The influence of bend amplitude and planform morphology on flow and sedimentation in submarine channels

Marine and Petroleum Geology 27 (2010) 1431e1447

Contents lists available at ScienceDirect

Marine and Petroleum Geology journal homepage: www.elsevier.com/locate/marpetgeo

The influence of bend amplitude and planform morphology on flow and sedimentation in submarine channels Kathryn J. Amos a, *, Jeff Peakall a, P. William Bradbury b, Mat Roberts a, Gareth Keevil a, Sanjeev Gupta c a

Earth and Biosphere Institute, School of Earth and Environment, University of Leeds, Leeds, West Yorkshire LS2 9JT, UK Rock Deformation Research Ltd., School of Earth and Environment, University of Leeds, Leeds, West Yorkshire LS2 9JT, UK c Department of Earth Science and Engineering, South Kensington Campus, Imperial College London, SW7 2AZ, UK b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 30 September 2009 Received in revised form 16 April 2010 Accepted 8 May 2010 Available online 15 May 2010

We present a series of experiments that investigate the morphology of sediment deposits within sinuous submarine channels of different sinuosity (S ¼ 1.14e1.94) and planform (symmetric and asymmetric bends), generated by bedload-dominated turbidity current flows. Flows were generated by releasing dense saline gravity currents over a mobile sediment bed through pre-formed sinuous channels. Flows had a basal-outwards helicity and produced a characteristic bed morphology with point bars downstream of the bend apex at the inside of bends and scour at the outside of bends. An increasing loss of fluid through overspill with increasing channel sinuosity results in a decreasing magnitude of crossstream velocity downstream, a decreasing amount of erosion and deposition, and decreasing transverse slopes of in-channel deposits. Basal fluid from within the channel is transported over the outer-levee at bends, implying that proximal outer-bend levee deposits will have similar sediment composition to that within the channel. More deposition of coarse material might be expected on levees and in overbank regions close to higher amplitude bends. No simple relationship was observed between superelevation and sinuosity, probably due to changes in the relative influences of downstream velocity and bend curvature on centrifugal force and inertial run-up. In the channel with the tightest initial bend curvature, overspill fluid from Bend 1 re-entered the channel at Bend 2, dominating flow characteristics and disrupting the basal-outwards helicity observed in the other channels. Higher sinuosity channels and those with shallow regional and levee slopes are thus more likely to have a higher proportion of anomalous flow and sedimentation patterns due to the influence of overspill fluid re-entry into the channel. The results of this investigation are combined with published observations to enable the synthesis of a new model for sedimentation in sinuous submarine channels. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Submarine channel Gravity current Turbidity current Point-bar Sinuosity Planform Flow dynamics Lateral accretion package

1. Introduction Turbidity current flow through submarine channels is a primary mode of clastic sediment transport to the deep ocean. Submarine channel systems form major geomorphological features on continental slopes and on the ocean floor, and their deposits can form important hydrocarbon reservoirs (Mayall et al., 2006; Wynn et al., 2007). Understanding the sedimentary processes occurring within submarine channels will provide information about the controls on submarine channel formation and evolution, and the pathway of terrestrial sediments to the ocean. This will enable better

* Corresponding author. Present address: Centre for Tectonics, Resources and Exploration (TRaX), Australian School of Petroleum, The University of Adelaide, Adelaide, SA 5005, Australia. Tel.: þ61 8 83034309; fax: þ61 8 83034345. E-mail address: [email protected] (K.J. Amos). 0264-8172/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpetgeo.2010.05.004

interpretation of ancient submarine channel deposits and the development of more accurate reservoir models. However, very few observations of turbidity currents from modern channels exist (Prior et al., 1987; Normark, 1989; Khripounoff et al., 2003; Paull et al., 2003; Xu et al., 2004; Vangriesheim et al., 2009). Even though geophysical data are an important source of information, the resolution of both surface-derived multibeam bathymetric data from modern systems and seismic data from ancient channels are typically too low to enable study of in-channel deposit features. Outcrop studies provide such resolution but rarely provide certainty of the location of observed features in relation to the channel planform; thus detailed observations of submarine channel deposits are relatively rare (Wynn et al., 2007). Due to the limited availability of direct observations and spatially detailed bathymetric and outcrop data, numerical and physical models of submarine channels are used to provide greater insight into sedimentary processes (e.g. Corney et al., 2006, 2008; Keevil et al.,

1432

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

2006, 2007; Imran et al., 2007; Peakall et al., 2007a; Islam and Imran, 2008; Islam et al., 2008; Kane et al., 2008; Straub et al., 2008; Kane et al., 2010). However, none of these studies has specifically addressed the influence of sinuosity on flow and depositional patterns, and consequently present understanding is over-reliant on a disparate group of single-sinuosity experiments, that collectively cover a range of experimental conditions. Here we address this limitation through a systematic analysis of the influence of sinuosity on flow processes and sedimentation.

2. Background Wynn et al. (2007) and Kolla et al. (2007) review published descriptions of submarine channel architectures and processes, and both conclude that there are significant differences between fluvial and deep-water sinuous channels, in terms of flow dynamics, internal architectures, and modes of evolution. Observations from the rock record, seismic studies and bathymetry data have described three deposit features within sinuous submarine channels; i) inner-bend ‘point-bar’ deposits, ii) outer-bend deposits that mantle the outer bank, called outer-bank bars, and iii) outer-bend deposits that form discrete pod-like elements, termed nested mounds. Inner-bend deposits (lateral accretion packages) have been inferred from low-angle reflectors dipping towards the channel in ancient deposits (e.g. Cook et al., 1994; Elliott, 2000; Haughton, 2000; Mayall and Stewart, 2000; Kolla et al., 2001; Abreu et al., 2003; Kneller, 2003) and have also been interpreted from seismic cross-sections of modern channels (e.g. Hesse and Rakofsky, 1992; Antobreh and Krastel, 2006) and bathymetric planform data (e.g. Klaucke and Hesse, 1996; Schwenk et al., 2003). Outer-bank bars have been recognized in seismic from the latter stages of submarine channel infill and are thought to be coarsegrained deposits related to rapid suspension fallout (Nakajima et al., 2009). Lastly, nested mounds are also coarse-grained deposits associated with the outer bank of channel bends (Phillips, 1987; Timbrell, 1993; Clark and Pickering, 1996), however, these are smaller features and it has been proposed that they might be related to levee collapse (Peakall et al., 2000a,, 2007a). Recent physical experiments and numerical simulations have successfully modelled density current flow, however, these studies present contradictory flow (e.g. Kassem and Imran, 2004; Corney et al., 2006; Imran et al., 2007; Keevil et al., 2007) and deposit (e.g. Peakall et al., 2007a, b; Straub et al., 2008; Kane et al., 2008) observations. Some experiments have shown flow in sinuous submarine channels to exhibit a helical flow circulation with an opposite direction of rotation to subaerial flows (Corney et al., 2006, 2008; Keevil et al., 2006, 2007; Peakall et al., 2007a, b) whilst others have shown river-like secondary circulation (e.g. Imran et al., 2007; Islam and Imran, 2008; Islam et al., 2008). Analytical modelling has shown that a primary control on the direction of helicity is the downstream velocity profile and the height of the velocity maximum above the bed (Corney et al., 2006,, 2008) although other factors such as cross-sectional geometry greatly influence flow processes (Islam et al., 2008; Straub et al., 2008). In density currents, where the velocity maximum is close to the bed (e.g. Tesaker, 1969; Stacey and Bowen, 1988; Garcia and Parker, 1993; Buckee et al., 2001; Felix, 2002, 2004; Corney et al., 2006) basal flow will likely be directed towards the outer bank, with an upper return flow towards the inner bank; the opposite to that observed in rivers, which have a downstream velocity maximum near the water surface. In cases, where the velocity maximum is higher in the flow, then flows may typically show a more river-like helical behaviour. For the remainder of this paper, the sense of helicity of flow circulation will be described as basal-inward and basal-

outward, to avoid use of terms such as ‘normal’ or ‘reversed’ circulation. Peakall et al. (2007a), Straub et al. (2008) and Kane et al. (2008) present the first experimental studies that successfully reproduce the constructional intra-channel architecture of sinuous submarine channels in small-scale physical experiments. These studies focus on significantly different phases of turbidity current sedimentation; Peakall et al. (2007a) describe deposition from bedloaddominated flows, whereas Straub et al. (2008) examine suspension-dominated flows, and Kane et al. (2008) describe flows that exhibit components of both suspension and traction. Peakall et al. (2007a) ran solute-driven density currents through pre-formed leveed sinuous channels containing mobile sediment beds, in which sediment was entrained, transported and deposited. These flows formed in-channel deposits with areas of inner-bend accumulation and outer-bend erosion that developed with time, similar to the ‘point-bar and pool’ bed morphology of sinuous rivers. As in rivers, these results demonstrate the dominance of the longitudinal flow over the cross-stream flow. However, differences were observed in the position of the inner-bend sediment accumulations between these experiments and in rivers, with the submarine ‘point-bars’ forming slightly downstream of the bend apex (Peakall et al., 2007a), as has been described from some natural submarine channels (Schwenk et al., 2003; Abreu et al., 2003). The experimental flows of Straub et al. (2008) modified a preexisting channel by spatially varying patterns of sedimentation. These flows deposited greater thicknesses of sediment towards the outer banks of bends, and exhibited enhanced sedimentation in the channel bottom than on the levees, successive flows reducing channel relief and sinuosity. Straub et al. (2008) concluded that in these flows, sedimentation rates were greatest where near-bed suspended sediment concentrations were greatest, thicker outerbank deposits reflecting the influence of a run-up associated with the momentum of the current as it flows around a bend. These experimental outer-bend deposits may be analogous to outer-bank bars and nested mound deposits described from natural systems (Timbrell, 1993; Nakajima et al., 2009). Kane et al. (2008) ran flows through curved channels of different depths, and found that with increasing confinement, more sediment was transported downstream within the channel as less sediment was lost overbank, and in-channel deposit thicknesses were greatest along the inner-bend (when flow height/levee height ¼ 1.7e2.5). In their less-confined flows (flow height/levee height ¼ 5e10), more sediment was deposited overbank and in-channel deposit thicknesses were greatest along the outer-bend. These results appear to conflict with those of Straub et al. (2008) in which flow height/levee height was 1.1e1.2, and flows deposited preferentially around the outside bend. Flow velocities in these two sets of experiments were similar, but density excess, grain size and channel cross-section were different. One key difference between the experiments however is that tractional reworking occurred in the more confined experiments of Kane et al. (2008) [see Figs. 7.10 and 7.15 of Kane (2007)], whilst the deposits described by Straub et al. (2008) contain no internal laminae, and so no evidence for traction. Consequently, the more channelised flows in the Kane et al. (2008) experiments exhibit inner-bend accumulations, and only larger, less constrained flows exhibit pure suspension deposition. In this respect both sets of experiments show outer-bend accumulations associated with suspension-dominated deposition. Interestingly, Straub et al. (2008) focus on describing the deposits at the second bend apex, whereas Kane et al. (2008) ran a one-bend experiment. The cumulative deposit thickness data presented by Straub et al. (2008; their Fig. 6C) show that at the first bend of their channel, sediment was preferentially deposited along the inside bend. It is possible that some tractional reworking occurred at the proximal, higher

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

energy bend-1 location in the Straub et al. (2008) experiments accounting for this difference. The alternative argument that the first bend is unrepresentative is dismissed here given that the Kane et al. (2008) experiments fit with i) other experimental work that examined subsidiary bends (Peakall et al., 2007a, b), ii) the influence of traction on inner-bend point-bar deposits (Dykstra and Kneller, 2009; Pyles et al., 2010), and, iii) the formation of outerbank bars by late-stage flows inferred to be strongly suspension dominated (Nakajima et al., 2009). In summary, experimental and theoretical studies indicate that the deposits of turbidity currents in sinuous submarine channels will vary significantly based on whether in-channel deposition is dominantly from bedload or suspension transport. This will vary between channels, as well as spatially (i.e. downstream) and temporally (i.e. vertically through a channel-fill succession). 3. Aims In this paper, we focus on extending the understanding of bedload-dominated turbidity current flows through sinuous submarine channels, and present results from an investigation of flow dynamics and deposit characteristics within channels of varying planform morphology. In nature, the planform morphology of channels varies; channels are often highly sinuous, with asymmetric-bend planforms. Lateral migration, cut-off loops and scroll bars are commonly observed in modern deep-water sinuous channels (e.g. Wynn et al., 2007). Submarine channel sinuosities of approximately 1 to greater than 3 have been reported (e.g. see Clark et al., 1992; Kolla et al., 2007; Wynn et al., 2007). In the literature, sinuous channels have been defined as having a sinuosity of greater than 1.2 (Wynn et al., 2007) and 1.15 (Clark et al., 1992; Clark and Pickering, 1996). Clark et al. (1992) showed that for a given submarine channel system, channel sinuosity increased as axial (valley) gradient decreased until sinuosity reached a maximum, and then decreased with further decreases in gradient, analogous to fluvial sinuous channels (Schumm and Khan, 1972; Schumm et al., 1972). As well as the position of the submarine channel point bars being different to those observed in rivers (Peakall et al., 2007a), the overall bed morphology and deposit characteristics are also different. Submarine point-bars are thus distinctive in terms of their morphology, position and deposit character. This study aims to investigate the influence of bend amplitude and planform morphology on the fluid dynamics and deposit topography of gravity current flows in submarine channels. Using the same experimental set-up and conditions of Peakall et al. (2007a), we present results from experiments investigating flow dynamics and deposits in channels of differing bend amplitude, and within a channel with an asymmetric-bend planform. 4. Methodology Three sets of experiments were run using different channel models. Each examined cross-stream flow fields and deposit topography following three separate flows of approximately 60 s over the same bed. Experiments were conducted within a square flume tank (1.8 m by 1.8 m by 1.7 m depth) with a 2 m long input channel centered on one face, set to a slope of 3 (Fig. 1A). Model channels (made from glass-reinforced plastic) were placed within the flume on a raised platform (0.4 m above the base of the tank). Saline fluid was mixed in a separate tank and pumped into the flume tank through an expansion pipe and into a straight channel (0.16 m wide and 1 m long) that was aligned flush with the input channel. Inserts within the straight section of channel achieved a smooth and gradual transition from the rectangular cross-section

1433

of the input channel to the curved cross-section of the sinuous channel. Saline fluid was pumped into the flume tank at a constant discharge of 1.25 l s-1 (0.05 l s-1) controlled by electromagnetic flow meters. Gravity current reflections were minimised by the incorporation of a sump area beneath the platform, and by balancing input rates with pumped output. In the first series of experiments, two channels were used which had simple planform geometries. These geometries were based on the UK Flood Channel Facility Fluvial Series B experiments (e.g. Sellin et al., 1993; Willetts and Rameshwaran, 1996), enabling comparison with fluvial channel characteristics, including those of compound channels. The geometry of these channels consists of a series of bends of constant radius, separated by straight sections; each channel consisting of two complete wavelengths (three bends). Channels have a constant ratio of channel axis length to straight section length, enabling sinuosity to be varied by only altering the bend radius. This planform geometry is applicable to a wide range of channel types, both in subaerial and submarine environments. In these experiments, the arc radius varied between channels, producing channel sinuosities (S) of 1.14 and 1.78 (Fig. 1B, C). These channels have the same geometry as that used in the experiments of Keevil et al. (2006) and Peakall et al. (2007a), which had a sinuosity of 1.36 (Fig. 1D), and thus our results will be directly comparable. The results from the experiments of Peakall et al. (2007a) are included here, in order that our new data can be directly compared. The second series of experiments used a third channel model with the same cross-section geometry, but an asymmetric-bend planform, based on that of a natural channel seen in seismic data from offshore Gabon, west Africa (Peakall et al., 2005). This channel model planform was an approximation based on that observed in seismic data, rather than attempting to take a full three-dimensional pick of the basal channel surface. This channel model includes an asymmetric compound bend, and has a sinuosity of 1.93 (Fig. 1E). The channel cross-section is constant along the channel length of all the models, with a levee height of 75 mm and width of 150 mm between levee crests, and curved edges in order to minimize corner-induced circulation (Naot and Rodi, 1982; Nezu and Nakagawa, 1993; Fig. 1E). Dimensions of the experimental channel models are presented in Table 1. Before each set of experiments, the sinuous channel model and the input channel were filled to a depth of 25 mm with low-density sediment (PolyVÒ, 1.1e1.2 g cm-3) that is angular, polydisperse, and has a median grain size of 175 mm (Fig. 2). The PolyVÒ sediment was soaked overnight in water containing a drop of dispersant prior to being used in these experiments. A weir, equal in height to the sediment bed, was added at the downstream end of the channel to prevent sediment from escaping, in order that an initial bed of uniform thickness could be achieved. Initial bed topography and bed topography after each flow period was measured using an ultrasonic bed profiler (Best and Ashworth, 1994) at a grid spacing of 10 mm. Ultrasonic Doppler Velocity Profiling (UDVP) was used to measure cross-stream flow velocities at the apex of the central bend of each channel. For a general description of UDVP, see Best et al. (2001). Velocities were recorded at 128 positions across the channel from each of ten 4-MHz ultrasonic transducers mounted flush with the channel cross-section. Transducers were mounted at the inner bend in order to minimize disturbance to the flow field, and at heights of 6,17, 28, 39, 49, 59, 70, 80, 90 and 101 mm above the base of the channel. Velocities were measured from each probe in series; one complete measurement of cross-stream velocities from all ten transducers took 260 ms. Velocities were measured for 60 s, starting when the flow front reached the end of the channel model. The pumped input of dense fluid to the flume tank was stopped immediately after the UDVP velocity recording period ended.

1434

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

Fig. 1. A. Schematic plan view of the experimental set-up, illustrating the placement of channel models within the flume tank. The platform within the flume tank (and thus the valley slope of the channel models) had a 3 slope. B. Planform geometry of the S ¼ 1.14 channel, with a bend amplitude (a) of 40 . C. Planform geometry of the S ¼ 1.78 channel, a ¼ 80 . D. Planform geometry of the S ¼ 1.36 channel, a ¼ 60 . E. Planform geometry of the asymmetric channel, modelled on a natural channel planform; S ¼ 1.94. This is the same channel model used in the experiments of Keevil et al. (2007) and Peakall et al. (2007a,b). Bend wavelength (l) is the same for all channels, as indicated in (B). Dashed lines depict the location of velocity measurements recorded using the UDVP. F. Cross-sectional geometry of the channel models and the initial sediment fill placed within each channel. Adapted from Keevil et al. (2006).

As the UDVP could not be used to directly measure flow velocities at the base of the channel model, basal flow directions within the S ¼ 1.14 channel were visualised during a separate experiment. Saline flow was pumped into the channel model as described above, but in this instance there was no sediment bed, and short pieces (30 mm) of polyester thread were attached to the channel floor using adhesive tape. These threads were arranged in 16 rows between the apices of bends 1 and 3, with thread placed at four evenly spaced positions across the width of the channel floor in each row. The orientation of these threads was recorded using a digital video camera placed directly above the channel model, above a floating glass window placed on the free water surface to

prevent any distortion from surface waves. The channel was illuminated from both sides using 500 W lamps. Three still frames were captured from the digital video footage at 30, 40 and 50 s after the head of the current reached the end of the channel model, and the orientations of each thread in these still frames were digitised to produce a visualisation of basal channel flow direction during this flow period. These can be directly compared with the basal flow directions observed from the S ¼ 1.36 channel presented in Keevil et al. (2006). This experimental investigation utilises the generic modelling approach, which aims to reproduce similarity of processes and features of a natural system without being scaled using specific

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447 Table 1 Summary of experimental channel form.

Table 2 Summary of experimental parameters.

Channel Name

40Degree

60Degree

80Degree

Bend amplitude

40

60

80

Sinuosity Wavelength Total valley length Total channel axis length Straight segment length Ratio channel axis length: straight section length Axis-to-axis system width

Asymmetrical

Bend 1 ¼ c. 90 1.138 1.355 1.784 1.936 764 mm 764 mm 764 mm e 1529 mm 1529 mm 1529 mm 1529 mm 1740 mm 2072 mm 2728 mm 2953 mm 118 mm 141 mm 186 mm e 2.674 2.674 2.674 e 182 mm

1435

302 mm

477 mm

Initial density Input discharge a Densiometric pffiffiffiffiffiffiffiffi Froude number Fr ¼ U g1 h where g1 ¼ g(rs-ra)/ra Reynolds numbera Re ¼ rsUh/m Particle grain size (D50) Particle specific gravity a

2.5% 1.25  103 m3 s1 0.63 7831 175 mm 1.2

Based on experimental data presented in Keevil et al. (2006).

637 mm

prototype or generic data (Hooke, 1968; Chorley, 1967; Peakall et al., 1996, 2007b). All flows were sub-critical and turbulent, ensuring similarity of mixing characteristics with previous experimental studies of density currents (Table 2). 5. Results from Series 1 experiments: varying channel sinuosity 5.1. Flow dynamics Gravity current flow thickness was greater than the channellevee height, causing overspill of dense fluid along the entire length of the channel models. Overspill was most prominent along the outside edge of channel bends just downstream of bend apices. A greater volume of overspill was observed in the S ¼ 1.78 channel than the S ¼ 1.14 channel. Sixty-second means of the cross-stream flow velocities at bend apices are presented in Fig. 3, for the first, second and third minutes of cumulative experimental flow for the 40, 60 and 80 degree channels. Data from the S ¼ 1.36 channel were presented in Peakall et al. (2007a), plotted here using a different scale (normalised velocities) to enable direct comparison with data from the S ¼ 1.14 and S ¼ 1.78 channels. Due to accumulation of sediment close to the inside of the bend, flow velocities in the outer-bend region of the channel below the maximum thickness of the deposit could not be

Fig. 2. Particle grain size distribution of the Poly VÒ sediment used in these experiments.

recorded, because of the position of the probes at the inner-bend. Flows within these channels (all sinuosities) have a strong basal movement of cross-stream flow towards the outer bank, and a weaker movement of the fluid above moving towards the inner bank. In general, the majority of channelised flow was directed towards the outer bank. These figures depict a region of elevated cross-stream velocity over the outside levee, and an isovel gradient from the basal velocity core to the outside levee. The 0.7 normalised velocity contour (an arbitrary value) is depicted in Fig. 3 in order to better highlight the shape of the velocity core. The data presented depict no clear trends in the minimum and maximum cross-stream velocities or of the height of the basal velocity core with time (Fig. 4). Maximum velocities are generally higher in the S ¼ 1.14 channel than in the S ¼ 1.36 channel, and maximum velocities are substantially lower in the S ¼ 1.78 channel. Basal flow vectors visualised for the S ¼ 1.14 channel depict an overall movement of fluid towards the outside bank upstream of the bend apex, and a shift in direction towards the inside bank downstream of the bend apex (Fig. 5). These are similar to basal flow vectors recorded in the S ¼ 1.36 channel model presented by Keevil et al. (2006; reproduced in Fig. 5).

5.2. Deposits Deposits formed within the S ¼ 1.14 and S ¼ 1.78 channels resulting from cumulative flow periods of 1, 2 and 3 min over an initial bed 25 mm thick are presented in Fig. 6 alongside those presented by Peakall et al. (2007a) for the S ¼ 1.36 channel. Net change is presented in Fig. 7 for all three channels, depicting areas and amounts of erosion and deposition that occurred between flow periods. In the S ¼ 1.14 channel, the main zones of deposition occurred just downstream of each bend apex, along the inside of the bend (Figs. 6A and 7A). A substantial region of erosion occurred around the apex of Bend 1, spanning the width of the channel at the bend apex and continuing along the outside bank to the apex of Bend 2 (Fig. 7A). At Bend 2, the zone of erosion did not extend across the channel after the first minute of flow, although a small zone of erosion occurred along the outer bank just downstream from the apex of Bend 2. During the second and third minutes of flow, the zone of erosion at Bend 2 extended across the channel and expanded in size with time. The amount of erosion at Bend 1 was greatest during the second minute of flow, but the erosion at Bend 2 was greatest during the third minute of flow. A zone of erosion also occurred on the outer bank at the apex of Bend 3. With time, the topographical lows generally became lower, and the highs became higher (Fig. 6A). Overall, the greatest amount of erosion occurred at Bend 1, the smallest at Bend 3, and the proportion of the bed that is erosive increased with time (Fig. 7A). Similarly, the greatest amount of deposition occurred downstream of Bend 2. Net erosion and deposition are presented for the S ¼ 1.36 channel using the deposit topography data presented in Peakall

1436

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

Fig. 3. Cross-sectional velocities recorded at the apex of Bend 2. These plots present velocities averaged over 60 s for the periods 0e60 s, 60e120 s and 120e180 s of cumulative flow over the same deposit in each of the three symmetric-planform channel models. Velocities have been normalised by the maximum mean velocity to enable better comparison of cross-stream flow fields in these different channels. In these figures, the inside levee (I) is to the left and the outside levee (O) is to the right. Downstream flow direction is away from the reader. Negative normalised mean cross-stream velocities (U norm ) indicate flow towards the inside levee, and positive U norm indicate flow towards the outside levee. The locations of these velocity transects are depicted in Fig. 1. The velocities closest to the UDVP transducers mounted within the inside levee and those below the sediment bed could not be recorded. Initial sediment thickness was 25 mm.

et al. (2007a; Fig. 7B). These data depict a similar trend to that described in Peakall et al. (2007a) from deposit topography data (Fig. 6B), with highest rates of deposition occurring close to the bank, just downstream of the apices of bends 1 and 2. Erosion dominantly occurs close to the bend apices, towards the outside bank. As observed in the S ¼ 1.14 channel deposit, a substantial region of erosion occurs around the apex of Bend 1, and this zone of erosion migrates downstream around the bend with time. In the S ¼ 1.78 channel, a considerable amount of erosion occurred within the initial ¼ wavelength of the channel in the first minute of flow (Fig. 7C). Downstream of this, erosion occurred along the inside of the bend, although with a greatly reduced magnitude, and after the second minute of flow was detached from the bank by a region of deposition at the apex of the bend. A zone of sediment accumulation formed at each bend apex, and erosion occurred at the outside of the bends. The majority of erosion and deposition occurred around Bend 1, where a bank-attached accumulation formed at the apex and extended only a short distance around the bed. Downstream from Bend 1, patterns of erosion and deposition are patchy, and indicate that the dominant mode was by a train of low-amplitude bedforms (Fig. 7C). Within these deposition zones, the region of maximum deposition is bank-attached at the bend apices, and detaches from the bank downstream of the bend apex, tending towards the centre of the channel and reaching the opposite bank at the inflection point between bends. During the first and second minutes of flow, as well as the larger zones of erosion observed at each bend, small pockets of erosion occurred towards the centre of the channel at Bend 2, and a series of three parallel elongate regions of erosion had developed towards the outside of Bend 3. By the third minute of flow, the bed of the initial quarterwavelength of channel had been almost completely scoured to the non-erodible channel floor, and the zone of erosion along the inner bank upstream of Bend 1 had extended across the channel and

merged with the zone of erosion at the outer bank. At Bends 2 and 3, the pockets of erosion were larger, and bedform features were apparent from the patterns of erosion and deposition on the outside of the bends, detaching from the bank and crossing the channel just upstream of the bend apex. As observed in the S ¼ 1.14 channel, the proportion of the bed that is eroded increases with time. Comparison of deposits in these three channels of varying sinuosity shows that the magnitude of the bed topography (Fig. 6) and the amount of erosion and deposition (Fig. 7) increase with decreasing sinuosity. At Bend 2, the transverse slope between banks is greater in the lower sinuosity channels. The amount of scour in the proximal channel bend increases with increasing sinuosity, as does the height of the bar that forms at Bend 1. 6. Results from Series 2 experiments: asymmetric channel planform 6.1. Flow dynamics As in the Series 1 experiments, overspill occurred along the entire length of the channel model. The dynamics of flow within this channel were more complex than in the symmetric channels, as some of the overspill from the first corner of the bend was observed to re-enter the channel downstream within the meander loop. The cross-stream velocity data (Fig. 8A) show that much of the flow within the measurement window is directed towards the outside of the bend, with flow at the base of the channel moving from the outside towards the inside of the bend, opposite to that observed in the symmetric-bend channels. The maximum crossstream velocities are outward directed and occur towards the top of the inside levee, and decrease with distance from the inside bank. This flow pattern is the result of the observed overspill re-entry into the channel from over the inside levee crest.

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

1437

Fig. 4. A. Maximum and minimum mean cross-stream velocities, and B. Height of the 0.7 U norm contour (used to represent the basal velocity core), recorded in each of the symmetric channel models for the three flow periods (0e60, 60e120 and 120e180 s).

6.2. Deposits

Fig. 5. Basal flow directions as derived from visualisation of flow using cotton threads adhered to the channel floor. Line thickness’ have been increased in order to improve figure clarity. Relative positions and lengths of the lines remain proportionate. A. S ¼ 1.14 channel. These were derived by digitising three stills of video footage each taken 10 s apart, starting at the point the flow front reached the end of the channel. B. S ¼ 1.36 channel (from Keevil et al., 2006). Inset is an illustration of helical flow circulation at the bend apices. Flow direction is from left to right.

Deposits formed within the asymmetrical channel model resulting from cumulative flow periods of 1, 2 and 3 min over an initial bed 25 mm thick are presented in Fig. 8B. The topography of the initial sediment bed is also presented, as it was not completely flat and thus influences the interpretation of the subsequent bed topographies. The deposits within this channel are dominated by a substantial amount of erosion within the initial section of the channel upstream of Bend 1. This zone of erosion is concentrated at and just upstream from the apex of Bend 1, and extends across the channel from the inside to the outside of the bend. The lowest part of the bed occurs on the outer bank of Bend 1, just upstream of the bend apex. On the inside of this bend, a zone of deposition extends from just upstream of the bend apex, downstream to Bend 2a. The highest part of this depositional-zone extends from the inside bend at the apex of Bend 1, across the channel to join the opposite bank at the inflection point between bends. Two zones of erosion occur at Bend 2a, one attached to the inside bend, the other against the outside bend, just downstream of the bend apex. The zone of erosion on the inside of Bend 2a is at the same location as a topographic low in the initial bed. Downstream of Bend 2a, the bed is not much altered; a train of low-amplitude (c. 1e3 mm) bedforms were observed in all deposits, and a region of deposition occurred on the inside bend just downstream of Bend 3.

1438

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

Fig. 6. Contour plots of deposit topographies after cumulative time periods of 1, 2 and 3 min of flow. A) S ¼ 1.14 channel; B) S ¼ 1.36 channel; C) S ¼ 1.78 channel. The topography is presented as distance to the top of the deposit from an arbitrary datum in tenths of a millimetre; blue represents the lowest areas of deposit and red the highest areas. Initial bed thickness was 25 mm. Flow direction is towards the top of the page.

7. Discussion 7.1. Flow dynamics Velocity data recorded in the S ¼ 1.14 and S ¼ 1.78 symmetrical experimental channels indicate that these channelised gravity

flows had basal-outward helical circulation, as has been previously observed in the S ¼ 1.36 experimental channel (Fig. 3; Peakall et al., 2007a). This basal-outward circulation has been shown to occur adjacent to the channel bed through visualisation of basal flow in the S ¼ 1.14 and S ¼ 1.36 channels (Fig. 5 and Keevil et al., 2006). The highest cross-stream velocities were recorded at measuring points

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

1439

Fig. 7. Net change between deposits after flow periods of 1 min, measured in tenths of a millimetre. A) S ¼ 1.14 channel; B) S ¼ 1.36 channel; C) S ¼ 1.78 channel. Erosion is depicted in greenseblues, deposition in yellows-reds. Some irregularities in data occur along the channel edges due to the steepness of the levee slopes. Flow direction is to top of page.

closest to the bed of the channel, showing movement of fluid from the inside to the outside. A region of elevated cross-stream velocities close to the crest of the outer-bend levee is consistent with movement of this basal fluid up the outside bank and over the outside levee. The inclined isovels depict the superelevation of flow at the bend apex, and the enhanced overspill of fluid at the outer bank relative to the inside bank. The shape of the velocity core (the

region of highest velocities) is linked to the movement of flow over the outside-bend levee as overspill (Fig. 3). Velocity data in these experiments clearly show that discharge loss at bends is related to movement of basal fluid out of the channel along the outside bend (Fig. 3). This indicates that the resultant overbank flow along the outside of bends has a similar composition to that of the entire channelised current, rather than just the upper parts, with

1440

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

important implications for estimating the grain size of sediment transported overbank. This has also been observed for channelised experimental flows around bends by Corney et al. (2006), Keevil et al. (2006, 2007), Peakall et al. (2007a) and Straub et al. (2008). In the Series 2 experiments with the asymmetric channel, the dynamics of flow recorded at Bend 2 were dominated by re-entry into the channel of overspill fluid from the upstream bend (Fig. 8A). The cross-stream velocities of overspill fluid entering the channel decrease with distance from the inside bank, probably due to overall deceleration of the fluid and/or a change in direction of the re-entered fluid as it mixes with the channelised flow. The crossstream velocity data (Fig. 8A) showed an opposite trend to that observed in the symmetric-bend channels, with flow at the base directed towards the inside bend. This movement of basal fluid towards the inside bank is probably the result of a circulation induced by the deflection of fluid re-entering the channel, against the outside bank. The extent to which overspilling fluid interacts with in-channel flow at downstream locations will be influenced by the magnitude of the overbank slope away from the channel, and the topographic expression of the channel boundary relative to the regional slope (Keevil et al., 2007). The velocity results presented here for the Series 1 channels (Fig. 3) are comparable to those from experimental sediment-free

gravity flows in the S ¼ 1.36 channel, in which the three-dimensional flow dynamics around a channel bend were investigated in detail (Keevil et al., 2006) and it was demonstrated that that the direction of helical flow rotation was controlled by the downstream velocity profile (Corney et al., 2006; 2008). Detailed three-dimensional velocity fields presented by Keevil et al. (2006; 2007) are most likely driven by a combination of centrifugal force and inertial run-up (see Straub et al., 2008). This run-up of basal flow in these cases enhances the secondary circulation-induced movement of basal flow towards the outer bank in the upstream part of channel bends (Wynn et al., 2007). The results presented here confirm that the basal-outwards helicity observed in sediment-free experiments is applicable to turbidity currents that are reworking their bed. However, as discussed by Peakall et al. (2007a), it is likely that the accretion of inner-bend deposits over time will add a spatial (convective) acceleration, which will increase the strength of the outwarddirected flow at the upstream end of the deposit and inwarddirected flow at the downstream end of the deposit relative to that observed in the sediment-free experiments (Fig. 5 and Keevil et al., 2006). Different bed topographies may result if substantial volumes of overspill fluid re-enter the channel downstream, capable of suppressing the amount of sediment accretion adjacent to the

Fig. 8. Results from the asymmetric channel model. A) Cross-sectional velocities recorded at the apex of Bend 2. These plots present velocities averaged over 60 s for the periods 0e60 s, 60e120 s and 120e180 s of cumulative flow over the same deposit. Velocities have been normalised by the maximum mean velocity. The inside levee (I) is to the left and the outside levee (O) is to the right. Downstream flow direction is away from the reader. Negative normalised mean cross-stream velocities (U norm ) indicate flow towards the inside levee, and positive U norm indicate flow towards the outside levee. The locations of this velocity transect is depicted in Fig. 1E. The velocities closest to the UDVP transducers mounted within the inside levee and those below the sediment bed could not be recorded. Initial sediment thickness was 25 mm. B) Contour plots of deposit topographies after cumulative time periods of 1, 2 and 3 min of flow. The topography is presented as distance to the top of the deposit from an arbitrary datum in tenths of a millimetre; blue represents the lowest areas of deposit and red the highest areas. Initial bed thickness was 25 mm. Flow direction is towards the top of the page.

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

inside bend and outer-bend scour, and potentially resulting in cessation of deposition or erosion of the bed. 7.2. Influence of sinuosity on flow dynamics In the Series 1 experiments, the magnitude of cross-stream velocity decreases with increasing sinuosity, due to increasing loss of fluid through overspill with increasing sinuosity. The initial bend of the Series 2 asymmetric channel was tighter than that of the Series 1 channels (S ¼ 1.94; Fig. 1E, Table 1) and overspill was even greater in flows through this channel model, resulting in the re-entry of overspill into Bend 2 from the upstream bend. Comparison of basal flow directions visualised for the S ¼ 1.14 channel and S ¼ 1.36 channel (Fig. 5) show a similar pattern, with basal flow directed towards the outside bank around the bend, with the inflection point upstream of the mid-point between bends. The only notable difference is that basal flow directions in the S ¼ 1.14 channel are directed more steeply towards the outer bank upstream of the bend apices (Bends 2 and 3) than in the S ¼ 1.36 channel. Keevil et al. (2006) observed that in their experiments, in which cross-stream, downstream and vertical velocities were recorded, the downstream velocity core was offset towards the inside of each bend, and adjacent to the channel wall at the inflections upstream of the bend apexes. This is thus likely to be the case in the experiments presented here. This flow dynamic is contrary to the suspension-fallout-dominated turbidity currents of Straub et al. (2008) in which the path of the velocity core was deflected towards the outside of each bend, just downstream of the bend apex. These differences in flow dynamics are probably due to the relative influences of flow helicity and run-up, resulting from differences in flow energy, density and bend geometry. A cross-section of downstream velocities would be needed in these experiments to determine the transverse location of the basal velocity core, and hence any contribution of run-up to the cross-stream velocity profile. However, if the angle of the 0.7 U norm velocity contour (Fig. 3) is used as a proxy for superelevation, along with magnitudes of U norm over the outside levee (Fig. 3), it appears that superelevation is greatest in the channel with intermediate sinuosity (S ¼ 1.36). Centrifugal force and run-up in sinuous channels will increase with increasing downstream velocity and with increasing bend curvature. The interaction of these forces in the fixedchannel models used here provides no simple relationship with sinuosity, as might be expected in nature. As sinuosity increases, the amount of overspill increases, decreasing the cross-stream velocity at Bend 2, and most likely the magnitude of the downstream velocity also. A high superelevation in the intermediate sinuosity channel could be explained by the low sinuosity dominating over the higher velocities producing lower superelevation in the low-sinuosity channel, and the decreased velocities due to overspill dominating in the high-sinuosity channel. 7.3. Deposits The experimental submarine channel deposits presented here comprise two primary zones with corresponding morphological features: 1) a zone of sediment accretion related to ‘point-bar’ deposits that form just downstream of bend apices adjacent to the inside bank, and 2) regions of scour that correspond to topographical lows or ‘pools’ along the outside of each bend (Figs. 6 and 7, and at Bend 1 in Fig. 8b). The Series 1 experimental results depict a decreasing amount of deposition and erosion with time, indicating that these deposits are moving towards a steady-state deposit that is in equilibrium with flow conditions (Fig. 7). The

1441

deposits of the asymmetric channel planform are dominated by the tightness of the initial bend rather than channel planform asymmetry, which will be discussed further below. At Bend 2a, the bed topography may also be influenced by the unevenness of the initial pre-flow bed (Fig. 8b). As described by Peakall et al. (2007a), the ‘point-bar’ and ‘pool’ units observed in these experimental deposits broadly correspond to their fluvial counterparts (e.g. Dietrich, 1987), with key differences being that the inner-bend sediment accumulations form further downstream in these experimental submarine channel deposits than in fluvial channels (e.g. Hooke, 1975; Dietrich, 1987; Wormleaton et al., 2004, 2005), with partially related changes in the widths and transverse slopes of the inner bank accumulation as a function of bend position (e.g. Leopold et al., 1964, p. 317). As discussed by Peakall et al. (2007a), the variation in point-bar position relative to the bend apex between these submarine channel experiments and fluvial systems may be related to two mechanisms, 1) a variation in the location of flux convergence caused by the lowermost part of the flow cell (Nelson and Smith, 1989), and 2) an area of slow downstream velocity close to the inner bank, related to the curvature of the inner bend at the apex (Keevil et al., 2006; Islam et al., 2008). The results presented for these experimental channels indicate that sediment flux convergence occurs further downstream in these submarine channel experiments relative to river channels, as a result of basal-outwards helical motion up to and just beyond the bend apex, followed by flow convergence towards the inner bank at a point beyond the bend apex (Fig. 5). Visualisation of the sediment particles eroded into these experimental flows demonstrated no permanent zone of recirculation as is observed in rivers with sharp bends (e.g. Bagnold, 1960; Leopold et al., 1960; Ferguson et al., 2003), although occasional upstream movement of particles at the inner bend was observed. The presence of a slow velocity zone close to the inner bank downstream of the bend apex will enhance the effects of flux convergence in this area and encourage development of a point bar. Comparison with compound channel experiments (channels with a narrow floodplain that becomes submerged during flood events; e.g. Wormleaton et al., 2004; 2005) demonstrates that basal-outward flow helicity (in both submarine channel and fluvial channel models with confined overbank flow) results in the position of the point-bar having its upstream end just beyond the apex (i.e. further downstream than occurs under ‘normal’ fluvial basalinwards helicity). During overbank flow in compound channels, shear between the predominantly downstream flowing overbank flow and the sinuous within-channel flow results in a reversal of helicity in parts of the channel bends (Knight and Shiono, 1990; Sellin et al., 1993; Wormleaton et al., 2004, 2005). At bank-full and lower discharges, compound channels exhibit point-bar deposits that begin upstream of the bend apex, as is typical of meandering rivers (e.g. Wormleaton et al., 2004, 2005). However, during overbank flow, point-bar deposits have been observed to form downstream of the bend (Wormleaton et al., 2005), in a similar position between the bend apex and inflection that was observed in the submarine channel experiments presented here. That this same pattern has been observed in compound channel experiments and all the experimental submarine channel deposits presented here confirms that this is a true product of flow dynamics and flowesediment interaction. Strong similarities between these submarine channel experiments and the larger experimental compound channel experiments of Wormleaton et al. (2005) and some (limited) evidence for asymmetric ‘point-bars’ with respect to bend apices in natural channels (Schwenk et al., 2003; Abreu et al., 2003) also demonstrates that these small-scale physical models of submarine channels are likely to replicate the gross dynamics of natural channels.

1442

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

7.4. Influence of channel sinuosity on deposits

7.6. Placing results in the context of natural flows

The amount of erosion and deposition recorded in these experimental deposits decreases with increasing channel sinuosity (Fig. 7, and comparison between Figs. 6 and 8). This is reflected in transverse slopes which increase with decreasing sinuosity (Figs. 6 and 8). In the Series 1 symmetric-planform channels, this is the result of decreased flow velocities being recorded in higher sinuosity channels due to increasing overspill with bend amplitude. As overspill increases, flow will decelerate more rapidly with distance downstream due to loss of fluid and increased mixing, reducing the ability of the flow to transport grains. These trends in sediment deposition will also be influenced by the increased transport of eroded proximal material overbank with increased fluid overspill. This is implied for the Series 2 asymmetric-planform experiments in which re-entry of overspilled fluid into the channel dominated flow characteristics at Bend 2, where substantial overspill occurred at Bend 1 and very little sediment reworking occurred downstream (Fig. 8b). It is also possible that the influence of re-entry of overspilled fluid on flow dynamics also acted to suppress sediment reworking downstream. The proximal scour in these experimental channel beds increased with increasing sinuosity for the Series 1 experiments (Figs. 6 and 7). This is probably related to the increase in centrifugal force with bend amplitude, increasing the amount of overspill and sediment erosion at this location. There was substantially less initial scour in the proximal section of the asymmetric channel (Fig. 8b). This is due to the increased distance between the start of the channel and the first bend. The Series 1 experimental channels begin at the apex of a bend, whereas the Series 2 channel had an initial straight section before the first channel bend. Sufficient energy will have been lost through overspill in this initial straight section of channel that less scour occurred at Bend 1 compared with the most proximal sections of channel in the Series 1 experiments. A greater amount of inner-bend deposition was observed in the S ¼ 1.36 channel than the S ¼ 1.14 channel at Bend 1, but at Bend 2, more inner-bend deposition was observed in the S ¼ 1.14 channel than the S ¼ 1.36 channel. This is likely to reflect a combination of the amount of sediment being eroded immediately upstream, and the amount of sediment being lost through overspill. Overspill increased with sinuosity, which will result in a more rapid decrease in flow competence in the higher sinuosity channels. This could explain why the S ¼ 1.14 channel here has greater sediment deposition at Bend 2 than the S ¼ 1.36 channel. The amount of deposition occurring at Bend 1 appears to be most directly related to the amount of scour occurring immediately upstream, at the apex of the bend.

As described above, these experiments show similar basaloutward helicity to the detailed investigations on flow dynamics of Corney et al. (2006) and Keevil et al. (2006, 2007) and similar flow and deposit characteristics to the sediment-laden experiments of Peakall et al. (2007a). This strengthens the hypothesis that in many natural turbidity currents which have velocity maxima close to the base of the flow, and are capable of eroding sediment from their bed and transporting sediment by traction, flow circulation is basaloutward. It is important to note though that since longitudinal flow dominates point-bar deposits will accrete on the inside of bends whether basal helical flows are inward- or outward directed; they will vary in their position though (Peakall et al., 2007a). Natural flows evolve with time, however, and can both increase and decrease their erosive energy, velocity and suspended load. As a turbidity current slows and its sediment load decreases, the helicity of flow might change from basal-outward to basal-inward (e.g. Imran et al., 2007), driven by an increase in the height of the vertical velocity maximum (Corney et al., 2006; 2008). However, this is likely to have a reduced impact on the channel-fill deposit due to its low energy and lower sediment concentration. Experiments on sediment-laden flows through sinuous channels that were depositing primarily by suspension fallout (Straub et al., 2008) identify inertial-dominated run-up of gravity currents at the outer bank of sinuous channel bends as a key influence on flow superelevation and sediment deposition, as sediment was deposited on the outside of bends. As the energy of a natural turbidity current wanes with time and distance downstream, it is likely that most channelised flows will change temporally and spatially from traction-dominated sediment deposition to suspension-falloutdominated deposition, resulting in the accretion of a channel-fill which is initially accreted as point bars close to the inside bend forming Lateral Accretion Packages (LAPs), and then a switch to sedimentation on the outside of the bend. Even the most laterally migrated sequences are commonly aggrading, from the convex to the concave side of sinuous loops (Kolla et al., 2007). This implies that the bed morphology observed in our experiments, with innerbend accretion and outer-bend erosion, is representative of natural channels which are aggrading as they migrate laterally. It has been observed in these experiments and in those of Straub et al. (2008) that basal fluid within the channel is transported out of the channel over the outer-levee crest. Overbank deposits were not investigated in these experiments, although sediment was transported out of the channel onto the platform on which the channel models were mounted. We anticipate that the sediment transported overbank on the outside of each bend would have the same grain size range of that being reworked within the channel, as was observed in the suspension-fallout-dominated experiments of Straub et al. (2008). Similarity of grain size between the levee and channel has been concluded from an investigation from the RamPowell field in the Gulf of Mexico (Clemenceau et al., 2000). Similarly, an outcrop study from the Ross Formation in southwest Ireland has interpreted levee crest deposits that are as coarse as the channelised deposits (Lien et al., 2003). Sinuosity evolution of subsurface submarine channels is thought to be due to repeated channel aggradation and subsequent lateral migration, rather than lateral migration alone, as seen in fluvial channels (Peakall et al., 2000b; Kolla et al., 2001). Submarine channels have a range of migration patterns. Erosion-dominated channels may show most similarity with meandering rivers with downstream migration and bend cut-offs (Abreu et al., 2003). Aggradational channels have greatly reduced bend cut-offs (Peakall et al., 2000b), although the Bengal Fan is an exception to this (Schwenk et al., 2003). Aggradational channels show increases in

7.5. The influence of channel planform asymmetry on flow and sediment dynamics The experimental investigation of asymmetric channel planforms on flow and sediment dynamics is substantially more difficult than for symmetrical-bend channels such as those used in our Series 1 experiments. Flow and sediment characteristics recorded within the asymmetric-planform channel model used in these experiments are dominated by the high sinuosity of the channel, more specifically by the tightness of its initial bend and the influence of this on overspill volumes. Subsequent experimental investigation of channels with asymmetric planform would need to investigate flows with lower initial discharge, mounting of the channel on a sloping platform, and/or necessitate a larger flume.

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

1443

Table 3 A general model for the characteristics of turbidity current deposits in sinuous submarine channels. Note: this model for deposition within sinuous submarine channels is applicable to flows with basal-outwards or basal-inwards helicity, because the spatial variations in-channel deposition and evolution are dominated by longitudinal flow. Although some differences will occur (e.g. in the position of the inside bend point-bar relative to the bend apex), the patterns of deposition will be broadly similar. Energy

High

Low

Dominant sediment transport mechanism(s) Dominant mechanism of sediment deposition

Traction, saltation, suspension Traction

Suspension Suspension-fallout

In-channel deposit topography

- inner-bend point bars form just downstream of apex - scour occurs around outside bend

- greatest deposition occurs around outside of bends

In-channel deposit architecture

Planform evolution

References

- sinuosity likely to increase with time This Study; Kane et al., 2008; Peakall et al., 2007a;

- sinuosity might decrease slightly Straub et al., 2008; Kane et al., 2008;

Schematic model depicting evolution of a sinuous submarine channel with decreasing energy

- levee sediments indistinguishable from in-channel material on outer-levee, finer on inside levee, fining upwards as flow energy wanes. - Coarser grain sizes are more likely to be found on the outer-bank levee proximal to higher amplitude bends

bend amplitude but limited downstream sweep. Factors important in controlling submarine channel sinuosity evolution include seafloor gradients, flow velocities, sediment load and concentration, sediment grain size, flow density, flow frequency and duration, the dominant mechanism of sediment transport and deposition (bedload vs. suspended load), flow dynamics (influenced by vertical velocity profile, centrifugal and Coriolis forces),

channel shape, bank cohesiveness and stability, and base-level changes. Key to the observations presented here regarding changing channel planform is that in nature, channel and flow are intrinsically linked and self-regulating. As discussed above, these experiments show increasing amounts of overspill lost at bends as the bend amplitude increases (with the same initial discharge). This

1444

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

effect can also be observed in the experiments of Straub et al. (2008) where larger flows lost greater amounts of fluid through overspill at channel bends. In nature, this would be mitigated somewhat by the channels ability to adjust its cross-section. As discharge or sinuosity increase, increased overspill is likely to result in increased sedimentation on levees and levee build-up such that the channel equilibrates to the flow conditions. Under the same flow conditions, levees around more sinuous bends would be expected to be higher, acting to constrain more flow within the channel and increase efficiency of transport downstream. This would result in a shift in the location of overspill and subsequent levee accretion downstream with time. One of the key implications of these experiments for natural channels is that more deposition of coarser material might be expected on levees and in overbank regions close to higher amplitude bends. However, in nature this will be mitigated by the building of higher levees. Nonetheless, during unusually large flow events, more material is likely to be transported overbank close to tighter bends, thus a more rapid downstream decrease in sediment load and grain size might be expected in higher sinuosity channels. Thus higher sinuosity channels will ‘tune’ outsize flows down to a characteristic flow scale more effectively and more rapidly than less sinuous channels. Dependant on the regional slope and levee slope, higher sinuosity channels are also more likely to have within-channel interaction between re-entry of overspilled fluid and channelised flow. This might result in a higher proportion of anomalous flow dynamics and sedimentation patterns in higher sinuosity channels. As sinuosity increases and bend asymmetry grows, then flows may begin to switch their secondary flow orientation, with progressively more flows or greater temporal periods of flows becoming reversed relative to their initial orientation. It has been previously noted that certain types of submarine channels, particularly aggradational high-sinuosity systems, reach a point where there is a near cessation of planform movement (ossification; Peakall et al., 2000b; Wynn et al., 2007). The underlying mechanisms controlling ossification are poorly understood although a number of potential factors have been identified (Peakall et al., 2000b). Such change in flow-cell polarity, as identified through this experimental programme, is one putative mechanism since this would affect point-bar sedimentation and outer-bank erosion. If the intensity or frequency of flow-cell reversal is linked to the sinuosity, bend asymmetry and complexity of the meander bend, there may be a negative feedback mechanism that serves to stabilise meanders at certain highly sinuous planform geometries. 7.6.1. Summary of results In all of the experimental channels investigated in this study, basal-outwards flow helicity produced a characteristic bed morphology different to that which forms under most conditions in meandering rivers. This was previously observed using one of these experimental channels by Peakall et al. (2007a). Point bars form at the inside of bends, and initiate at or just downstream of the bend apex, further downstream than fluvial point bars in equivalent channel bend geometries. The rate of erosion and deposition decreased with time, indicating that these deposits are moving towards a steady-state deposit that is in equilibrium with flow conditions. Discharge loss at bends through overspill is important. An increasing loss of fluid through overspill with increasing channel sinuosity results in a decreasing magnitude of cross-stream velocity downstream. This results in a decreasing amount of erosion and deposition with increasing channel sinuosity, and decreasing transverse slopes of in-channel deposits. No simple relationship was observed between superelevation and sinuosity, probably due

to changes in the relative influences of downstream velocity and bend curvature on centrifugal force and inertial run-up. As sinuosity increases, overspill will increase, thus higher sinuosity channels will ‘tune’ outsize flows down to a characteristic flow scale more effectively and more rapidly than less sinuous channels. As sinuosity increases, the energy of channelised flow downstream will decrease and thus it is more likely that sedimentation becomes dominated by suspension fallout, and thicker deposits accrete around the outside of bends. This could potentially form a mechanism for the onset of channel ossification. In the channel with the tightest initial bend curvature, overspill fluid from Bend 1 re-entered the channel at Bend 2. This dominated flow characteristics at this bend, disrupting the basal-outwards helicity observed in the other channels. With similar regional slope and levee slope, higher sinuosity channels are more likely to have within-channel interaction between re-entry of overspill fluid and channelised flow, which might result in a higher proportion of anomalous flow dynamics and sedimentation patterns, and possibly form another mechanism for the onset of ossification. 8. An updated model for sediment deposition in sinuous submarine channels The results of this experimental investigation and comparison of these with published literature on fluvial and submarine channels as discussed above enables the synthesis of a new model for sedimentation in submarine channel bends (represented schematically in Table 3):  In flows with sufficient energy that sedimentation is dominated by traction, sediment accumulates on the inside bend and is eroded from the outside bend, similar to fluvial pointbars and pools (this study; Peakall et al., 2007a). Helical flow is likely to have a basal-outwards sense of rotation, and point bars form further downstream than in fluvial systems, downstream of the apex of the bend. This results in low cross-channel gradients that are relatively steep at the upstream parts of point bars. Downstream of the bend apex, basal flow is inwardly dominated (this study and Keevil et al., 2006), which might result in more rapid downstream change to finingupwards successions than in fluvial channel deposits (Peakall et al., 2007a).  As the turbidity current energy wanes with time and distance downstream, flow is likely to change temporally and spatially from traction-dominated sediment deposition to suspensionfallout-dominated deposition. In suspension-fallout-dominated flows, inertial-dominated run-up at the outer-bank influences sediment deposition, with sediment being deposited on the outside of bends where near-bed suspended sediment concentrations are greatest (Das et al., 2004; Straub et al., 2008).  As the turbidity current slows and its sediment load decreases, the helicity of flow might change from basal-outward to basalinward, driven by an increase in the height of the vertical velocity maximum (e.g. Imran et al., 2007; Corney et al., 2006; 2008). Due to their low energy and sediment load, these periods of flow are likely to have a reduced influence on deposits.  A channel-fill sequence is thus likely to initially accrete as point bars close to the inside bend forming lateral accretion packages (LAPs). During this period of channel development, bends are likely to migrate laterally, increasing channel sinuosity. As flows decrease in energy, this style of deposition will be followed by a switch to sedimentation on the outside of the bend due to suspension-fallout-dominated deposition (e.g. Straub

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

et al., 2008). Channels are then most likely to aggrade vertically with little change in the channel planform, and might reduce their sinuosity (Straub et al., 2008; Kane et al., 2008). Such a pattern is commonly observed in the natural environment (e. g. Peakall et al., 2000a,b). In both erosive and depositional, traction-dominated and fallout-dominated flows, overspill over the outside levee includes basal fluid from within the channel, and the sediment transported out of the channel will have the same grain size range of that being reworked within the channel (this study; Straub et al., 2008). Within this overall pattern of accretion and channel filling, the precise characteristics of accretion will vary as the result of relative influences of flow helicity and run-up, resulting from differences in flow characteristics (energy, density) and bend geometry. This model for sedimentation in sinuous submarine channels is applicable over different temporal and spatial scales. As described above, the model applies to flow and deposition at a point during a single flow event with decreasing energy. The model can also be applied to a series of flow events through the same channel that decrease in energy over time. Such longer time-scale decreases in the energy of flows through a submarine channel will occur during abandonment following avulsion, and during marine transgression as the distance between the channel location and the sediment source increases. The model can also be applied spatially, to help understand changes in flow energy and deposition downstream. The proposed model highlights a distinct difference between the mechanisms influencing slope:sinuosity relationships in rivers and submarine channels. In both rivers and submarine channels, sinuosity has been observed globally to increase to a threshold and then decrease with decreasing valley slope downstream (Schumm et al., 1972; Clark et al., 1992). In rivers, suspension-dominated channels have higher sinuosity thresholds than bedload-dominated rivers (Schumm, 1981). With this model, we propose that the opposite is true for submarine channels, where bedload-dominated channels are able to laterally migrate and are thus likely to have higher sinuosities than suspension-dominated channels which will accrete vertically with little or no lateral migration. Another difference between rivers and submarine channels not previously mentioned is that when sinuosity/fluvial style is plotted against slope, rivers generally depict a trend from braided at high slopes, to meandering and then straight at lowest slopes (Schumm, 1981). Submarine channels tend to have highest sinuosity at midfan (mid-gradient) slopes (e.g. Damuth and Flood, 1984; Clark et al., 1992; Pirmez and Imran, 2003). In addition, well constrained examples of channel braiding or intra-channel braid bars have not been described from sinuous modern deep-water channels (Wynn et al., 2007). The model of channel evolution postulated here can be applied to help understand spatial changes in sinuosity on submarine fans. In a simple fan system with decreasing slope and turbidity current energy downstream, in which bypass is most likely in proximal locations: i) Channels in proximal locations will show limited migration due to lack of sedimentation. ii) Further down the system there is less bypass and deposited sediments may be tractionally reworked. This results in the growth of point-bars and lateral migration of channels and thus an increase in sinuosity down-fan. iii) Even further down the system, as energy decreases there will be increasingly less traction and more suspension-fallout. This will result in decreasing lateral migration of bends down-fan. This simple model provides a mechanism by which sinuosity increases to a mid-fan peak, and then decreases downstream, as observed in the natural environment (e.g. Damuth and Flood, 1984; Clark et al., 1992; Pirmez and Imran, 2003). However, the factors controlling sinuosity in the natural environment are complex. For example,

1445

Babonneau et al. (2002) conclude that for the present Zaire Channel, although down-slope decreases in energy of turbidity currents induces a decrease in sinuosity, the decrease in sinuosity down-slope is mainly due to channel avulsion and the relative maturity of channel sections. After avulsion the channel is relatively straight and slope gradient is relatively high, which is compensated for by a subsequent increase in sinuosity. As the distal channellevee system is younger (more recently avulsed) than the proximal channel-levee system, sinuosity decreases with distance. This model and ours are not mutually exclusive: i) a local increase in slope may increase tractional reworking, point-bar development and channel sinuosity; ii) given time, in a prograding system there will be an increase in the influence of tractional reworking at a point, and therefore an increase in sinuosity. Other factors shown to influence sinuosity in rivers include valley slope, stream power, discharge, sediment load, bed material size, channel width:depth ratio and the characteristics of the underlying material being eroded into by the channel (e.g. Bridge, 2003 and references therein). Many of these are likely to be influencing factors in submarine systems. Clark et al. (1992) found no obvious trends between channel width:depth and sinuosity, and suggest that this is because unlike rivers, turbidity currents generally overspill the confines of the channel by heights and widths much greater than that of the channel. However, it has been shown that degree of confinement does influence sedimentation within sinuous submarine channels, which is likely to influence sinuosity (Kane et al., 2008). Submarine channel sinuosity has been observed to increase in response to a decrease in slope gradient (Amazon Fan, Damuth et al., 1983; late Miocene Lower Congo Basin, Ferry et al., 2005), related to a decrease in flow energy and erosional intensity, and coincident with levee development and shallower incision of the channel complex (Ferry et al., 2005). This relationship is similar to that between slope and sinuosity observed in rivers (Schumm, 1986, 1992). The opposite occurs where a channel adapts to an increase in slope by decreasing sinuosity (Ferry et al., 2005). These sinuosity changes could be related to changes in the influence of tractional reworking vs. suspension fallout, as described above. A decrease in flow energy due to a decreased slope will result in greater aggradation. If the increase in aggradation results in levee development, such as has been observed in a late Miocene channel system in the Lower Congo Basin (Ferry et al., 2005), this would result in increasing flow confinement, tractional reworking, lateral migration and thus higher sinuosities. Alternatively, an increase in inner-bend deposition downstream of the change in slope will result in increased lateral migration and thus higher sinuosities.

8.1. Influence of sinuosity on sedimentation During the traction-dominated phase of turbidity current flow, as discharge and/or sinuosity increase, there is increasing loss through overspill and corresponding decreases in the magnitude of flow velocities and amount of erosion and deposition at downstream locations. For a flow with the same initial energy, transverse bed slopes increase with decreasing sinuosity. The amount of sediment deposited around bends is related to the amount of sediment being eroded from immediately upstream and the amount of sediment being lost through overspill, which are influenced by bend amplitude. Flow dynamics around downstream bends will be influenced by flow energy and bend amplitude, such that increased overspill at a tight initial bend might reduce flow energy sufficiently that superelevation is lower in downstream bends than would occur in bends downstream of a lower-amplitude initial bend.

1446

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447

Increasing amounts of fluid and suspended sediment are transported overbank in more sinuous channels. This is likely to result in increased sedimentation on levees and levee build-up. One of the key implications of these experiments for better understanding natural sinuous submarine channels is that more deposition of coarse material might be expected on levees and in overbank regions close to higher amplitude bends. Levees around more sinuous bends would be expected to be higher, increasing efficiency of transport downstream and resulting in a shift in the location of overspill and subsequent levee accretion downstream with time. Channel planform geometry probably reflects the most frequent flow magnitude passing through the system (Kane et al., 2008), yet during infrequent or unusually large flow events, more material is likely to be transported overbank close to tighter bends than in less tight bends. A more rapid downstream decrease in sediment load and grain size might thus be expected in higher sinuosity channels, and higher sinuosity channels will ‘tune’ outsize flows down to a characteristic flow scale more effectively and more rapidly than less sinuous channels. In rivers, sinuosity has been shown to vary with changes in sediment load and discharge downstream of tributary confluences (e.g. Timár, 2003). We discussed above that increased overspill at high amplitude bends will influence the discharge and sediment load of the downstream reach, resulting in cross-section change. It is also likely that downstream changes in discharge and sediment load due to overspill will result in sinuosity changes. Higher sinuosity channels are more likely to have withinchannel interaction between re-entry of overspilled fluid and channelised flow. This might result in a higher proportion of anomalous flow dynamics and sedimentation patterns in higher sinuosity channels. Re-entry of overspill fluid into the channel downstream might completely dominate flow dynamics, resulting in a change of flow-cell polarity, and result in suppression of sediment reworking. Acknowledgements This work was funded by the UK Natural Environment Research Council (NERC) and Total (grant NER/T/S/2000/01400 Ocean Margins with Total as a LINK partner). The development of the UDVP’s was funded through NERC grant GR3/10015. Development of the laboratory facilities was funded by NERC and a consortium of oil companies comprising Amerada Hess, BG, BHP, BP, Chevron, ConocoPhillips, ExxonMobil, Total and Shell. We would like to thank Jo Ann Hegre of Total and Alick Leslie of NERC for their support during this work. Thanks to Mark Franklin for assistance in the laboratory. We thank Kyle Straub for discussions on intrachannel flow processes. We also thank Associate Editor Allard W. Martinius, reviewer Sverre Henriksen and an anonymous reviewer for their consideration of this manuscript. ‘KJA’s contribution forms TRaX Record #97. References Abreu, V., Sullivan, M., Pirmez, C., Mohrig, D., 2003. Lateral accretion packages (LAP’s): an important reservoir element in deep water sinuous channels. Marine and Petroleum Geology 20, 631e648. Antobreh, A.A., Krastel, S., 2006. Morphology, seismic characteristics and development of Cap Timiris Canyon, offshore Mauritania: a newly discovered canyon preserved-off a major arid climatic region. Marine and Petroleum Geology 23, 37e59. Babonneau, N., Savoye, B., Cremer, M., Klein, B., 2002. Morphology and architecture of the present canyon and channel system of the Zaire deep-sea fan. Marine and Petroleum Geology 19, 445e467. Bagnold, R.A., 1960. Some aspects of the shape of river meanders. US Geological Survey. Professional Paper 282E135e144. Best, J.L., Ashworth, P.J., 1994. A high resolution ultrasonic bed profiler for use in laboratory flumes. Journal of Sedimentary Research A64, 674e675.

Best, J.L., Kirkbride, A.D., Peakall, J., 2001. Mean flow and turbulence structure of sediment-laden gravity currents: new insights using ultrasonic Doppler velocity profiling. In: McCaffrey, W.D., Kneller, B.C., Peakall, J. (Eds.), Particulate Gravity Currents. International Association of Sedimentology Special Publication 31, pp. 159e172. Bridge, J.S., 2003. Rivers and floodplains: forms, processes, and sedimentary record. Blackwell Science Ltd., UK. Buckee, C., Kneller, B., Peakall, J., 2001. Turbulence structure in steady solute-driven gravity currents. In: McCaffrey, W.D., Kneller, B.C., Peakall, J. (Eds.), Particulate Gravity Currents. International Association of Sedimentologists Special Publication, vol. 31, pp. 173e188. Chorley, R.J., 1967. Models in Geomorphology. In: Chorley, R.J., Haggett, P. (Eds.), Models in Geography. Methuen, London, pp. 59e96. Clark, J.D., Pickering, K.T., 1996. Submarine channels: processes and architecture. Vallis Press, London, 231 pp. Clark, J.D., Kenyon, N.H., Pickering, K.T., 1992. Quantitative analysis of the geometry of submarine channels: implications for the classification of submarine fans. Geology 20, 633e636. Clemenceau, G.R., Colbert, J., Eddens, J., 2000. Production results from levee-overbank turbidite sands at Ram/Powell Field, deepwater Gulf of Mexico. In: Weimer, P., Slatt, R.M., Coleman, J., Rosen, N.C., Nelson, H., Bouma, A.H., Styzen, M.J., Lawrence, D.T. (Eds.), Deep-water Reservoirs of the World, GCSSEPM Foundation 20th Annual Bob F. Perkins Research Conference, pp. 241e251. Cook, T.W., Bouma, A.H., Chapin, M.A., Zhu, H., 1994. Facies architecture and reservoir characterization of a submarine fan channel complex, Jackfork Formation, Arkansas. In: Weimer, P., Bouma, A.H., Perkins, B.F. (Eds.), Submarine Fans and Turbidite Systems. Gulf Coast SEPM 15th Annual Research Conference, pp. 69e81. Corney, R.K.T., Peakall, J., Parsons, D.R., Elliott, L., Amos, K.J., Best, J.L., Keevil, G.M., Ingham, D.B., 2006. The orientation of helical flow in curved channels. Sedimentology 53, 249e257. doi:10.1111/j.1365-3091.2006.00771.x. Corney, R.K.T., Peakall, J., Parsons, D.R., Elliott, L., Best, J.L., Thomas, R.E., Keevil, G.M., Ingham, D.B., Amos, K.J., 2008. Reply to discussion of Imran et al. on “The orientation of helical flow in curved channels” by Corney, et al. Sedimentology 53, 249e257. Sedimentology, 55, 241e7. Damuth, J.E., Kolla, V., Flood, R.D., Kowsmann, R.O., Monteiro, M.C., Gorini, M.A., Palma, J.J.C., Belderson, R.H., 1983. Distributary channel meandering and bifurcation patterns on the Amazon deep-sea fan as revealed by long-range side-scan sonar (GLORIA). Geology 11, 94e98. Damuth, J.E., Flood, R.D., 1984. Morphology, sedimentation processes and growth pattern of the Amazon Deep-Sea Fan. Geo-Marine Letters 3, 109e117. Das, H.S., Imran, J., Pirmez, C., Mohrig, D.C., 2004. Numerical modeling of flow and bed evolution in meandering submarine channels. Journal of Geophysical Research-Oceans C10, C10009. Dietrich, W.E., 1987. Mechanics of Flow and Sediment Transport in River Bends. In: Richards, K. (Ed.), River ChannelsdEnvironment and Process. Blackwell, Oxford, pp. 129e227. Dykstra, M., Kneller, B., 2009. Lateral accretion in a deep-marine channel complex: implications for channelized flow processes in turbidity currents. Sedimentology 56, 1411e1432. Elliott, T., 2000. Depositional architecture of a sand-rich, channelised turbidite system: the Upper Carboniferous Ross Sandstone Formation, western Ireland. In: Weimer, P., Slatt, R.M., Bouma, A.H., Lawrence, D.T. (Eds.), Deep-water Reservoirs of the World. Gulf Coast Section SEPM 20th Annual Research Conference, pp. 342e373. Felix, M., 2002. Flow structure of turbidity currents. Sedimentology 49, 397e419. Felix, M., 2004. The significance of single value variables in turbidity currents. Journal of Hydraulic Research 42, 323e330. Ferguson, R.I., Parsons, D.R., Lane, S.N., Hardy, R.J., 2003. Flow in meander bends with recirculation at the inner bank. Water Resources Research 39 (11), 1322. Ferry, J.-N., Mulder, T., Parize, O., Raillard, S., 2005. Concept of equilibrium profile in deep-water turbidite systems: effects of local physiographic changes on the nature of sedimentary process and the geometries of deposits. In: Hodgson, D. M., Flint, S.S. (Eds.), Submarine Slope Systems: Processes and Products. Geological Society London Special Publication, vol. 244, pp. 181e193. Garcia, M., Parker, G., 1993. Experiments on the entrainment of sediment into suspension by a dense bottom current. Journal of Geophysical Research 98, 4793e4807. Haughton, P.D.W., 2000. Evolving turbidite systems on a deforming basin floor, Tabernas, SE Spain. Sedimentology 47, 497e518. Hesse, R., Rakofsky, A., 1992. Deep-sea channel/submarine Yazoo system of the Labrador Sea: a new deep-water facies model. AAPG Bulletin 76, 680e707. Hooke, R.L., 1968. Model geology: prototype and laboratory streams: discussion. Geological Society of America Bulletin 79, 391e394. Hooke, R.L., 1975. Distribution of sediment transport and shear stress in a meander bend. The Journal of Geology 83, 543e565. Imran, J., Islam, M.A., Huang, H., Kassem, A., Dickerson, J., Pirmez, C., Parker, G., 2007. Helical flow couplets in submarine gravity underflows. Geology 35, 659e662. Islam, M.A., Imran, J., 2008. Experimental modeling of gravity underflow in a sinuous submerged channel. Journal of Geophysical Research 113, C07041. doi:10.1029/2007JC004292. Islam, M.A., Imran, J., Pirmez, C., Cantelli, A., 2008. Flow splitting modifies the helical motion in submarine channels. Geophysical Research Letters 35, L22603. doi:10.1029/2008GL034995.

K.J. Amos et al. / Marine and Petroleum Geology 27 (2010) 1431e1447 Kane, I.A., 2007. Architecture and sedimentology of submarine channel-levee systems: insights from the Upper Cretaceous Rosario Formation, Baja California, Mexico, and from laboratory experiments. Unpublished Ph.D thesis, University of Leeds, Leeds, UK, 310 pp. Kane, I.A., McCaffrey, W.D., Peakall, J., 2008. Controls on sinuosity evolution within submarine channels. Geology 36, 287e290. Kane, I.A., McCaffrey, W.D., Peakall, J., Kneller, B.C., 2010. Submarine channel levee shape and sediment waves from physical experiments. Sedimentary Geology 223, 75e85. Kassem, A., Imran, J., 2004. Three-dimensional modeling of density current, II. Flow in sinuous confined and unconfined channels. Journal of Hydraulic Research 42 (6), 591e602. Keevil, G.M., Peakall, J., Best, J.L., Amos, K.J., 2006. Flow structure in sinuous submarine channels: velocity and turbulence structure of an experimental submarine channel. Marine Geology 229, 241e257. doi:10.1016/j. margeo.2006.03.010. Keevil, G.M., Peakall, J., Best, J.L., 2007. The influence of scale, slope and channel geometry on the flow dynamics of submarine channels. Marine and Petroleum Geology 24, 487e503. Khripounoff, A., Vangriesheim, A., Babonneau, N., Crassous, P., Dennielou, B., Savoye, B., 2003. Direct observation of intense turbidity current activity in the Zaire submarine valley at 4000 m water depth. Marine Geology 194, 151e158. Klaucke, I., Hesse, R., 1996. Fluvial features in the deep-sea: new insights from the glacigenic submarine drainage system of the Northwest Atlantic Mid-Ocean Channel in the Labrador Sea. Sedimentary Geology 106, 223e234. Kneller, B.C., 2003. The influence of flow parameters on turbidite slope channel architecture. Marine and Petroleum Geology 20, 901e910. Knight, D.W., Shiono, K., 1990. Turbulence measurements in a shear layer region of a compound channel. Journal of Hydraulic Research 28, 175e196. Kolla, V., Bourges, P., Urruty, J.-M., Safa, P., 2001. Evolution of deep-water Tertiary sinuous channels offshore Angola (west Africa) and implications for reservoir architecture. AAPG Bulletin 85, 1373e1405. Kolla, V., Posamentier, H.W., Wood, L.J., 2007. Deep-water and fluvial sinuous channelseCharacteristics, similarities and dissimilarities, and modes of formation. Marine and Petroleum Geology 24, 388e405. Leopold, L.B., Bagnold, R.A., Wolman, M.G., Brush, L.M., 1960. Flow resistance in sinuous and irregular channels. US Geological Survey. Professional Paper 282D111e134. Leopold, L.B., Wolman, M.G., Miller, J.P., 1964. Fluvial Processes in Geomorphology. Dover Publications, New York, 522 pp. Lien, T., Walker, R.G., Martinsen, O.J., 2003. Turbidites in the Upper Carboniferous Ross Formation, western Ireland: reconstruction of a channel and spillover system. Sedimentology 50, 113e148. Mayall, M., Stewart, I., 2000. The architecture of turbidite slope channels. In: Weimer, P., Slatt, R.M., Bouma, A.H., Lawrence, D.T. (Eds.), Deep-water Reservoirs of the World. Gulf Coast Section SEPM 20th Annual Research Conference, pp. 578e586. Mayall, M., Jones, E., Casey, M., 2006. Turbidite channel reservoirsdkey elements in facies prediction and effective development. Marine and Petroleum Geology 23, 821e841. Nakajima, T., Peakall, J., McCaffrey, W.D., Paton, D.A., Thompson, P.J., 2009. Outerbank bars: a new intra-channel architectural element within sinuous submarine slope channels. Journal of Sedimentary Research 79, 872e886. Naot, D., Rodi, W., 1982. Calculation of secondary currents in channel flow. Journal of the Hydraulics Division, Am. Soc. Civ. Eng 108, 948e970. Nelson, J.M., Smith, J.D., 1989. Evolution and stability of erodible channel bends. In: Ikeda, S., Parker, G. (Eds.), River Meandering. AGU, pp. 948e970. Nezu, I., Nakagawa, H., 1993. Turbulence in Open-channel Flows. In: International Association for Hydraulic Research Monograph Series. A.A. Balkema, 281 pp. Normark, W.R., 1989. Observed parameters for turbidity-current flow in channels, Reserve Fan, Lake Superior. Journal of Sedimentary Petrology 59, 423e431. Paull, C.K., Ussler, W., Greene, H.G., Keaten, R., Mitts, P., Barry, J., 2003. Caught in the act: the 20 December 2001 gravity flow event in Monterey Canyon. Geo-Marine Letters 22, 227e232. Peakall, J., Ashworth, P., Best, J., 1996. Physical Modelling in Fluvial Geomorphology: Principles, Applications and Unresolved Issues. In: Rhoads, B.L., Thorn, C.E. (Eds.), The Scientific Nature of Geomorphology. John Wiley and Sons, Chichester, pp. 221e253. Peakall, J., McCaffrey, W.D., Kneller, B.C., Stelting, C.E., McHargue, T.R., Schweller, W. J., 2000a. A process model for the evolution of submarine fan channels: implications for sedimentary architecture. In: Bouma, A.H., Stone, C.G. (Eds.), Fine-grained Turbidite Systems. AAPG Memoir 72/SEPM Special Publication, vol. 68, pp. 73e88.

1447

Peakall, J., McCaffrey, W.D., Kneller, B.C., 2000b. A process model for the evolution, morphology, and architecture of sinuous submarine channels. Journal of Sedimentary Research 70, 434e448. Peakall, J., Amos, K.J., Keevil, G., Gupta, S. and Laursen, Y., 2005. Final Report to Total: Architecture and controls of sinuous submarine channels. Unpublished Confidential Industry Report. Peakall, J., Amos, K.J., Keevil, G.M., Bradbury, P.W., Gupta, S., 2007a. Flow processes and sedimentation in submarine channel bends. Marine and Petroleum Geology 24, 470e486. Peakall, J., Ashworth, P.J., Best, J.L., 2007b. Meander-bend evolution, alluvial architecture, and the role of cohesion in sinuous river channels: a flume study. Journal of Sedimentary Research 77, 197e212. doi:10.2110/jsr.2007.017. Phillips, S., 1987. Dipmeter interpretation of turbidite-channel reservoir sandstones, Indian Draw Field, New Mexico. In: Tillman, R.W., Weber, K.J. (Eds.), Reservoir Sedimentology. SEPM Special Publication, vol. 40, pp. 113e128. Pirmez, C., Imran, J., 2003. Reconstruction of turbidity currents in Amazon Channel. Marine and Petroleum Geology 20, 823e849. Prior, D.B., Bornhold, B.D., Wiseman, W.J., Lowe, D.R., 1987. Turbidity current activity in a British Columbia fjord. Science 237, 1330e1333. Pyles, D., Jennette, D.C., Tomasso, M., Beaubouef, R.T., Rossen, C., 2010. Concepts learned from a 3D outcrop of a sinuous slope channel complex: Beacon channel complex, Brushy canyon formation, west Texas. Journal of Sedimentary Research 80, 67e96. Schumm, S.A., 1981. Evolution and response of the fluvial system, sedimentologic implications. In: Ethridge, F.G., Flores, R.M. (Eds.), Recent and Ancient Nonmarine Depositional Environments: Models for Exploration. SEPM Special Publication, vol 31, pp. 19e29. Schumm, S.A., 1986. Alluvial River Response to Active Tectonics. In: Wallace, R.E. (Ed.), Studies in Geophysics e Active Tectonics. National Academy Press, Washington DC, pp. 80e84. Schumm, S.A., 1992. River response to baselevel change: implication for sequence Stratigraphy. Journal of Geology 101, 279e294. Schumm, S.A., Khan, H.R., 1972. Experimental study of channel patterns. Geological Society of America Bulletin 88, 1755e1770. Schumm, S.A., Khan, H.R., Winkley, B.R., Robbins, L.G., 1972. Variability of river patterns. Nature, Physical Science 237, 75e76. Schwenk, T., Spieß, V., Hübscher, C., Breitzke, M., 2003. Frequent channel avulsions within the active channel-levee system of the middle Bengal Fan e an exceptional channel-levee development derived from Parasound and Hydrosweep data. Deep-sea Research Part II 50, 1023e1045. Sellin, R.H.J., Ervine, D.A., Willetts, B.B., 1993. Behaviour of meandering two-stage channels. Proceedings of the Institute of Civil Engineers, Water, Maritime and Energy 101, 99e111. Stacey, M.W., Bowen, A.J., 1988. The vertical structure of density and turbidity currents: theory and observations. Journal of Geophysical Research 93, 3528e3542. Straub, K.M., Mohrig, D., McElroy, B., Buttles, J., 2008. Interactions between turbidity currents and topography in aggrading sinuous submarine channels: a laboratory study. GSA Bulletin 120, 368e385. Tesaker, E., 1969. Uniform turbidity currents. Unpublished doctoral thesis, Technical University of Norway, Trondheim. Timár, G., 2003. Controls on channel sinuosity changes: a case study of the Tisza River, the Great Hungarian Plain. Quaternary Science Reviews 22, 2199e2207. Timbrell, G., 1993. Sandstone architecture of the Balder Formation depositional system, UK Quadrant 9 and adjacent areas. In: Parker, J.R. (Ed.), Petroleum Geology of Northwest Europe, Proceedings of the Fourth Conference. Geological Society of London, pp. 107e121. Vangriesheim, A., Khripounoff, A., Crassous, P., 2009. Turbidity events observed in situ along the Congo Submarine Channel. Deep Sea Research II 56, 2208e2222. Willetts, B.B., Rameshwaran, P., 1996. Meandering overbank flow structures. In: Ashworth, P.J., Best, J.L., McLelland, S.J. (Eds.), Coherent Flow Structures in Open Channels. John Wiley and Sons, pp. 609e629. Wormleaton, P.R., Sellin, R.H.J., Bryant, T., Loveless, J.H., Hey, R.D., Catmur, S.E., 2004. Flow structures in a two-stage channel with a mobile bed. Journal of Hydraulic Research 42, 145e162. Wormleaton, P.R., Hey, R.D., Sellin, R.H.J., Bryant, T., Loveless, J.H., Catmur, S.E., 2005. Behaviour of meandering overbank channels with graded sand beds. Journal of Hydraulic Engineering 131, 665e681. Wynn, R.B., Cronin, B.T., Peakall, J., 2007. Sinuous deep-water channels: genesis, geometry and architecture. Marine and Petroleum Geology 24, 341e387. Xu, J.P., Noble, M.A., Rosenfeld, L.K., 2004. In-situ measurements of velocity structure within turbidity currents. Geophysical Research Letters 31, L09311. doi:10.1029/2004GL019718.