Marine Geology 337 (2013) 80–97

Contents lists available at SciVerse ScienceDirect

Marine Geology journal homepage: www.elsevier.com/locate/margeo

U-shaped slope gully systems and sediment waves on the passive margin of Gabon (West Africa) Lidia Lonergan ⁎, Nur Huda Jamin, Christopher A.-L. Jackson, Howard D. Johnson Department of Earth Science and Engineering, Imperial College London, London SW7 2AZ, UK

a r t i c l e

i n f o

Article history: Received 23 June 2012 Received in revised form 6 February 2013 Accepted 10 February 2013 Available online 19 February 2013 Communicated by D.J.W. Piper Keywords: Gabon slope gullies sediment waves turbidity current deposition 3D-seismic data slope sedimentation

a b s t r a c t 3-D seismic reflection data has enabled the documentation of a system of remarkable modern and buried u-shaped gullies which intimately co-exist with upslope migrating sediment waves along 80 km of the Gabon continental slope. The modern gullies occur on a silty mud-dominated slope in water depths of 150–1500 m on an ~50 km wide slope with a gradient of 4.5° decreasing to 1.5°. The gully sets persist laterally for distances of at least 40 km and extend downslope for distances of up to 60 km. The gullies are u-shaped in crosssection, with a relief of 5–30 m, and widths of 50–400 m. Intriguingly, the gullies become narrower and shallower with distance down the slope, as well as increasing in number down slope. The majority of the gullies appear to be erosional but some appear to have resulted from simultaneous aggradation along inter-gully ridges and non-deposition along the adjacent gully floor. Hence, these gullies are interpreted to have formed mainly in response to spatially-variable deposition, rather than erosion. Upslope migrating sediment waves occur in close proximity to the gullies. Gullies cross fields of sediment waves and waves are observed to migrate up-slope locally within both the erosional and aggradational gullies. Evidence is lacking for any slumping or headward erosion in the headwall region of the gullies, which discounts formation by very local sediment gravity flows originating from shelf-edge collapse, as has been observed in other v-shaped gully systems. Based on our new data, and supported by theoretical studies on the mechanics of turbidity currents, we propose that the gullies and related sediment waves were formed by diffuse, sheet-like, mud-rich turbidity currents that presumably originated on the shelf. Instabilities in the turbidity currents generated a wave-shaped perturbation in a crossflow direction leading to regularly spaced regions of faster and slower flow. For the non-aggradational and erosional gullies it is inferred that gully axes experienced flow velocities that mainly exceeded the settling velocity of the sediment in suspension, and thus no deposition occurred. In contrast, the aggradational gullies indicate lower flow velocities with sediment deposition both within the gully axes and on the gully flanks. Mixed mode gullies are also found which indicate that successive flows can experience variations in flow properties leading to interspersed erosional and depositional events. © 2013 Elsevier B.V. All rights reserved.

1. Introduction The majority of recent research on the sedimentology of continental margins has focussed on areas of high sediment-supply where slope canyon systems typically feed sediment-rich submarine fans, often with well-developed channel-levee systems (e.g. Piper and Normark, 1983; Damuth et al., 1988; Pickering et al., 1995; Weimer and Slatt, 2007, amongst many others). In contrast, the fine-grained, lower-sediment supply parts of these margins exhibit a different range of less wellunderstood sediment supply systems. The widespread increase in both multibeam seafloor imaging and 3D seismic reflection data is improving our knowledge of continental slope systems and allowing greater insight into the sedimentary processes and products that occur in lower sediment-supply areas (e.g. Pilcher and Argent, 2007; ⁎ Corresponding author. Tel.: +44 20 75946465; fax: +44 20 7594 7444. E-mail addresses: [email protected] (L. Lonergan), [email protected] (C.A.-L. Jackson), [email protected] (H.D. Johnson). 0025-3227/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.margeo.2013.02.001

Straub and Mohrig, 2009; Jobe et al., 2011). Here we use 3D seismic reflection data to describe suites of closely-spaced, sub-parallel, u-shaped, slope gullies and associated sediment waves imaged at the seabed in a low-sediment supply portion of the Gabonese continental margin (Fig. 1). An unusual feature of this area is the close spatial and temporal relationship between seabed sediment waves and u-shaped gullies, a feature not documented by any previous studies of gully systems. We document the morphology, fill, and down-dip expression of modern and buried gullies in the study area revealing a greater variety in gully morphology and organisation than previously observed for u-shaped gullies. We investigate the relationship between sediment waves and gullies and propose a model for gully formation. 1.1. Slope gullies Gullies are commonly observed on submarine slopes along many on continental margins. They are important areas of submarine erosion and conduits for downslope sediment transfer from the outer shelf to

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

81

17°E

9°E 7°N

Africa 7°E

9°E

Atlantic Ocean

11°E

Eq. Guinea

Gabon Libreville

Ogooué River

0

Study area

Ogooué Fan

Congo

7°S

Port Gentil

Nkomi

2°

2°

Gabon

Ndogo

Study Area

.

aR

ng

a Ny

Congo

4°

4°

20

00

m

Congo River

Congo Fan 300 km 6°

6°

7°E

9°E

11°E

13°E

Fig. 1. Location map. Basemap made using Google Earth; Marine bathymetry from the Global Topography V12.1 dataset from http://topex.ucsd.edu/marine_topo/mar_topo.html using the Google Earth kmz file; and land topography is the ETOPO-1 dataset from Amante and Eakins (2008) from http://www.ngdc.noaa.gov/mgg/global/global.html. Location of Congo and Ogooué fans after Séranne and Nzé Abeigne (1999).

the base-of-slope and basin floor. A gully, while not formally defined, is a term used for steep-sided, relatively straight channels, which are up to a few tens of metres deep. Based on their cross-sectional geometry two types are commonly recognised: (1) v-shaped (e.g. Twichell and Roberts, 1982; Pratson and Coakley, 1996; Pratson et al., 2007; Dowdeswell et al., 2008; Micallef and Mountjoy, 2011), and (2) u-shaped (Flood, 1983; Chiocci and Normark, 1992; Field et al., 1999; Spinelli and Field, 2001; Chiocci and Casalbore, 2011). The gullies described in this paper are of this second type. V-shaped gullies typically originate at, or close to, the shelf edge, or at changes of slope gradient on steeper parts of the slope (e.g. Micallef and Mountjoy, 2011). They are frequently associated with, and feed into, larger canyon or channel systems as is well documented on the New Jersey continental margin (Twichell and Roberts, 1982). They are also widespread on continental margins affected by glacial processes and have been found on the surface of slope fans deposited at the mouths of major ice streams (e.g. Belgica Fan, Antarctica; Dowdeswell et al., 2008). They may be organised into dendritic networks in the upslope reaches or headwalls of larger canyons, or form parallel channels in inter-canyon areas (Pratson and Coakley, 1996; Pratson et al., 2007). There is general consensus that this type of gully originates by mass wasting processes, triggered by failure events at the shelf edge.

The gullies may be further modified by later turbidity currents, or other near seabed currents. Seepage-induced slope failure can explain the regular spacing (e.g. Orange et al., 1994). More recently, Micallef and Mountjoy (2011) proposed that the source of the gravity flows responsible for initiating v-shaped gullies in the Cook Strait, New Zealand is cascading, dense-water flows driven by seawater density contrasts. U-shaped gullies are less-well understood and typically form in straight, sub-parallel sets which rarely branch or feed into larger channels. These type of gullies appear to be less commonly developed, or perhaps less frequently recognised along continental margins, but they have been described from the Eel river area on the northern Californian continental margin (Field et al., 1999; Spinelli and Field, 2001) and on the upper slope of the Tyrrhenian sea, seaward of the Tiber delta (Chiocci and Normark, 1992; Chiocci and Casalbore, 2011). Linear, spaced, erosional grooves that form in fine-grained, cohesive sediments, termed furrows, have been recognised since the 1970s (Dyer, 1970; Coleman et al., 1981; Flood, 1983). They are morphologically broadly similar to the features described as gullies in later papers, except that they tend to be almost an order of magnitude smaller with widths in the range of 1–20 m, and depths of less than 10 m. They have been recognised from tidal estuaries (Dyer, 1970), the shelf edge to slope transition

82

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

(Coleman et al., 1981) and on the abyssal plain (Flood, 1983), in areas where aligned currents flowing parallel to the furrows were also measured. Flood (1983) and Coleman et al. (1981) both proposed that the furrow bedforms were erosional features that formed due to the action of a longitudinal helical motion in the benthic boundary layer of the current. An unusual feature of the Californian and Tyrrhenian Sea u-shaped gullies is that they are aggradational in form. Models proposed for these types of gullies have tried to address how they appear to be both erosional, and aggradational. Chiocci and Normark (1992) suggested that the Tiber delta related upper slope gullies in the Tyrrenhian Sea are net-depositional features that are growing because of faster aggradation in inter-gully areas with respect to gully axes. They related gully evolution to late Quaternary sea-level changes and proposed that during sea-level lowstands, when the Tiber river mouth was closer to shelf edge, the steeper gradient of the continental shelf and higher sediment discharge led to local bedload transport of coarser material that initiated gully development. The movement of coarser/ and or thicker sediment flows along gully floors resulted in slower accumulation rates than on adjacent gully highs. They suggested that the uniform nature of the sediments in the inter-gully areas indicates that muddy gravity flows covered large parts of the gullied area and contributed to the aggradation of the gully system. More recently Chiocci and Casalbore (2011) have reassessed these data as well as new data from upper slope gullies related to the smaller Volturno river, located to the south of the Tiber system, and the Simento river system in Sicily. Because the gullies are forming in highstand conditions in the Volturno and Simento systems, the authors propose that the gullies can also form by hyperpycnal flows generated when large-eruptions overwhelm the river catchments with fine-grained volcanoclastic sediment; thus implying that gully formation can be independent of sea-level cycles. Field et al. (1999) and Spinelli and Field (2001) proposed a hybrid model of an initial erosional gravity flow that erodes and excavates gullies during periods of low, or falling sea-level to explain the northern Californian slope gullies. During higher sea-levels, when nepheloid deposition dominates, sediment is draped over the gullies, preserving the gully morphology and resulting in gradual aggradation upwards through time. This brief review of the few observational studies of u-shaped gullies indicates that the exact processes responsible for the formation of this gully type are not well understood and it is probably reasonable to assume that there may be different processes operating in different submarine environments. More recently, Izumi (2004), Hall et al. (2008) and Fedele and Garcia (2009) have used theoretical mathematical analyses to explore how broad, sheet-like turbidity currents might be responsible for initiating slope gullies. Using numerical simulations Fedele and Garcia (2009) showed that sheet-flow gravity currents crossing a bedslope transition (e.g. a concave transition at the shelf edge) develop centrifugal instabilities at interfaces within the flow. This creates a wave-like deformation of the current interface in the cross-stream direction with higher flow velocities and bottom shear stress in regions of smaller current thickness. They predict that these higher velocities and shear stresses will trigger the formation of longitudinal bedforms, either by erosion or differential deposition. At higher velocities the flow can erode the substrate, or not deposit. This contrasts with inter-gully areas which will have lower flow velocities and the potential for net-deposition. Their analysis predicts that the instabilities, and hence gullies, should be regularly spaced. Izumi (2004) presented a mathematical model for the formation of purely erosional gullies by wide unconfined, subcritial turbidity currents, but it is not clear whether this model is appropriate for gully systems that aggrade through time. Hall et al. (2008) conducted a linear stability analysis of the three dimensional Navier–Stokes equations which takes account of the coupled interaction of the fluid and particle motion inside the current with the erodible bed below it. In contrast to the analyses of Izumi (2004) and Fedele and Garcia (2009), the

instability mechanism identified by Hall et al. (2008) does not require any assumptions about subcritial flow nor does it require the presence of a slope break. Hall et al. (2008) demonstrate that a flow-transverse, wave-shaped perturbation is generated within the turbidity current which results in higher shear stresses, particle concentrations and erosion in the wave troughs. Similarly to the analyses of Fedele and Garcia (2009), the wave-shape nature of the flow perturbation predicts that the gullies will be regularly spaced. Hall et al. (2008) also note that once the trough-like geometry is imposed on the substrate the initial perturbation should amplify so later flows would be likewise perturbed and the bedforms generated can continue to grow. Previously published studies describing the detailed geometry of u-shaped slope gullies are limited to seabed and very shallowly buried examples, developed over relatively small areas. Here we provide descriptions of a wider range of u-shaped gullies, with varying architectures and developed over a large slope area. The Gabonese gullies also show more a complex plan-view organisation than that observed in other areas, with the gully spacing and density increasing down slope. Additionally our examples include much older gully systems found at depths of up to ~ 800 m beneath the modern seabed. In contrast to previous studies we also describe, and attempt to explain, an intriguing relationship between the seabed gullies and coeval sediment waves. 2. Regional setting and study area The study area is located to the south of Port Gentile, on the Gabonese continental slope. The dataset covers the edge of the modern shelf and slope extending c. 70 km along the shelf edge and 80 km down the slope. The slope trends in a N–NNW direction and water depths range from 80 to 90 m at the shelf edge to 2500 m on the lower slope (Fig. 1). The coastline, some 45 km inboard of the study area, is dominated by beaches, barriers, spits and lagoons. The spits are strongly aligned parallel to the coast indicating that both alongshore sediment transport to the NW and coastal wave processes dominate over any fluvial sediment transport at the present day (Biscara et al., 2012). The amount of sediment transport linked to longshore drift has been documented to be between 300,000 and 400,000 m 3/year (Bourgoin et al., 1963). This is consistent with an expected reduced fluvial sediment supply linked to Holocene sea-level rise and transgression. The Ogooué River enters the sea through Port Gentil which is located some 130 km to the North of the study area (Fig. 1). The location of the Ogooué Fan suggests that it has been located in broadly the same position since the Pliocene (Séranne and Nzé Abeigne, 1999; Wonham et al., 2000). However during periods of lower sea-level some of the distributary rivers of the Ogooué system may have supplied some sediment to the shelf directly landward of our study area. The modern slope profile is concave upward, similar to that of many other passive margins (e.g. Pratson and Haxby, 1996). The shelf edge occurs at about 80 m water depth and the slope gradient increases rapidly to a maximum of 4.5°, and decreases seaward to 2° on the mid slope, at water depths of c. 1500 m (Fig. 2). 2.1. Geological setting The continental margin of the Gulf of Guinea, from Cameroon to Angola, formed due to rifting and opening of the South Atlantic in Neocomian to Lower Aptian times. Syn-rift deposits are buried by mid-late Cretaceous transgressive sediments consisting initially of evaporites followed by platform carbonates. Since Senonian times the passive margin succession has been dominated by prograding clastic wedges, with periods of major uplift and canyon incision occurring during Eocene to Lower Miocene times (Séranne et al., 1992; Rasmussen, 1996; Wonham et al., 2000). The sediments of the Ogooué and Congo prograding deltas and slope fans form the majority of the Neogene to Holocene stratigraphy (Fig. 1 and Séranne and Nzé Abeigne, 1999).

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

Neogene terrigenous sediments between the two major fan systems are mainly fine-grained (Rasmussen, 1997). The studied gullies are located on the continental slope adjacent to the south-eastern edge of the Neogene-modern Ogooué Fan (Fig. 1; Séranne and Nzé Abeigne, 1999; Wonham et al., 2000). The edge of the most recently active fan system is imaged on the NW edge of the seismic dataset, and comprises several stacked channel-level complexes including a recently active sinuous deepwater channel that has a bathymetric expression at seabed. The gullies we describe here are unrelated to the Ogooué fan and they do not supply sediment to any major slope or basin-floor fans in the study area. Data from a commercial borehole located within the study area on the slope shows that the top 500 m of the stratigraphy is Pliocene to Holocene in age and comprises green–grey silty clays, with rare, thin beds of fine- to medium-grained sandstone bands rich in shell debris (Perenco, internal report). This borehole occurs in an area of parallel seismic reflections, with continuous strong amplitude events, which continue with little change over the entire gully field, suggesting that fine-grained sediments characterize the gullied interval. Additionally,

83

a recent study by Jobe et al. (2011) using 3D seismic reflection data from the slope offshore Equatorial Guinea, (located to the north of our study area) described a very similar seismic facies within the top few 100 ms TWT of their data. Core data to 120 m beneath seabed confirmed that this interval is dominated by clays with minor silt. 2.2. Previous work Compared to the extensive literature on the Congo fan and the Angolan margin to the south, the Neogene sediments of the Gabon passive margin are relatively poorly studied. Wonham et al. (2000) described a Lower Miocene upper slope and canyon fill underlying the modern Ogooué delta to the north of our study area, c. 40 km south of Port Gentil. Rasmussen (1994, 1997) and Séranne and Nzé Abeigne (1999) described and proposed alternative models for a suite of enigmatic, u-shaped, straight, 2–3 km wide channels that occur in the Miocene to recent slope sediments of southern Gabon, located some 250 km to the SE of our study area. An unusual feature of these channels is the margin parallel NW migration of their axes through time.

(a) 70 km

Sediment waves

2

1 Submarine channel Fan limit

Pock mark field

N

Ogouué Fan

(b) Distance (km) 0 Depth (m)

7 km

0

10

20

0

10

20

30

40

50

60

70

30

40

50

60

70

500 1000 1500 2000 2500

Gradient (°)

5

N

4 3 2 1 0 Distance (km)

Fig. 2. (a) Colour relief map of seabed bathymetry showing Areas 1 and 2; (b) Bathymetry and slope gradient profiles. Profile located on 3D relief image of Area 1 in (a).

84

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

Rasmussen (1994, 1997) called the channels ‘linear canyons’ and attributed their formation to erosion by sedimentary gravity flows, followed by deposition and subsequent infill during cycles of sea-level rise. In contrast Séranne and Nzé Abeigne (1999) called the same features ‘furrows’ and proposed that they are formed, and still maintained, by coastal upwelling-related bottom currents that form normal to the shelf edge. Fluid escape pockmarks at the seabed that are both randomly distributed and aligned in linear arrays have been described by Pilcher and Argent (2007) on the continental slope to the north of Port Gentil and in an area further to the north in Equatorial New Guinea. They argued that the aligned pockmarks form in response to slumping in an area of slope instability, which causes the pockmarks to be localized into linear arrays which ultimately coalesce to form pockmark gullies. 3. Dataset and methods A 3D time-migrated seismic reflection survey covering an area of 3200 km2 was used for this study. It has a bin size of 12.5 m × 12.5 m, which gives a maximum horizontal resolution of 12.5 m at the seabed, but which is likely to decrease to ~19 m in the shallow subsurface. The dominant frequency in the top 500 ms TWT of data is 40–45 Hz giving a vertical resolution of between 8 and 10 m close to the seabed, assuming a near seabed sediment velocity of 1700 ms −1. The seabed bathymetry (Fig. 2) has been extracted from the data by mapping the strong seabed reflection and depth converting it using a velocity for water of 1490 ms−1. Edge and dip magnitude maps were computed from the bathymetry map to assist in mapping more detailed features. These maps remove the effect of the large bathymetry range across the dataset and allow the smaller-scale features in the bathymetry to be illustrated in detail. Thicknesses quoted for the immediate sub-sea sediment units were calculated using sediment velocities of 1700 ms−1. The seabed gullies were mapped and measurements made of the main morphological characteristics, including width, depth, gradient and spacing. The widths, depths and gradients were measured along the lengths of nine representative gullies, which were selected for detailed analyses (Fig. 3). Spacing was measured across three transects at distances of 10, 20 and 30 km, respectively, from the shelf edge (Fig. 3). The seabed gullies and associated sediment waves vary in morphology and organisation along the slope. Here we describe them from two areas to illustrate these variations (Areas 1 and 2 in Fig. 2). 4. Results 4.1. Area 1-gullies and sediment waves 4.1.1. Gullies The seabed gullies are well-developed over an area of ~ 700 km 2. They occur as a set of dominantly straight and sub-parallel, evenlyspaced, gullies the longest of which extend some 60 km from the shelf edge (80 m) to the base of slope (2000 m) (Fig. 3). They are oriented perpendicular to the slope contours. Locally, in the north of the dataset, the shelf break and slope changes to a more north-westerly direction and in this area some of the gullies converge and join down dip as a function of this change in slope trend (Figs. 2, 3). In terms of their point of initiation three groups of gullies are recognised (Fig. 3): (1) Those that start at, or close to the shelf edge (blue); (2) those that start on the upper slope, c. 6 km from the shelf edge (green) and (3) those that start on the middle and lower slope, c. 10–20 km from the shelf edge (red). Consequently the gullies become more closely spaced and increase in density in a seaward direction. This relationship has not been documented in previous studies, and its significance is discussed later. Three seismic sections (Fig. 4) through the gully field illustrate the change in gully morphology down the slope. The gullies are 160–615 m wide and 8–43 m deep. Those closest to the shelf edge (Fig. 4(a), Section 1) are widest and deepest, and they become narrower and

shallower in a down-slope direction (Fig. 4(a), Sections 2 and 3). In cross-section the gullies have flat bases and gently dipping margins with slopes of ~10° (Fig. 4(b)). The spacing between gullies also decreases down-slope as new gullies (green and then red) infill the space between existing gullies (Figs. 3, 4). Measurements of width and depth were made on gully-perpendicular transects, 2 km apart, along nine representative gullies. For the same gullies the gradient along the gully axis was also measured. The results illustrate broadly similar width/depth characteristics between gullies, with most variation evident in the up-dip reaches of the gullies (Fig. 5). The gradient along the gully axes mirrors that of the general slope, with gradients of 3–4° in the uppermost reaches decreasing to 2° at 30 km from the shelf edge. Small (b0.5°) variations in the gradient profiles (e.g. at c. 7 km from shelf edge) are due to the presence of sediment waves with a positive relief in the gully thalweg. Gully spacing measured on three 20 km long transects at distances of 10, 20 and 30 km, respectively, from the shelf edge is illustrated in Fig. 5. The raw spacing data (Fig. 5(d)) shows that overall the gullies become more closely spaced with increasing distance down slope. The gullies on transects 2 and 3 (most down dip positions) are very regularly spaced, with an average spacing of 150–200 m for transect 3 (30 km from the shelf edge) and 300–400 m for transect 2 (20 km from the shelf edge) (Fig. 5(d), (e)). In contrast on transect 1, closest to the shelf edge, the gully spacing has a bimodal distribution and most gullies are spaced at around 300 m and 700 m. This suggests that geometrically, in the most distal reaches, the gully distribution is approaching saturation; i.e. spacing is just larger than gully width. Were the gullies to be any closer spaced, then one gully would capture another. It is difficult to be certain whether the gully bases are erosional. Some gullies (arrowed in Section 1, Fig. 4(a)) appear to truncate an otherwise continuous reflection at their base. However, it is also possible that the gullies are largely growing by lateral accretion of sediment on the flanks and that no vertical accretion occurs in the gully axes. Higher resolution data (for example bottom profiler data) would be required to confirm whether this is the case or not. The strong reflection from the seabed in the more distal sections obscures any smaller scale stratal reflections. 4.1.2. Sediment waves Three sets of asymmetric, upslope migrating sediment waves occur in Area 1. An example from two of the sets is illustrated here: one on the lower slope (Fig. 6) and the other on the upper slope (Fig. 7). For each example, two seismic lines are shown, illustrating how the same sediment wave located outwith the gully also occupies the gully axis. On the modern sea floor the average wavelength and relief of the upper slope sediment wave set are 200–300 m and 11–23 m respectively (Fig. 7). The set just beneath the shelf edge has wavelengths of 500–700 m, with relief of 9–15 m, while those at the base of the slope are longer (wavelengths 600–800 m) and have slightly less relief (7–13 m) (Fig. 6). Wave crests range in length from b1 km to 15 km (Fig. 3). The lower slope waves are strongly asymmetric with long, down-slope dipping limbs and shorter, upslope-dipping limbs (Fig. 6). The upper slope sediment waves are more symmetric, but both sets exhibit well-developed upslope migrating geometries. The upslope climbing bedforms are up to 60–70 m thick on the upper slope (Fig. 7) and 85 m thick on the lower slope (Fig. 6); thus thicker than the maximum relief of the gullies. In the upper and lower sediment wave fields, there are two stacked sets of sediment waves, with a total thickness of 130 m on the upper slope, and c. 200 m on the lower slope (see Fig. 7(b)). On the lower slope the two sets are laterally offset, with the uppermost set forming a lensoid shaped sedimentary package with a very high width-to-height aspect ratio and a maximum thickness of 85 m and lateral extent of at least 20 km. In general the sediment wave crests are straight to undulose and show little to no bifurcation along strike (Fig. 3). The sediment waves

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

85

(a) N

Fig. 6 Gullies used for measurements in Fig. 5

Fig. 7

Fig.

Fig. 8

4a-1

Fig. 4a-2

Fig. 4a-3

Approx. locations of transects for spacing measurements in Fig. 5 (d,e)

6 km

(b) N k re a fB el Sh

6 km

Gullies starting at/close to shelf edge Gullies starting on upper slope

Sediment wave troughs

Gullies starting on mid and lower slope

Fig. 3. (a) Seabed relief map of Area 1 (located in Fig. 2). Position of seismic sections and maps in figures shown. (b) Map of gullies and sediment wave fields in Area 1. Horizontal lines on the shelfal part of the seabed relief image are acquisition artefacts.

closest to the shelf edge are less continuous than the other sets. They are best developed within gully axes and are sometimes absent in intergully areas (Fig. 3).

4.1.3. Relationship between sediment waves and gullies An unusual feature of the sediment waves on this part of the Gabon continental slope is their relationship with the gullies. The

86

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

(a) 1 NW

SE

see (b)

75 m

2

75 m

1 km

3

75 m

V:H = 6:1

(b) 40m

400m

Fig. 4. (a) Seismic sections illustrating gully cross-sectional geometries at three positions on the slope; (1) close to shelf break, (2) mid-slope, (3) base of slope. See Fig. 3 for section locations. Vertical exaggeration of sections is approximately ×6. (b) Enlarged view of a single gully from section (1) (at box) shown with no vertical exaggeration.

gullies clearly incise into the sediments containing the sediment waves, suggesting that they post-date the sediment waves. However the presence of sediment waves crossing and occupying the gullies suggests that the sediment waves have continued to form and grow after the gullies have formed. Within the gullies the sediment waves often exhibit more irregular geometries with deeper troughs that locally form sub-circular depressions (Fig. 7(a), (c)). This may suggest that the dynamics of the currents forming the sediment waves is altered in some way by the local change in slope topography in the vicinity of the gullies. There are numerous places on the slope outside of the three main zones of sediment waves where trains of very small sediment waves are found restricted to within gullies (Figs. 7(c), 8). These sediment waves have wavelengths of between 300 and 500 m and form in the top 35 m or so of sediment. They thus appear younger than the larger sediment wave fields, and are either synchronous with, or post-date gully formation. Although individual migrating upslope dipping foresets are not resolved at the resolution of the data, the characteristic asymmetric sediment package, with long down-slope dipping limb and short steeper up-slope dipping limb located between the seabed and the base gully reflections, suggests that these are sediment waves. Despite the apparent similarity of the within-gully sediment wave troughs to the map pattern of the pock mark gullies described further to the north along the Gabon margin (Pilcher and Argent, 2007) their subsurface seismic expression is clearly very different (Figs. 7, 8) and there is no seismic evidence for fluid flow or faulting associated with the Gabonese sediment waves.

4.2. Area 2: gullies and sediment waves Area 2 also has two sets of gullies visible at the seabed; one on the upper slope, and the other located on the mid-lower slope and associated with sediment waves. The upper slope set resembles those identified in Area 1 (Fig. 9). Additionally, buried gullies within the shallow Plio-Pleistocene sediments also occur in Area 2. Both erosional and aggradation types are observed and are described further below. 4.2.1. Seabed gullies The two sets of seabed gullies have differing morphologies and geometries. Those on the upper slope are comparable to the Area 1 gullies, but they do not start at the shelf break, and they are shorter, with the longest gully only extending 6.5 km down slope (Figs. 9, 10(b)). The gullies range in width between 72 and 300 m and the maximum relief in the gully troughs is 22–47 m, exhibiting similar dimensions to those of Area 1 (Fig. 10). They also have erosional bases, with clear stratal truncations on the gully margins (Fig. 10(a)). Furthermore a seismic profile down a gully axis very clearly illustrates erosion along the gully base, with erosion of underlying strata (see arrows at reflection terminations in Fig. 10(b)). At the down dip end of the gullies, where they die out, incipient sediment waves are present (Figs. 9(a), 10(b)), as was seen in some places in Area 1. Lower on the slope there are another three gullies that are more irregular in form. These gullies vary in width from 250 to 1100 m and their present-day relief at the sea bed is only 20–40 m, because they are almost completely filled by a train of sediment waves (Figs. 9, 11).

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

87

(a) Width (m)

600 400 200 0

(b)

50

Depth (m)

40 30 20 10 0

(c) Gradient (°)

4 3 2 1 0

5000

10000

15000

20000

25000

30000

Distance from Shelf Edge (m)

(d)

1400 Transect 3

1200

Transect 1

Spacing (m)

1000 Transect 2 800 600 400 200 0 0

4000

8000

12000

16000

20000

Distance (m)

Cummulative Frequemcy (%)

(e)

100 90 80 70 60 50 40

Transect 1

30

Transect 2

20

Transect 3

10 0 0

200

400

600

800

1000

1200

1400

Spacing (m) Fig. 5. (a)–(c) Gully width, depth and gradient graphs. Gullies used for measurements are located on Fig. 3. (d) and (e) Gully spacing data for three 20 km long transects spaced at 10 km intervals from the shelf edge, where transect 1 is closest to shelf edge and transect 3 is furthest down slope.

The infilling sediment waves are restricted to within the gullies and are absent from inter-gully areas. The sediment wave troughs have the same characteristic sub-circular geometry as seen in some places in Area 1 (e.g. Fig. 7(a)). Two closely spaced seismic sections across the

gullies, which are located through an adjacent wave trough and crest respectively, illustrate the morphology in more detail (Fig. 11(a), (b)). The section through the wave trough shows the remnant gully morphology (Fig. 11(a)). However in the second section Gullies 1 and 2

88

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

1 km

(c) (b) (a) (b)

Outside gully

75 m

500m

(c)

Within gully

75 m

500m Fig. 6. Sections through sediment wave set on the lower slope. (a) Seabed dip map; blue high dip; white low dip. (b) Profile in intra-gully area. (c) Profile within a gully.

are entirely infilled by the sediment wave, whereas Gully 3 has some preserved relief (Fig. 11(b)). Most notable on these sections is that there are two gully sets that are generally co-aligned vertically through the section with the trough of the oldest gullies, occurring at ~250 ms TWT/~215 m beneath the sea bed. Gully 3 however has some lateral offset in trough axis from the underlying gully at the same location. Two sections located along the axes of Gullies 1 and 3 (seismic lines (c) and (d) in Fig. 11) illustrate that sediment waves infill the stacked gullies, with a preserved thickness of 200–220 m. The sediment wave bedforms display a characteristic asymmetric shape illustrating upslope migration (shorter and steeper up-dip limbs) and a crest-to-crest spacing of 700–1000 m (Fig. 11(c)). In contrast, the sediment wave internal geometries within Gully 3 appear to be less regular, especially towards the base of the gully system. This is a function of the fact that the lower gully in this location is offset from the youngest/modern gully so the section is not passing through the centre of the lower set of sediment waves. When examined in cross-section these gullies show evidence for both aggradation of the sediment on the gully flanks and some minor erosion on the gully walls and bases, as well as the onlapping reflections from the sediment wave fill (Fig. 11(a), (b)). These geometries suggest a complicated interaction between (1) erosion to form the initial gully, (2) enhanced deposition on the gully flanks with no, or less, deposition in the gully axes to maintain the gully form through time, and (3) either concurrent, or final, infill of the gully by upslope migrating sediment waves. In the area of Gullies 1–3 this process occurred twice within the Plio-Pleistocene to recent succession.

4.2.2. Buried gullies Buried, infilled gullies are also found in Area 2. A section parallel to the strike of the upper slope (Figs. 9, 12(a)) shows that these gullies have similar dimensions to the erosional gullies expressed at the seabed to the north in Area 1. A further set of mixed aggradational/erosional gullies, more akin to seabed gullies 1–3 is found even deeper in the succession (between 0.7 and 0.9 s TWT on Fig. 12(a), (c)) Three seismic units are distinguished in the strike parallel section (Fig. 12(a), (b)): Z (oldest), Y, and X (youngest). The buried gullies are located within units Z and X. A seismic profile perpendicular to the slope (Fig. 12(c)) illustrates the spatial relationships between the seismic units. The gullies within Unit X erode into recent shelf-edge-slope clinoforms, which we suggest formed during a near recent (last?) sealevel lowstand, prior to the present Holocene highstand. There is an erosional boundary—surface y′, within Unit Y that marks the base of the Unit X prograding sequence. A set of older clinoforms forms the underlying Unit Z. The Unit X gullies exhibit a clear erosional morphology (Fig. 12(a), (b)). They have widths and maximum reliefs that fall within the range measured for the seabed gullies in Areas 1 and 2 (i.e. widths from 220 to 400 m and a maximum relief of 20–30 m). The lack of vertical stacking of gully troughs suggests that the Unit X gullies were not long-lived. It appears that a set formed, was subsequently in-filled and then another cycle of gully incision began. There are no sediment waves associated with the Unit X gullies. In contrast, Unit Z is characterized by two or three sets of stacked gullies. Mapping of the trough axes, such as those at reflections ‘a’ and

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

89

(a)

(b) (c) 2 km

(b)

seabed sed. waves older waves

Outside gully

75 m

750m

(c)

Within gully

Waves restricted to within gully 75 m

750m Fig. 7. Sections through sediment wave set on the upper slope. (a) Seabed dip map; blue high dip; white low dip. (b) Profile in intra-gully area. (c) Profile within a gully.

‘b’ on Fig. 12(b) and (c), confirms this. The troughs at ‘b’, for example, have a longer down-slope length than the overlying troughs at ‘a’. The gullies at the level of reflection ‘a’ have typical widths of 400–600 m and can be traced 2–3 km in a down-slope direction. The gullies near the base of Unit Z (i.e. marker ‘b’ on the section) have widths of 275–410 m with a down-dip length of 3–4 km. Maximum gully relief is ~30 m. Gullies within Unit Z appear to aggrade through time; this is most apparent around 0.8 s TWT. However, the upward growth is not as a result of a single phase of trough incision near the base of the unit followed by deposition, as suggested by the complexity of the reflections within this interval. Gully flanks and troughs aggrade at different rates, but there is also truncation of reflections on gully margins and bases. Gully troughs are also laterally offset from underlying gullies and some, such as those at ‘c’ and ‘d’ on Fig. 12(b), have been completely infilled. It is thus evident that multiple phases of gully erosion and infill occurred throughout the deposition of unit Z and it is only at the top of the unit that significant infill and drape of remnant topography are observed. Throughout this older progradational succession, gullying has been a long-lived process on the upper slope, with flows that varied spatially and temporally between erosion and deposition. In contrast to the modern aggradational gullies in Area 2 there are no sediment waves associated with the Unit Z gullies.

5. Interpretation and discussion 5.1. Gully geometry The seabed and buried gullies show some similarities to other smaller u-shaped gullies previously described in the literature. However we document a greater variety in gully morphology and organisation than those described previously from northern California or the Tyrrhenian Sea and the Gabon gullies are the first to be described with related sediment wave fields. The relatively simple, self-similar upward growth of the gullies as seen in Californian and Italy appears less prevalent in the modern examples in Gabon. This might be due to the fact that the system is relatively young and/or is forming too slowly to show significant aggradation due to sediment starvation. However the lack of clearly visible aggradation in the Area 1 gullies could also be a function of the vertical resolution of the seismic data. The Tyrrhenian and Californian examples were imaged with high-frequency, and hence higher resolution, seismic techniques (Sparker and CHIRP), and the gullies themselves have less relief (max 10 m) than the Gabon examples. Buried gullies in the Tyrrhenian area, at c. 100 m below the seabed, have similar relief to the seabed examples in the same area (Chiocci and Normark, 1992).

90

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

(a) (b)

(c)

1.5 km

(b)

Within gully

85 m

750 m

(c)

Outside gully

85 m

750 m Fig. 8. Sections through sediment waves within gullies. (a) Seabed dip map; blue high dip; white low dip. (b) Profile within a gully. (c) Profile in intra-gully area; note that sediment waves are restricted to within the gully.

The vertical resolution of commercial 3D seismic data, such as that used in our study, which typically has frequencies in the 10–80 Hz range (mean 40–45 Hz), means that aggradational geometries at the scale of the Californian or Tyrrenhian examples would not be imaged. We thus cannot preclude that the Gabon gullies in Area 1 do not exhibit aggradational geometries at a scale beneath the seismic resolution of our dataset. The larger Area 1 gullies (Figs. 3 and 4), the shorter seabed gullies in Area 2 (Fig. 10), and the shallowly buried gullies in Area 2 (in Unit X. Fig 12) all exhibit erosion at the gully bases and flanks. The second set of partially buried Area 2 gullies (gullies 1, 2 and 3; Fig. 9) and the fully buried Area 2 gullies within Unit Z (Fig. 12) are more complex in morphology, with mixed erosion, irregular aggradation and even local infill and drape, and consequently are somewhat more similar to the other documented examples of u-shaped gullies.

The second unusual feature of the Gabon gullies in Area 1 is their plan view organisation and the increase in density and resultant reduction in spacing that occur in a basinward direction. Both the Californian and Tyrrhenian gully systems are shorter, and extend for distances of less than 10 km across the slope, with no change in density in a downslope direction. The Tyrrhenian examples all start at approximately the same isobath, while the Californian examples show some variation in gully start positions on the slope, which varies between c. 250 m and 400 m isobaths (Spinelli and Field, 2001). Additionally some of the Californian gullies terminate between 400 and 500 m but others continue down slope and appear to stop abruptly, although this may be due to data truncation. Buried gullies in the same location show a broadly similar planform organisation with a similar spacing on differing mapped surfaces (Spinelli and Field, 2001).

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

91

(a)

sediment waves

(b)

N

Fig 1

Fig 10b

2a c

Fig 12

G1 Fig 11c

G2

0a

Fig 1

Fig 11d

G3 Fig 11a

Fig 11b

4 km

Fig. 9. (a) Seabed dip map of Area 2. (b) Dip-map overlain with seabed contours of seabed; Location of seismic section in Figs. 10–12 is shown.

5.2. Gully formation processes The frequently observed truncation of reflections on gully flanks and bases suggests that erosion is an important process in gully formation and the most likely agent of erosion on a submarine slope is sediment gravity flows. The absence of any evidence for slumping at gully headwalls discounts the possibility that the sediment gravity flows are triggered by slope failure events, as they occur in v-shaped systems. It would therefore appear that individual, narrow, turbidity currents, restricted to single gullies, were not the dominant erosive process within the Gabonese slope gullies. Borehole data through the slope succession shows it to be dominated by silty mudstones and there appears to be no coarser clastic apron developed at the base of the gully system. This implies that the turbidity currents responsible for gully formation were dominantly fine-grained, and may have been wide, sheet-like sediment gravity flows that formed a suite of gullies simultaneously. The location of the gullies down-dip of the shelf edge suggests that the fine-grained turbidity currents most likely originated on the shelf. Hemipelagic deposition is an important process in some muddominated slope settings, and has been invoked as the mechanism responsible for the aggradation seen in the gully systems in the

Tyrrenhian Sea and Northern California. However, laterally extensive horizons that drape pre-existing topography which typifies hemipelagic deposition are absent from both the modern gullies, and the shallowly buried erosional gullies (e.g. Fig. 12(a), (b); Unit X) in Gabon. Infill and drape of gullies do occur, but it appears to be very spatially localized as is seen in Gullies 1–3 in Area 2 and the Unit Z buried gullies (Fig. 12(a), (b)). Additionally, the geometries of the draping reflections in the aggrading gullies are irregular, varying in thickness from gully flank to trough (see Fig. 11(a), (b); Fig. 12(a), (b)) ruling out hemipelagic drape as a mechanism for gully aggradation. We therefore conclude that the aggradational geometries must be an intrinsic part of the gully formation process, possibly representing lateral variations in the flow velocity within the sheet-like turbidity current, transverse to the current flow direction. The theoretical works of Hall et al. (2008) and Fedele and Garcia (2009) have shown that instabilities in a sheet-like turbidity current result in evenly spaced perturbations with higher flow velocities and basal shear stresses occurring adjacent to regions of lower velocities and basal shear stresses in the cross-stream direction. Thus flow variations within the turbidity current itself could account for the variety of gully geometries observed on the Gabon margin. It seems most likely

92

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

line (b) intersection

(a)

S

N

85 m

750 m

E

W

(b) line (a) intersection

1000 m

85 m

Fig. 10. (a) Seismic profile parallel to slope strike illustrating seabed gullies in Area 2; (b) Intersecting seismic profile down the length of the a seabed gully. Erosion of underlying reflections is evident in the gully thalweg. Arrows mark position of reflections that have been truncated by erosion along the gully axis. Blue vertical lines on (b) are boundaries of seismic profile segments where digitized in 3D volume.

that gullies initiated due to a wave-like perturbation in the turbidity current and that erosion, or simply no deposition, occurred in the perturbation ‘troughs’ where higher velocities, basal shear stresses and sediment concentrations occur (Fig. 13). For the purely erosional u-shaped gullies the resultant higher flow velocities in gully axes were large enough to erode, and the inter-gully flow velocities appear also to have been high enough to allow little to no-sedimentation in the inter-gully areas (Fig. 13(b)). For gullies in which little or no erosion in the gully axis is evident the resultant higher flow velocities in gully axes exceeded sediment settling velocity and there was no deposition within the gully; there may however have been some deposition on the gully flanks. At lower flow velocities there could be some deposition within the gully and enhanced sediment deposition on gully flanks which, through time, led to the formation of constructional bedforms characterized by net-deposition and aggradation such as the gullies seen in Area 2 (Fig. 13(c)). Variation in current strength between subsequent flows can lead to a range of differing stratal geometries, such as those expressed in the partially buried gullies in Area 2 and the buried Unit Z gullies in Area 2. Periods of lower overall flow velocities may lead to some aggradation, with enhanced deposition on the gully flanks. Subsequent faster flows may promote erosion and truncation of underlying strata, but with the gully axes remaining in broadly the same location. Flows with different velocities or widths may set up a new gully set with a different spacing. We note that this may also have been a shortlived process, as once the gully morphology was established, then self-channelling flows could have maintained the morphology through time. The appeal of the proposed sheet-like turbidity current model is that variations in current velocity or concentration can explain both the stratal geometries and the regular spacing of the gully systems. The mathematical linear stability analyses of a turbidity current by Hall et al. (2008) predict that current heights of between 10 and 100 m generate the most amplified current perturbations with a wavelength of between 250 and 2500 m, resulting in gully spacings in the same range. This would imply that the Area 1 seabed gullies, with spacings ranging from 150 to 700 m could have been formed by turbidity currents less than 100 m high. The change in down-slope gully spacing remains difficult to explain. The apparent saturation in spacing basinward suggests that geometrically at least the gully distribution is approaching some sort of equilibrium or limiting case. But, it is not clear what, if anything, this tells us about the physics of the flows responsible for the gully system. If instead

we think of the gully spacing as a function of the turbidity current height as suggested by the theoretical work of Hall et al. (2008), then the reduction in spacing might suggest that the flows are thinning in a downstream direction. Alternatively there may be some other aspect about the structure of the ocean water column, such as salinity or thermal changes, that forces a change in the dynamics of the turbidity currents as they meet an oceanographic boundary. Additionally, our work suggests that u-shaped gullying can be a long-lived process on continental slopes (c.f. the buried examples in Area 2). The variation in morphology of the gullies both through time, and spatially at any one time, indicates that the sediment gravity flows responsible for forming gullies can vary between an erosional, net-depositional or mixed mode. For example in Area 2 the oldest gullies are dominantly aggradational, and were followed by a shallower younger set, also buried, that are erosional. Yet on the present day seabed, Area 1 is dominated by erosional gullies which co-existed with a set of aggradational gullies in Area 2. Within an even shorter distance in Area 2, erosional gullies are forming up-dip of aggradational gullies. 5.3. Sediment waves Sediment waves of the type and dimensions observed on the Gabon slope are generally attributed to either turbidity currents or bottom currents (Wynn and Stow, 2002). All sediment waves are transverse bedforms with crests aligned perpendicular to the current flow direction. Most fine-grained sediment waves that form due to bottom current action occur as sediment drifts on the deep ocean floor or on the lower slope (e.g. Stow et al., 2002). Where bottom current-related sediment waves form on slopes, their crests are often aligned at low-angles (10–50°) to the regional slope contours. These oblique trending sediment waves mostly migrate both up-slope and up-current (Wynn and Stow, 2002). In contrast turbidity current-related sediment waves are typically found on slopes and deep-water channel levees. Their crests are aligned broadly parallel to the regional bathymetric slope, except on spill-over levees where the crests can form at between 90° and 45° to the direction of the turbidity current in the main channel (Migeon et al., 2001). The waves are observed to migrate both up-slope and up-current (Wynn and Stow, 2002). In Gabon the sediment waves are aligned parallel to the slope contours, which is indicative of a turbidity current, rather than a bottom current, origin. Were the Gabon examples to have formed from bottom currents then the occurrence of sediment

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

93

N

(a)

S

gully 1

gully 2

gully 3

1.6

1.8

base of sediment wave package

ms (twtt) N

(b)

1000 m

intersection with (c)

intersection with (d)

S

1.6

1.8 ms (twtt)

(c)

W axial section along gully 1

E intersection (b) (a)

1.4

1.6

1.8 ms (twtt)

(d)

1000 m

W axial section along gully 3

intersection (b)

l intersection (a)

E

1.4

1.6

1.8 ms (twtt) Fig. 11. (a), (b) Seismic sections across gullies 1, 2, and 3 (G1, G2 and G3) in Area 2. Section (a) is located within a sediment wave trough; and section (b) along the adjacent sediment wave crest. Green triangles mark the base of the sediment wave package. (c), (d) Sections down the axis of gullies 1 and 3 respectively. The dashed green line indicates the base of the sediment wave package within the gullies.

waves at depths of only 200–300 m water depth just beneath the shelf edge would require a very long, up-dip directed current. An up-slope directed upwelling current has been proposed by Séranne and Nzé Abeigne (1999) for an area at about 4°S, some 250 km to the south of our study area. These authors have suggested that Coriolis-induced Ekman effects deflect the N-directed, shallow-water, wind-driven current offshore, which is compensated for by deeper waters being drawn upslope in an onshore direction. We see several problems with invoking this model further to the north in Gabon; firstly, the wind-related longshore currents at most affect the water mass to a maximum of 200 m and Ekman effects are likewise limited to this depth (Brown et al., 2001); but more importantly the Coriolis force at the Equator reduces to zero and hence Coriolis-related Ekman transport will be very weak, to insignificant in our study area (at 2°S). Secondly, the ocean currents offshore Gabon are more complicated than that inferred by Séranne and Nzé Abeigne (1999). The Angola current

that affects the Gabon shelf is a shallow shelfal current that, in fact, flows south offshore Gabon and is limited to 250–300 m in depth (Moroshkin et al., 1970; Stramma and Schott, 1999). The major zone of coastal upwelling that draws up Antarctic intermediate water in the eastern Atlantic ocean is restricted to low latitudes offshore W Africa, S of 20°S (e.g. Hay and Brock, 1992; Lass et al., 2000). Thirdly, the deeper ocean currents that occur between 500 and 1200 m water depth offshore equatorial Gabon form part of the complex Equatorial gyre which has both E- and W-directed currents in the region between 5°N and 5°S, but none of which appear to come landward enough to affect the Gabon slope (Stramma and Schott, 1999). Furthermore the fact that some of the sediment wave trains in both the northern and southern parts of the study area are restricted to within gullies, might suggest that at least some of the sediment waves are related to the currents flowing within the gullies (i.e. downslope flowing gravity currents). It therefore seems more probable that the sediment waves described in

94

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

(a)

S

N

0.5

0.7

0.9

s (twtt) N

(b)

intersection with (c)

1 km

S

sea

X 0.5 Y y’ 0.7 a

Z

c

d

0.9

b

s (twtt)

(c)

E

intersection with (a,b)

W

sea

0.2 unit y

unit x

0.4 0.8

1.0

s (twtt)

y’

a

0.8

unit z

b

1 km Fig. 12. a) Seismic profile parallel to slope strike illustrating near seabed buried gullies in Area 2. (b) Line-drawing of seismic section in (a) identifying three seismic units, X, Y and Z. Blue dots indicate gully troughs. (c) Line drawing of a slope parallel seismic line that illustrates the geometries of the seismic units identified in (b). Note how unit X represents a younger prograding shelf-slope sequence.

this study area are turbidity current upslope-migrating sediment waves (sensu—Wynn et al., 2000a; Wynn and Stow, 2002).

discount the possibility that the seabed gullies and sediment wave fields were and continue to be active during the Holocene.

5.4. Turbidity current origin

5.5. Relationship between gullies and sediment waves

Both the bedform types documented in our study can be formed by sheet-like turbidity currents that may have been up to 100 m thick and perhaps tens of kilometres wide. These currents presumably originated on the shelf, possibly at times of lower sea-level prior to the Holocene. The evidence from the current shoreline morphology is that most of the sediment is being trapped on the shelf, near to the shoreline and reworked into long-shore bars and spits. However large storms may periodically transport muddy sediment, deposited in times of lower sea-level, to the shelf edge and trigger turbidity currents that spill over the shelf-edge. Alternatively some of the finer grained sediment fraction being carried in the alongshore currents on the shelf may spill over the shelf edge forming turbidity currents. We therefore cannot

The relationship between the gullies and sediment waves is puzzling, and the spatial link we describe in Gabon between gullies and sediment waves has not been documented previously. Turbidity current sediment waves have been reported from areas swept by unconfined turbidity currents such as on channel levees (Normark et al., 1980), flanks of volcanic islands (Wynn et al., 2000b), on continental slopes in association with large river-sourced turbidite systems (Ercilla et al., 2002), associated with sediment gravity flows fed through channels and erosional gullies up-slope of the wave field (Gong et al., 2012), and more rarely confined to within channels (Kidd et al., 1998; Wynn et al., 2000a; Heinio and Davies, 2009). An examination of the seabed bathymetry alone in Area 1 might suggest that the sediment wave field is forming on a gullied

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

(a)

Front of f low

f flo w To p o

H

H H

~50-600 m

(b) ~5-30 m

~250-1000 m

(c) ~ 30 m

Fig. 13. Turbidity current model for gully formation. (a) Schematic diagram illustrating geometry of gullies and perturbation of flow velocities within a sheet-like turbidity current. The large arrows within the gully troughs represent the regions of faster flow with either erosion or low sediment deposition rates; smaller arrows in inter-gully areas represent slower velocities and areas of sediment deposition. The streamlines show the transverse perturbation of the velocity field within the flow, with arrows showing flow directions, according to the theoretical analyses of Hall et al. (2008). Schematic diagrams illustrating stratal geometries in (b) gullies dominated by erosional flow in the turbidity current and (c) aggradational gullies with deposition on gully flanks and within troughs. Vertical scale greatly exaggerated. Dimensions given are representative of Area 1 for the erosional gullies and Area 2 for the aggradational gullies. These diagrams are proposed end-members of the spectrum of gully types documented on the Gabon slope.

seabed surface because wave crests can be traced continuously across the slope over large areas. However the seismic data show that the depositional package containing the sediment waves is thicker than the relief of the gullies indicating that sediment waves existed on the slope prior to the formation of the gullies, and that the sediment waves have subsequently continued to form. In Area 2 a close temporal relationship between Gullies 1–3 at the seabed and sediment waves is very evident. Gullies 1–3 persist through 220 m or so of stratigraphy as do the associated sediment waves. Furthermore this has been the case for a significant length of time as two sets of upward aggrading gullies, both filled with sediment waves, have formed in this location. The question then arises as to whether the same current responsible for forming gullies can also deposit sediment waves. Or do variations in the dynamics of the turbidity currents crossing the slope mean that some currents form gullies, while others deposit sediment waves? At times it would seem that conditions favour sediment wave deposition over gully formation and gullies are infilled by sediment waves as appears to be happening at the seabed in the case of Gullies 1–3 in Area 2. There are two other interesting features associated with the sediment waves; firstly the occurrence of the main sediment wave fields in three discreet bands on the slope in Area 1 and, secondly, the formation of sediment waves restricted to within gullies (e.g. Fig. 7(c) in Area 1; Gullies 1–3; Area 2). There is no obvious morphological feature on the slope, such as a gradient change for example, that coincides with the location of the lower, middle and upper slope sediment wave fields which occur at c. 1200–1000 m, 500–600 m and 300–200 m water depths respectively. Nor do the location of the sediment waves fields correspond

95

with the onset position of the infill gully sets, except for the middle set of sediment waves (at 500–600 m) which are broadly coincident with the start of the second set of gullies (green gullies in Fig. 2). Once a topographic irregularity occurs on the seabed and the wave-like morphology has started to form, one can then envisage how the bedform can continue to grow and migrate up slope. Any downslope-flowing current will decelerate as it encounters positive bathymetric relief resulting in greater sediment deposition from suspension on the upslope flank of the wave, and sediment bypass or erosion on the downslope flank. However the mechanism that triggers the formation of the original bedform remains much more problematic, although some recent works suggest that cyclic steps in a net-depositional, Froude-supercritical flow may give rise to trains of upstream-migrating sediment waves (Fildani et al., 2006; Cartigny et al., 2011; Kostic, 2011). Further exploration of these ideas is outside the scope of our work. We suggest that there may be other hydrodynamic factors that are important in the study area. For example salinity-related density variations or thermal stratification in the ocean water masses may affect the dynamics of the downslope flowing currents. Stramma and Schott (1999) describe five layers in the tropical Atlantic ocean between 20°N and 20°S. One of the boundaries between two layers, the upper Circumpolar Deep Water and cold North Atlantic Deep water, occurs at about 1200 m. Another important density boundary between South Atlantic Central Waters and underlying colder and fresher Antarctic Intermediate waters exists at ~500 m. Interestingly, these depths do coincide with the slope bathymetries at which the lower two sediment wave fields are located, which might lend some support to the idea that structure within the water mass itself may affect the behaviour of the sheet-like turbidity currents travelling down the slope, and thus influence the location of the sediment wave fields. However further detailed data on the water masses above the Gabon slope with which to explore this hypothesis are lacking, and existing maps of water masses (Stramma and England, 1999; Stramma and Schott, 1999) are acknowledged by the authors as not of being high resolution because of sparse control data. Data from other continental margins suggest that this hypothesis is at least plausible. For example, studies of currents in the Gulf of Cadiz have shown that the downslope limit of sediment wave fields coincides with where the Mediterranean outflow water lifts off the seafloor when it becomes neutrally buoyant due to mixing with the surrounding water mass (Habgood et al., 2003). Very short sediment waves that are restricted to within a channel are not as common as extensive fields of sediment waves. Where such sediment waves have been previously described they appear to be mostly dominated by coarse-grained sandy or gravelly turbidites, and have wavelengths up to 1 km and heights of up to 10 m (Wynn et al., 2000a; Wynn and Stow, 2002). Channel-confined sediment waves described by Wynn et al. (2000a) from the El Julan channel on the Western Flank of the Canary islands and those in buried Neogene submarine channels in the Espirito Santo Basin on the Brazilian continental margin (Heinio and Davies, 2009), have wavelengths and heights in a similar range to those of the sediment waves restricted to within the Gabon gullies. The El Julan channel sediment waves have wavelengths of 400–1200 m, with wave heights of 6 m and the Brazilian examples have wavelengths of 200–600 m, and typical heights of ~20 m. However the channels within which the Brazilian and Canary Island examples occur are substantially larger than the Gabon gullies (c. 0.7 km wide Brazil; 2 km wide El Julan), and they are either part of, or feed into a larger channel-levee system. The Brazilian examples are more complex in that they have pronounced sub-circular troughs (a feature of some of the Gabon examples, particularly in Area 2) but are then infilled by a mounded, onlapping fill between episodes that are characterized by the formation of sediment wave/asymmetric depressions within the same channel. Both Wynn et al. (2000a) and Heinio and Davies (2009) conclude that the sediment waves that they document are formed by turbidity current sediment gravity flows that are restricted to the channels. Heinio and Davies (2009) propose that flow instabilities

96

L. Lonergan et al. / Marine Geology 337 (2013) 80–97

or undulating topography within the channel leads to the development of wave-like bedforms in the channel. These waves are then amplified by preferential deposition on the shallower dipping up-slope flanks and erosion on the steeper down-slope flanks. Subsequently, the depressions between the waves deepen and become more circular, with more deposition on the lee side, but they note that the exact mechanism for the change from asymmetric wave to more circular trough depression is not understood. The presence of sediment wave trains restricted to within gullies on the Gabon margin may therefore indicate that at times some of the turbidity currents on the Gabon margin were much smaller than the flows responsible for the gullies themselves or the large sediment wave fields, and were restricted to within the gullies. 5.6. Comparison with other studies The Gabon aggradational gullies exhibit some similarities to recently described, long-lived aggradational ‘canyon’ systems in mud-dominated slope settings in Equatorial New Guinea (Jobe et al., 2011) and offshore Brunei Darussalam (Straub and Mohrig, 2009), albeit at a much smaller scale. The channels on the Equatorial New Guinea slope are broadly linear and extend for 50 km downslope in an area of low sediment supply. They are u-shaped, 1–2 km wide, with a present-day seabed relief of up to 150 m and spacing of c. 5 km. They are aggradational (> than 800 m of aggradation since the late Miocene), and they do not feed an associated sediment apron or fan on the lower slope or basin floor. Jobe et al. (2011) propose that the canyons were formed by thick, dilute, muddy turbidity currents, generated by storm waves on the shelf or increased river run-off which was interspersed with hemi-pelagic deposition. The Bruneian examples cut through a growing shale ridge, oriented parallel to the slope trend, in ~900 m water depth, some 20 km down dip of the present day shelf edge. The three canyons are 1–2 km wide and have a maximum relief of 165 m. Higher rates of sedimentation in the inter-canyon areas resulted in the preservation and upward aggradation of the canyons with time, and Straub and Mohrig (2009) suggest that the canyons were formed by sheet-like turbidity current flows that extended for many kilometres in the strike directions. Although the Bruneian examples are somewhat unusual in that the presence of the shale anticline exerts the primary control on the presence of canyons, both groups of authors invoke similar processes for the net-depositional canyons observed in the two areas. All of these recent discoveries of wellimaged net-depositional gully and canyon systems on mud-dominated continental slopes suggest a continuum of broadly similar features from b50 m to 2 km scale, whose formation is strongly controlled by the action of muddy, sheet-like turbidity currents. 6. Conclusions We have described a wider variety of slope gullies than hitherto documented over a large area of the Gabonese mud-dominated continental slope. The u-shaped gullies are of both aggradational, erosional and mixed aggradational/erosional form and occur in close proximity both at the modern seabed and buried within the Plio-Pleistocene to Holocene sediments, indicating that gully formation has been an ongoing process through recent geological time on the Gabon continental margin. The modern slope-gullies that continue for distances of up to 60 km down slope are closely associated with fields of upslope migrating sediment waves. Shorter sediment waves restricted to within gullies are also described for the first time, although sediment waves within submarine channels have been described before (e.g. Wynn et al., 2000a; Heinio and Davies, 2009). We have presented evidence to suggest that the two morphological bedforms, i.e. gullies and sediment waves, could be formed by sheetlike, relatively thin, sediment gravity flows dominated by fine-grained suspended load flowing down the slope. It remains unclear as to what aspect of the current flow dynamics promotes the formation of gullies as opposed to sediment waves, and whether other external factors,

such as density or thermal stratification of the water column might be important in triggering sediment wave formation. Recent mathematical analyses of sheet-like turbidity currents predict that wave-like perturbations in a cross-flow direction generate evenly spaced regions of higher and lower flow velocities (Hall et al., 2008; Fedele and Garcia, 2009), and such models provide an elegant mechanism for which slope gullies such as those we describe can form. In particular they predict the regular spacing that is typical of the gully sets. In erosional or non-constructional gullies (Area 1 and Area 2) the gully axes experienced flow velocities that mainly exceeded the settling velocity of the sediment in suspension, and no deposition occurred. In contrast the net-depositional gullies (Area 2) indicate lower flow velocities with sediment deposition both within the gully axes and on the gully flanks. Vertical variations in erosional versus depositional geometries are observed in some gully sets (e.g. Area 2) suggesting that successive flows can experience fluctuations in velocity or sediment concentration leading to interspersed erosional and depositional events. Theoretical analyses of Hall et al. (2008) predict that gullies with spacings in the range of those measured on the Gabon seabed (150–700 m) could form by turbidity currents that were up to 100 m thick. We propose that u-shaped gully systems are more likely to form on low-sediment supply continental margins dominated by sheetlike fine-grained muddy turbidity currents where entrainment of finegrained sediments from the continental shelf may be triggered by storm driven resuspension. Acknowledgements The seismic data for the study was provided by Perenco and we acknowledge the permission of Perenco Oil and Gas Gabon, Tullow Oil Gabon S.A. and Direction Generale des Hydrocarbures to publish the seismic data. NHJ was funded by a Malaysian government scholarship. We thank Landmark Graphics for a University Software Grant. LL acknowledges very helpful discussions with Dan Orange, John Decker and Phil Teas, and the added insights that data in the Timor Trough provided for some of the ideas developed herein. LL also thanks Peter Haughton for suggesting the potential importance of ocean water mass structure. We thank reviewers R. Wynn and T. Mulder for helpful comments and suggestions. References Amante, C., Eakins, B.W., 2008. ETOPO1 1 Arc-Minute Global Relief Model: Procedures, Data Sources and Analysis. National Geophysical Data Center, NESDIS, NOAA, U.S. Department of Commerce, Boulder, CO (August 2008. http://www.ngdc.noaa.gov/ mgg/global/relief/ETOPO1/docs/ETOPO1.pdf). Biscara, L., Hanquiez, V., Leynaud, D., Marieu, V., Mulder, T., Galissaires, J.-M., Crespin, J.-P., Braccini, E., Garlan, T., 2012. Submarine slide initiation and evolution offshore Pointe Odden, Gabon—Analysis from annual bathymetric data (2004–2009). Marine Geology 299–302, 43–50. Bourgoin, J., Reyre, D., Magloire, P., Krichewsky, M., 1963. Les canyons sous-marins du Cap Lopez (Gabon). Cahiers Océanographiques, 15ème Année 6, 372–387. Brown, E., Colling, A., Park, D., Phillips, J., Rothery, D., Wright, J., 2001. Ocean Circulation, second ed. Open University and Butterworth-Heinemann. Cartigny, M.J.B., Postma, G., van den Berg, J.H., Mastbergen, D.R., 2011. A comparative study of sediment waves and cyclic steps based on geometries, internal structures and numerical modeling. Marine Geology 280, 40–56. Chiocci, F.L., Casalbore, D., 2011. Submarine gullies on Italian upper slopes and their relationship with volcanic activity revisited 20 years after Bill Normark's pioneering work. Geopshere 7, 1284–1293. Chiocci, F.L., Normark, W.R., 1992. Effects of sea-level variations on upper-slope depositional processes offshore of Tiber delta, Tyrrhenian Sea, Italy. Marine Geology 104, 109–122. Coleman, J.M., Prior, D.B., Adams, J.R., 1981. Erosional furrows on continental shelf edge, Mississippi delta region. Geo-Marine Letters 1, 11–15. Damuth, J., Flood, R.D., Kowsmann, R.O., Belderson, R.H., Gorini, M.A., 1988. Anatomy and growth of Amazon deep-sea fan revealed by long-range side-scan sonar (GLORIA) and high resolution seismic studies. AAPG Bulletin 72, 885–911. Dowdeswell, J.A., Ó Cofaigh, C., Noormets, R., Larter, R.D., Hillenbrand, C.-D., Benetti, S., Evans, J., Pudsey, C.J., 2008. A major trough-mouth fan on the continental margin of the Bellingshausen Sea, West Antarctica: the Belgica Fan. Marine Geology 252, 129–140. Dyer, K.R., 1970. Linear erosional furrows in Southampton Water. Nature 255, 56–58.

L. Lonergan et al. / Marine Geology 337 (2013) 80–97 Ercilla, G., Alonso, B., Wynn, R.B., Baraza, J., 2002. Turbidity current sediment waves on irregular slopes: observations from the Orinoco sediment-wave field. Marine Geology 19, 171–187. Fedele, J.J., Garcia, M.H., 2009. Laboratory experiments on the formation of subaqueous depositional gullies by turbidity currents. Marine Geology 258, 48–59. Field, M.E., Gardner, J.V., Prior, D.B., 1999. Geometry and significance of stacked gullies on the northern California slope. Marine Geology 154, 271–286. Fildani, A., Normark, W.R., Kostic, S., Parker, G., 2006. Channel formation by flow stripping: large scale scour features along the Monterey East Channel and their relation to sediment waves. Sedimentology 53, 1265–1287. Flood, R.D., 1983. Classification of sedimentary furrows and a model for furrow initiation and evolution. Geological Society of America Bulletin 94, 630–639. Gong, C., Wang, Y., Peng, X., Li, W., Qiu, Y., Xu, S., 2012. Sediment waves on the South China Sea Slope off southwestern Taiwan: implications for the intrusion of the Northern Pacific deep water into the South China Sea. Marine and Petroleum Geology 32, 93–109. Habgood, E.L., Kenyon, N.H., Masson, D.G., Akhmetzhanov, A., Weaver, P.E., Gardner, J., Mulder, T., 2003. Deep-water sediment wave fields, bottom current sand channels and gravity flow channel-lobe systems: Gulf of Cadiz, NE Atlantic. Sedimentology 50, 483–510. Hall, B., Meiburg, E., Kneller, B., 2008. Channel formation by turbidity currents: Navier–Stokes-based linear stability analysis. Journal of Fluid Mechanics 615, 185–210. Hay, W.M., Brock, J.C., 1992. Temporal variation in intensity of upwelling off southwest Africa. In: Summerhayes, C.P., Prell, W.L., Emeis, K.C. (Eds.), Upwelling Systems: Evolution Since the Early MioceneGeological Society Special Publication 63, 463–497. Heinio, P., Davies, R.J., 2009. Trails of depressions and sediment waves along submarine channels on the continental margin of Espirito Santo Basin, Brazil. Geological Society of America Bulletin 121, 698–711. Izumi, N., 2004. The formation of submarine gullies by turbidity currents. Journal of Geophysical Research 109, C03048. http://dx.doi.org/10.1029/2003JC001898. Jobe, Z.R., Lowe, D.R., Uchytil, S.J., 2011. Two fundamentally different types of submarine canyons along the continental margin of Equatorial Guinea. Marine and Petroleum Geology 28, 843–860. Kidd, R.B., Lucchi, R.G., Gee, M., Woodside, J.M., 1998. Sedimentary processes in the Stromboli Canyon and Marsili Basin, SE Tyrrhenian Sea: results from sidescan sonar surveys. Geo-Marine Letters 18, 146–154. Kostic, S., 2011. Modeling of submarine cyclic steps: controls on their formation, migration, and architecture. Geosphere 7, 294–304. Lass, H.U., Schmidt, M., Mohrholz, V., Nausch, G., 2000. Hydrographic and current measurements in the area of the Angola–Benguela Front. Journal of Physical Oceanography 30, 2589–2609. Micallef, A., Mountjoy, J.J., 2011. A topographic signature of a hydrodynamic origin for submarine gullies. Geology 39, 115–118. Migeon, S., Savoye, B., Zanella, E., Mulder, T., Faugères, J.-C., Weber, O., 2001. Detailed seismic-reflection and sedimentary study of turbidite sediment waves on the Var sedimentary Ridge (SE France): significance for sediment transport and deposition and for the mechanisms of sediment-wave construction. Marine and Petroleum Geology 18, 179–208. Moroshkin, K.V., Bunov, V.A., Bulatov, R.P., 1970. Water circulation in the eastern South Atlantic Ocean. Oceanology 10, 27–34. Normark, W.R., Hess, G.R., Stow, D.A.V., Bowen, A.J., 1980. Sediment waves on the Monterey Fan levee: a preliminary physical interpretation. Marine Geololgy 37, 1–18. Orange, D.L., Anderson, R.S., Breen, N.A., 1994. Regular canyon spacing in the submarine environment: the link between hydrology and geomorphology. GSA Today 4, 35–39. Pickering, K.T., Hiscott, R.N., Kenyon, N.H., Ricci, Lucchi F., Smith, R.D.A. (Eds.), 1995. Atlas of Deepwater Environments: Architectural Style in Turbidite Systems. Chapman and Hall, London.

97

Pilcher, R., Argent, J., 2007. Mega-pockmarks and linear pockmark trains on the West African continental margin. Marine Geology 244, 15–32. Piper, D.J.W., Normark, W.R., 1983. Turbidite depositional patterns and flow characteristics, Navy submarine Fan, California borderland. Sedimentology 30, 681–694. Pratson, L.F., Coakley, B.J., 1996. A model for the headward erosion of submarine canyons induced by downslope-eroding sediment flows. Geological Society of America Bulletin 108, 225–234. Pratson, L.F., Haxby, W.F., 1996. What is the slope of the U.S. continental slope? Geology 24, 3–6. Pratson, L.F., et al., 2007. Seascape evolution on clastic continental shelves and slopes. In: Nittrouer, C.A., Austin, J.A., Field, M.E., Kravitz, J.H., Syvitski, J.P.M., Wiberg, P.L. (Eds.), Continental Margin Sedimentation: From Sediment Transport to Sequence Stratigraphy. IAS Special Publication, 37. Blackwell, pp. 339–380. Rasmussen, E.S., 1994. The relationship between submarine canyon fill and sealevel change; an example from middle Miocene offshore Gabon, West Africa. Sedimentary Geology 90, 61–75. Rasmussen, E.S., 1996. Structural evolution and sequence formation offshore South Gabon during the Tertiary. Tectonophysics 266, 509–523. Rasmussen, E.S., 1997. Depositional evolution and sequence stratigraphy of the shelf and slope area off South Gabon, West Africa. Journal of Sedimentary Research 67, 715–724. Séranne, M., Nzé Abeigne, C.-R., 1999. Oligocene to Holocene sediment-drifts and bottom currents on the slope of continental margin (West Africa)—consequences for sedimentation and SE Atlantic upwelling. Sedimentary Geology 128, 179–199. Séranne, M., Seguret, M., Fauchier, M., 1992. Seismic superunits and post-rift evolution of the continental passive margin of southern Gabon. Bulletin de la Société Géologique de France 163, 135–146. Spinelli, G.A., Field, M.E., 2001. Evolution of continental slope gullies on the Northern California margin. Journal of Sedimentary Research 71, 237–245. Stow, D.A.V., Pudsey, C.J., Howe, J.A., Faugères, J.-C., Viana, A.R. (Eds.), 2002. Deep-Water Contourite Systems: Modern Drifts and Ancient Series, Seismic and Sedimentary CharacteristicsGeological Society of London's Memoirs 22. Stramma, L., England, M., 1999. On the water masses and mean circulation of the South Atlantic Ocean. Journal of Geophysical Research (Oceans) 104, 20863–20883. Stramma, L., Schott, F., 1999. The mean flow of the tropical Atlantic Ocean. Deep-Sea Research II 46, 279–303. Straub, K.M., Mohrig, D., 2009. Constructional canyons built by sheet-like turbidity currents: observations from offshore Brunei Darussalam. Journal of Sedimentary Research 79, 24–39. Twichell, D.C., Roberts, D.G., 1982. Morphology, distribution and development of submarine canyons on the United States Atlantic continental slope between Hudson and Baltimore Canyons. Geology 10, 408–412. Weimer, P., Slatt, R.M., 2007. Petroleum geology of deepwater settings. AAPG Studies in Geology, n.57. AAPG, Tulsa. Wonham, J.P., Jayr, S., Mougamba, R., Chuilon, P., 2000. 3D Sedimentary evolution of a canyon fill (Lower Miocene-age) from the Mandrove Formation, offshore Gabon. Marine and Petroleum Geology 17, 175–197. Wynn, R.B., Stow, D.A.V., 2002. Classification and characterisation of deep-water sediment waves. Marine Geology 192, 7–22. Wynn, R.B., Weaver, P.P.E., Ercilla, G., Stow, D.A.V., Masson, D.G., 2000a. Sedimentary processes in the Selvage sediment-wave field, NE Atlantic: new insights into the formation of sediment waves by turbidity currents. Sedimentology 47, 1181–1197. Wynn, R.B., Masson, D.G., Stow, D.A.V., Weaver, P.P.E., 2000b. Turbidity current sediment waves on the submarine slopes of the western Canary Islands. Marine Geology 163, 185–198.