable nature of the lower Blue Unit allows fluid to flow laterally from Site U1324 to U1322 inducing higher excess pore pressures at this Site than should be expected given the sedimentation rate at this location (Fig. 5). This effect is better seen in terms of overpressure (λ). Figure 5e clearly shows similar, or sometimes higher, overpressure at Site U1322 compared to Site U1324. The 2-D simulation results agree better with the estimated and observed overpressures at both Sites. It also shows that the higher overpressures concentrate within the depth range of 100 to 200–350 mbsf.

Overpressure build up in Ursa Basin has probably played an important role in slope failure generation. Major additional controls in slope stability in Ursa Basin include variations in slope angle due to depositional processes and salt tectonics. Presently the regional slope in Ursa Basin does not exceeds 2°, and local slopes rarely exceed 4°. The Gulf of Mexico is an area where large seismic ground motions are not probable (Petersen et al. 2008). No large earthquakes have been reported recently with the exception of two 5 < Mw < 6 earthquakes (Preliminary Determination of Earthquakes (PDE) Catalog 1973-present). Recent studies, suggest however that these are a result of shallow slippage (most probably large-scale gravitational sliding) rather than deep-seated tectonic processes (Nettles 2007, Dewey and Dellinger 2008), and therefore we will not consider seismic ground motions as potential triggering mechanism.

The main characteristics of MTDs in Ursa Basin are: (a) they occur on a more or less uniform slope, (b) the failure planes are subparallel to the seafloor and (c) the length of the failure surfaces is large compared to the failure thickness (Figs. 1 and 2). Therefore, it is considered that an “infinite slope” approach provides a first approximation to the margin stability. Urgeles et al. (2007) show the results of a drained slope stability evaluation in Ursa Basin using the parameters identified from the CIU triaxial tests (c = 7 kPa, Φ = 28°). Urgeles et al. (2007) show that for the range of overpressures observed in Ursa Basin and the present regional slope gradient, the slope can be considered safe. For failure to occur overpressure values close to 0.9 are needed, or slope gradients need to approach angles between 7.5 an 11° depending on the overpressure. Using undrained shear strengths the slope appears in the stable to metastable conditions (Urgeles et al. 2007).

Using the output from the overpressure simulations it is also possible to investigate the margin stability conditions during the last 100 ky. Figure 6a shows that the margin remained relatively stable with a Factor of Safety (FoS) ~6 until ~45 ka, and then an overall decrease in margin stability occurred. Despite several oscillations, the margin's FoS (Fig. 6a) has remained lower at the site of lower overburden (Site U1322) until about 20 ka, in agreement with core evidence showing that most failures are found at this Site. From 20 ka onwards the FoS became similar for both sites. Using the output from the overpressure simulations it is also possible to investigate the depth at which failure could most likely have occurred. Figure 6b indicates that the reduction in the margin's FoS was accompanied by an increase in thickness of the potentially unstable sedimentary package. It is also found that the potentially unstable sediment package is thinner at Site U1322 than at Site U1324 in agreement with the thicker MTD found at the latter site (Fig. 6b).

Fig. 6
figure 6_15

(a) Ursa Basin FoS at Sites U1324 and U1322 using the output from overpressure (λ) simulations performed with “BASIN”. (b) Depth to the potential failure surface (depth of lowest FoS) at Sites U1324 and U1322

Full size image

Conclusions

Physical property data and geotechnical measurements indicate that sediments from Ursa Basin are largely overpressured, inducing effective stresses that are only 25% of those that would exist under hydrostatic conditions. At Sites U1324 and U1322 similar overpressures are found despite large differences in sedimentation rates. Modeling simulations show that this most probably results from excess pore pressure being laterally transferred through the Blue Unit from places of high to low overburden. The highest overpressures concentrate in the depth range between ~50 and 250– 350 mbsf. 2-D modeling results indicate that pore pressure started to build up with onset of deposition of the South West Pass Canyon Formation ~53 ka ago.

Basic regional slope stability analysis suggests that the slope is stable under current conditions. For the slope to fail it is necessary that the overpressure exceeds a value of ~0.9 or the regional slope steepens above 7°. Using the output from the margin stratigraphic and hydrodynamic evolution it is found that the margin's fluid flow pattern induced lower stability at the foot of the slope where sedimentation rates are lower. The margin stability reduced significantly from ~45 ka onwards. The decrease in margin's stability was accompanied by increased thickness of the potential failure package.