at SciVerse ScienceDirect

Marine and Petroleum Geology 41 (2013) 7e34

Contents lists available

Marine and Petroleum Geology

journal homepage: www.elsevier .com/locate/marpetgeo

The diversity of deep-water sinuous channel belts and slope valley-fill complexes

M. Janocko a,*, W. Nemec a, S. Henriksen b, M. Warcho1 c

aDepartment of Earth Science, University of Bergen, Allégaten 41, 5007 Bergen, Norwayb Statoil Research Centre, Arkitekt Ebbels Veg 10, Rotvoll, 7005 Trondheim, Norwayc Statoil Research Centre, Sandsliveien 90, Sandsli, 5020 Bergen, Norway

a r t i c l e i n f o

Article history:Received 2 August 2011Received in revised form11 June 2012Accepted 26 June 2012Available online 6 July 2012

Keywords:Continental slopeOffshore West Africa3D seismicsTurbiditeMeanderingLevéePoint barLateral accretion packageChannel-bend mound

* Corresponding author. Present address: Statoil R90, Sandsli, 5020 Bergen, Norway. Tel.: þ47 94166972

E-mail addresses: mjan@statoil.com, mikejanocko@

0264-8172/$ e see front matter � 2012 Elsevier Ltd.http://dx.doi.org/10.1016/j.marpetgeo.2012.06.012

a b s t r a c t

The study combines interpretation of 3D seismic imagery of submarine sinuous channel belts in offshoreWest Africa with observations from a range of outcrop analogues. Five main architectural elements ofslope channel belts are recognized: lateral-accretion packages (LAPs), channel-bend mounds, levées,non-turbiditic mass-transport deposits (MTDs) and last-stage channel-fills. Channel belts differ in theirplanform, cross-section and the range of architectural elements involved. Four end-member types ofsinuous channel belts are distinguished, formed by meandering non-aggradational channels, levéedaggradational channels, erosional cut-and-fill channels and hybrid channels. Analysis indicates thatmeandering channels formwhen system is near its potential equilibrium profile. They evolve from nearlystraight to highly sinuous by increasing first the bend amplitude and then the conduit length. Levéedchannels are thought to evolve from incipient meandering conduits perturbed by aggradation anderosional channels to evolve from either levéed or meandering conduits, inheriting their sinuosity.Hybrid channels signify a failed or incomplete transformation. The channel belts occur isolated orstacked into multi-storey complexes, unconfined or formed within incised valleys. Unconfinedcomplexes, composed of levéed channel belts, are relatively uncommon. Valley-confined complexespredominate and are overlain by isolated channel belts, often confined by the valley external levées.

Valley-fill complexes are characterized by an upward fining and a general decrease in sandstone net/gross. The majority of slope valley-fills in the study area and other reported cases show a developmentfrom deep incision to a transient equilibrium state recorded by the deposition of coarse sediment lag ornon-aggradational channel belts, which are commonly overlain by MTDs emplaced when the valleyreached its maximum relief. The middle to upper part of valley-fill consists of levéed channel beltsrecording aggradation, with possible development of non-aggradational meandering channel belts in theuppermost part prior to the valley abandonment. Similar meandering channel belts may alsooccasionally occur in the middle part of valley-fill succession. It is suggested that the variation amongvalley-fills can be due to external factors, such as slope tectonics and salt movements, or to an internalforcing through the interplay of valley incision depth, base-level change, turbiditeesystem equilibriumprofile and slope general aggradation rate.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The last decade saw significant advances in the sedimentolog-ical understanding of deep-water sinuous channels and theirfeatures. Detailed studies of side-scan sonar and 3D seismic-reflection imagery have revealed a range of architecturalelements associated with sinuous channels, such as lateral-accretion packages (LAPs) (Abreu et al., 2003; Mayall et al., 2006;Kolla et al., 2007; Labourdette, 2007), nested mounds (Clark and

esearch Centre, Sandsliveien; fax: þ47 55996076.hotmail.com (M. Janocko).

All rights reserved.

Pickering, 1996; Peakall et al., 2000), outer-bank bars (Nakajimaet al., 2009), non-turbiditic mass-transport deposits (Deptucket al., 2003; Samuel et al., 2003; Heiniö and Davies, 2007;Armitage et al., 2009), levées (Clemenceau et al., 2000; Skeneet al., 2002; Babonneau et al., 2004; Hubbard et al., 2009),crevasse splays (Demyttenaere et al., 2000; Mayall and Stewart,2000; Posamentier and Kolla, 2003; Cross et al., 2009) and last-stage channel-fills (Kneller, 2003; Wynn et al., 2007). Most ofthese elements have been recognized in outcrops as sandy togravelly deposits (e.g., Morris and Normark, 2000; Lien et al., 2003;Dykstra and Kneller, 2009; Kane et al., 2009; Kane and Hodgson,2011) and are considered to be important components of hydro-carbon reservoirs (Prather, 2003; Mayall et al., 2006).

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e348

However, the previous studies have also indicated that elementsof one type vary as sedimentary deposits and that it is unlikely forall architectural elements to occur within a single channel belt (e.g.,Abreu et al., 2003; Kane et al., 2008; Amos et al., 2010; Janockoet al., 2013). Although some elements may be genetically linked,the development of one type of element may require flow condi-tions that virtually preclude formation another element type. Thisdepositional variability of channelized flows and the variability ofelements as sedimentary deposits may have a direct bearing on theobserved diversity of deep-water channels (Abreu et al., 2003;Kneller, 2003; Nakajima et al., 2009). Studies of architecturalelements in connection with the planform and cross-sectionalgeometry of channels may thus shed more light on the formativeprocesses of these highly diversified systems and help predict theirreservoir properties.

2. The aim of the present study

The present study documents the seismic characteristics ofdeep-water sinuous channels in an upper- to middle-slope settingin offshore West Africa and supplements these observations withwell-core data and a range of outcrop analogues. We revisit furtherthe taxonomic concept of channel classification (cf. Mayall andStewart, 2000; Morris and Normark, 2000; Pirmez et al., 2000;Kneller, 2003), with a special focus on intra- and extra-channelarchitectural elements and the temporal changes in channeldevelopment within deep-water slope valleys.

The 3D seismic dataset used in the study extends about 25 kmseawards, from the West African palaeo-shelf edge to the middlezone of continental slope, and covers an area of 80 � 55 km(4400 km2). The stratigraphic interval studied is of Miocene age.The dataset is a post-stack time-migrated volume with a binspacing of 12.5 � 12.5 m and a sampling interval of 4 ms. Seismicfrequency ranges from 20 to 60 Hz, with an average of 40 Hz cor-responding to a vertical resolution of ca. 10 m. The volume has beenprocessed to zero-phase and displayed in SEG normal polarity, suchthat the positive amplitude (black or dark-blue hue in the display)reflects higher acoustic impedance. An average seismic velocity of2000 m/s was used in the conversion of two-way travel time tometric depth for the purpose of calculating rock thicknesses inmetres.

More than 1600 m of core samples were recovered from 29wells in the study area. However, the samples and gamma logs fromonly five wells are utilized in this study, because the majority of thedrilling targets are in areas with poor seismic resolution, whereboth seismic interpretation and well-to-seismic ties are extremelydifficult. The problems with resolution are due to salt diapirism.

The quality of seismic data allows recognition of such strati-graphic features as valley-fills, palaeochannels, channel belts andtheir main architectural elements. The seismic recognition andinterpretation of architectural elements have been bolstered byoutcrop analogue studies from the Miocene Mt. Messenger Fm. ofNew Zealand, the Eocene Kırkgecit Fm. of Turkey, the Late Creta-ceous Rosario Fm. of Mexico and the Late Carboniferous Ross Fm. ofIreland. The purpose of using outcrop analogues was to get aninsight in the facies composition and depositional process of theelements from which no drilling samples were available. Althoughthe selected field examples are often smaller in scale than theelements identified in seismic imagery, they are considered to bevalid analogues in terms of geometry, facies assemblages andformative processes. Suffice it to note that elements such as levées,point bars, nested mounds, intra-channel mass-transport depositsand last-stage channel-fills occur in a range of settings and ona wide range of scales (Phillips, 1987; Timbrell, 1993; Elliott, 2000;Abreu et al., 2003; Arnott, 2007; Cronin et al., 2007; Euzen et al.,

2007; Wynn et al., 2007; Dykstra and Kneller, 2009; Amos et al.,2010; Kane and Hodgson, 2011; Janocko et al., 2013; Janocko andNemec, in press).

3. Terminology

Descriptive sedimentological terminology is after Harms et al.(1982) and Collinson and Thompson (1982). Submarine channel isdefined as a conduit formed by and conveying sediment-gravityflows. Channelized flows deposit coarse sediment both inside anddirectly outside the conduit, which itself may migrate, and theresulting sand-prone and possibly gravel-bearing sedimentarybody is referred to broadly as a channel belt (Bridge, 2003). Channelbelts with a laterally-inactive sinuous planform, formed by simpledowncutting and vertical aggradation are referred to as erosionalchannel belts (Fig. 1A); those showing significant lateral accretionand conduit sideways migration are referred to as meanderingchannel belts (Fig. 1B; cf. Nanson and Knighton, 1996); and thosewith seismically detectable levées and relatively stable sinuousplanform are referred to as levéed channel belts (Fig. 1C). Somechannel belts show major vertical aggradation combined withlateral accretion of sediment, which is called aggradational lateralaccretion (Fig. 1C, lower part). Multi-storey channel belts, stackedvertically upon one another with or without significant offset, arereferred to as channel-belt complexes (Fig. 1AeC).

The deepest, hydraulic axial zone of a channel is referred to asthe channel thalweg (Bridge, 2003). It does not correspond strictlyto the plan-view geometrical axis, or centreline, of the channel(Fig. 1D), which is more convenient to use in the analysis ofchannel-belt seismic maps. Accordingly, the sinuosity index ofa channel or its particular segment is defined as the ratio of thecentreline length to the corresponding straight-line distance(Bridge, 2003). Channels with a sinuosity index equal or greaterthan 1.1 are considered to be sinuous, non-straight. Othergeometrical parameters of channel planform used in the study are(Fig. 1D):

� channel width e considered to be the maximum local distancebetween the channel banks;

� channel depth e measured as the vertical relief from thechannel base in axial zone to the bank or levée crest;

� channel bend amplitude (or radius of curvature)e defined as themaximum departure of channel centreline from a straight-linepath through the centreline inflection points; and

� channel bend half-wavelength e the distance between centre-line inflection points measured along the channel centreline.

A submarine incised valley (Carlson et al., 1982; Prather, 2003) isan underwater slope conduit incomparably deeper than the systemlargest channels, cut in earlier deposits by excessively erosivesediment-gravity flows. In contrast to the more permanent deepsubmarine conduits, such as bedrock canyons, the incised valleysare cut and filled by the channelized turbiditic system, possiblyseveral times over during the time-span of its activity. Submarineincised valleys may not necessarily be related to sea-level changesand the fluvial incised valleys formed by forced regressions(Dalrymple et al., 1994), but they similarly result from major re-adjustments of the system morphometric profile.

Large-scale levées that flank an incised valley are referred to asexternal levées, whereas the smaller-scale levées flanking individualchannels are called internal levées (Kane and Hodgson, 2011).Channel belts formedwithin the valley confinement are consideredto be erosionally confined (Fig. 1E), whereas those constrainedlaterally by external levées are considered to be levée-confined(Fig. 1F). A submarine incised valley-fill commonly evolves from

Figure 1. Schematic diagrams illustrating basic terminology used in the present study. (A) Erosional channel belts. (B) Non-aggradational meandering channel belts. (C) Aggra-dational levéed channel belts. (D) Descriptive geometrical parameters of sinuous channel planform. (E) Erosionally confined channel-belt complex. (F) Erosionally to levée-confinedvalley-fill complex. (G) Unconfined complex of vertically offset-stacked levéed channel belts. (H) Valley-fill complex set. For definitions and further explanation, see text.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 9

erosionally confined to levée-confined (Fig. 1F). An aggradationalstack of levéed channel belts unrelated to a valley is referred to asan unconfined channel-belt complex (Fig. 1G).

A valley-fill complex may comprise a complex of multi-storeychannel belts as well as isolated channel belts (Fig. 1E) or becomposed mainly or entirely of mud. Mud-prone abandonedvalley-fills occur in the study area, but are not considered here. Twoor more valley-fill complexes stacked upon one another (Fig. 1H)are referred to as a valley-fill complex set (cf. Sprague et al., 2002).

4. Architectural elements of sinuous deep-water channels

An architectural element is a depositional body defined by itsgeometry, facies assemblage, scale, a particular formative process or

suite of processes, and its depositional setting (Miall, 1985). Inseismic interpretation of ancient deposits, the recognition of archi-tectural elements is generally based on their geometry, scale anddepositional setting, whereas facies composition and processes areinferred from other geological data (e.g., outcrop analogues, labo-ratory experiments, numerical modelling). The elements describedhere occur within sinuous channel belts and indicate sites of pref-erential sediment deposition by the channelized flows involved.

4.1. Lateral-accretion packages (LAPs)

Lateral-accretion packages (Abreu et al., 2003) appear in attri-bute maps and time slices as features similar to fluvial scroll bars(Fig. 2B) or as crescent-shaped high-amplitude reflection patches

Figure 2. Seismic attribute maps (AeD) and corresponding vertical sections showing the planform and cross-sectional geometry of lateral accretion packages (LAPs) in the studyarea. (A) Map of LAPs manifested as high-amplitude reflection threads, showing channel-loop rotation and expansion combined with downstream translation. The LAPs havea thickness at the margin of seismic resolution and appear as a single high-amplitude reflection (see cross-section AeA0). (B) Map of LAPs resembling fluvial scroll bars, with thecrescent-shaped patches of high-amplitude reflections showing bend expansion followed by expansion with downstream translation. In vertical section, the LAPs show up asshingled reflections dipping towards the last-stage channel thalweg (see cross-section BeB0). (C) Map of a purely downstream-translated LAP, with high-amplitude reflectionthreads and with a pattern of shingled reflections in vertical section (see cross-section CeC0). (D) Map of LAPs with a scroll-like pattern showing bend expansion combined withdownstream and upstream translation; cross-section DeD0 shows shingled reflections. Note that the LAP bases and tops are generally flat and that the LAP planform developmentmay vary from one bend to another.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3410

(Fig. 2B). They may locally appear also as closely spaced, high-amplitude sinuous threads (Fig. 3A, C). In seismic profiles, LAPsare typified by discontinuous, offlapping shingled reflectionsdipping at 5e10� towards the last-stage channel thalweg (Fig. 2,sections BeB0, CeC0 and DeD0). In places where the LAP thickness isbelow the seismic tuning thickness (i.e., seismic wavelength), theinclined reflections are unrecognizable and the package appears asa single, continuous high-amplitude reflection (Fig. 2, sectionAeA0). The LAPs in such a case can only be inferred from attributemaps. The bases and tops of LAPs are generally flat and horizontal.The areal extent of LAPs is in the range of 40e480 m2 and theirthicknesses are up to 30 m.

The sedimentary facies of LAPs are inferred from drilling cores.Four separate cores from three different channel belts at the base ofvalley-fill complexes have been analysed (Figs. 3 and 4). On the basisof seismic sections and attribute maps, each of the cored intervals isconsidered to be a complete (cores A1, A2 and B1) or a partial (coreC1) single-storey channel-belt LAP. The two cores from channel beltA both show an overall fining-upward trend and similar facies, asthey consist of massive to planar parallel-stratified, normally-

graded sandstone beds with scattered mudclasts (Fig. 4, cores A1and A2). Sandstones are mainly coarse- to fine-grained, overallslightly coarser in core A1. Mudclasts are angular to subrounded and0.5e20 cm in length. They occur either at the bed base, where theyoften show imbrication, or in the bed middle part where they aremore scattered and lack preferential orientation. Core A2 shows alsonormally-graded beds of sand-rich mudclast conglomerate.

The core from channel belt B shows beds of sand-supported,extra- and intra-formational conglomerates in the lower part,whereas the upper part is dominated by planar parallel-stratified toripple cross-laminated sandstone beds (Fig. 4, core B1). Extra-formational lithic clasts are up to 5 cm in size, but the maximumsize of mudclasts reaches 22 cm. The conglomerate beds showplanar parallel stratification with clast imbrication and typicallypass upwards into a massive or crudely stratified sandstone. Thecore from channel belt C shows only the lower part of the LAP,which differs from the others in that it consists of thick, amal-gamated, massive to crudely stratified sandstone beds (Fig. 4, coreC1). The bed bases are erosional, commonly strewnwith imbricatedmudclasts up to 8 cm in length. These thick beds are intercalated

Figure 3. Seismic maps (upper row), their interpretation (middle raw) and the corresponding vertical sections (lower row) of meandering channel belts at the base of valley-fillcomplexes in the study area. The channel belts are interpreted to be single-storey non-aggradational meander belts that evolved by bend cut-off and lateral shifting. The seismicsections include gamma-ray (GR) well logs of the meander belts, with the corresponding core logs and facies details shown in Figure 4.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 11

with minor thin beds of planar parallel-stratified to ripple cross-laminated sandstone capped with siltstone.

The LAPs are interpreted to represent point bars formed by thelateral migration of an open channel. Similar features elsewherehave been attributed by some authors to the lateral stacking ofconsecutive cut-and-fill channels (Kolla et al., 2001, 2007; Mayallet al., 2006; Arnot et al., 2007; O’Byrne et al., 2007; Labourdetteand Bez, 2010). However, this interpretation is here disfavouredby the evidence of a systematic channel-bend expansion indicatedby scroll-bar patterns in attribute maps. It seems unlikely that thedevelopment of a long series of consecutive and unrelated scour-and-fill channels would result by sheer chance in such a system-atic stacking pattern of nearly identical erosional relics.

On the basis of their planformdevelopment, the point bars in thestudied channel belts can be classified as expansional, downstreamor upstream translational, rotational or representing a combinationof these three main modes of evolution (Fig. 2; terminology afterBrice, 1974). The development of point bars appears to vary fromone channel-belt segment to another and lacks any systematicspatial trend. This variability suggests that the planform evolutionof point barsmaydepend strongly on the local seafloor gradient andsubstrate cohesiveness, which would in turn control the planformof channel bends and curvature of their transitions.

The lower parts of point-bar LAPs are dominated by stratifiedsand-supported conglomerates and massive to crudely stratifiedsandstones. The stratified conglomerates and sandstones representthe turbidite division R1 of Lowe (1982) and division b of Bouma

(1962), respectively, and are interpreted to be tractional upperflow-regime deposits of low-density turbidity currents (sensuLowe, 1982). Although massive sandstones occur mainly in thelower part of LAPs, they can be found also in the middle and upperpart. They represent the turbidite division S3 of Lowe (1982) andindicate sand deposition by rapid dumping from a decelerated,high-density turbidity current (see also Lowe, 1988). The deceler-ation and abrupt basal densification of flow can be attributed to itsoblique climbing on the point bar, with flowline expansion towardsthe inner bank and frictional loss of energy (Janocko et al., 2013).The planar parallel-stratified to ripple cross-laminated sandstonebeds are classical Bouma-type turbidites Tbc, deposited by low-density turbidity currents and occurring mainly in the upper partof point-bar LAPs.

Point-bar deposits similar to those in the West African offshorechannel belts can be found elsewhere exposed on land. An examplewith similar geometry and facies assemblage is afforded by theWaikiekie South Beach cliff section of the Mount MessengerFormation in New Zealand’s North Island (Fig. 5). Although the LAPhere is smaller than the seismically recognized cases, it may stillserve as a valid analogue for larger point bars, because the process ofchannel meandering as such is not scale-limited (Janocko andNemec, in press).

The channel belt occurs at the base of a valley-fill complex(Fig. 5A, B), which is located in the lower part of a large valley-fillcomplex set (Arnot et al., 2007). The estimated channel-belt widthisw170m, with about two-thirds of it occupied by the LAP. The LAP

Figure 4. Sedimentological well-core logs and photographic facies details of the meander belts shown in Figure 3. Well logs A1eB1 show whole point-bar LAP successions and welllog C1 shows the lower part of a point-bar LAP. Log grain-size scale: m ¼ mud; vfs to vcs ¼ very fine to very coarse sand; gr ¼ granule gravel; pb ¼ pebble gravel. GR is the wellgamma-ray log.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3412

is 6 m thick, composed of beds with a mean inclination of 8�. Thedeposits are laterally-accreted couplets of massive mudclastconglomerate and planar parallel-stratified sandstone (Fig. 5C, D),with the last-stage aggradational channel-fill composed of amal-gamated, massive to crudely stratified sandstone beds occasionallybearing basal mudclast lags (Fig. 5A, B). The mudclast conglomer-ates have sandy matrix, consist of angular to subrounded clasts upto 30 cm in length and have a clast- to matrix-supported texture.Most of these conglomeratic divisions are normally graded, butsome show a coarse-tail inverse grading. They are thickest in theirdown-dip parts and tend to thin up-dip in the LAP cross-section,

where they also become finer grained and their bases lesserosional. The sandstone divisions, in contrast, are lenticular in theLAP section and their down-dip parts are generally thinner, trun-cated by the overlying bed (Fig. 5D). The top and base surfaces of theLAP are planar, originally horizontal, although the lateral migrationof channel thalweg involved uneven scouring and resulted in localmorphological irregularities of the channel-belt base (Fig. 5A, B).

The successive conglomerate-sandstone couplets in the LAP arethought to be products of density-layered bipartite flows (cf.Postma et al., 1988). Mudclasts were derived from erosion of anunderlying, rugged slump body of semi-consolidated slope mud,

Figure 5. Outcrop section of a meandering channel belt in the Mount Messenger Fm. at the Waikiekie South Beach, New Zealand’s Northern Island. (A) Cliff photomosaic and (B)overlay drawing of the point-bar LAP and associated last-stage channel-fill. The close-up details show: (C) a mudclast conglomerate-sandstone couplet in the LAP and (D) a down-dip pinchout of the sandstone division. Note the hydroplastic deformation caused by sediment loading.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 13

breached by the channel incision and lateral migration (see Janockoand Nemec, in press). The supply of mudclasts was local, limited tothe cut-banks of certain bends. An erosive turbidity current,abruptly overcharged with mudclasts at the bend, would spawna co-genetic basal debris flow and spread it over one or two bendsdownstream. The abundance of angular and subangular clastssupports the notion of a short transport. The loss of energy onerosion would decelerate the flow, whereas the entrainment ofcohesive material caused basal densification of the flow due torapid suspension fall-out (Lowe, 1988), leading to a cohesivefreezing of the bedload layer. The mudclast conglomerates are thusthought to represent co-genetic debris flows spawned and draggedbriefly along by the successive high-density turbidity currents.Weak inverse grading indicates that the size of mudclasts tended tobe diminished by the stronger frictional shear near the flow base.The overlying sandstone division Tab of each couplet was depositedby flow that rid itself of the excess basal load and kept dumpingsand directly from turbulent suspension before reversing todeposition from upper-stage plane-bed tractional transport (Harmset al., 1982; Lowe, 1982).

4.2. Channel-bend mounds

The sinuous channels in seismic-volume attribute mapscommonly show longitudinal patches of high-amplitude reflectionsin the apical zone of channel bends (Fig. 6). These features areassociated mainly with relatively sharp bends of high-sinuositychannels, levéed or non-levéed, and occur also in the last-stagechannels of some meander belts. Their occurrence seems to beindependent of the channel width/depth ratio. In vertical seismicsections, these features appear as high-amplitude horizontal

reflections at the base of channel belt, but may be indiscernible iftoo thin relative to seismic resolution, though visible in attributemap. Their areal extent is in the range of 5e60 m2 and thicknessesup to 30 m. No drilling cores of these deposits are available, buttheir laterally continuous high-amplitude seismic signature indi-cates coarse-grained deposits with little or no facies heterogeneity.

The high-amplitude reflection patches at channel thalwegbends are thought to represent deposits similar to ‘nested mounds’previously documented by Phillips (1987), Timbrell (1993) andClark and Pickering (1996). Nested mounds generally consist ofcoarse-grained sediment and have often been interpreted to resultfrom flow stripping at channel bends (Piper and Normark, 1983).However, the process of flow stripping has so far failed to producemound deposits in laboratory channels (see Peakall et al., 2000)which is why some authors have suggested alternative mecha-nisms, such as flow retardation in front of a collapsed channelmargin (Peakall et al., 2000, 2007;Wynn et al., 2007) and erosion ofouter-bank bars (Nakajima et al., 2009). The evidence from thepresent study indicates that the high-amplitude patches occurmainly at sharp channel bends, irrespectively of the channel aspectratio and presence of levées, which suggests that they representcoarse-grained deposits formed by an abrupt local deceleration offlow and are not necessarily related to flow overspill (cf. Clark andPickering, 1996).

An outcrop analogue of such deposits is afforded by the SanFernando canyon section of the Rosario Fm. in Baja California,Mexico (Fig. 7; Janocko and Nemec, in press). This example isconsiderably smaller than the analogous features observed inseismic imagery, but their formative process may be similar as theyall display comparable geometry and a similar scaling with channeldepth. Numerical simulations have also shown that the formation

Figure 6. Seismic maps of levéed aggradational channel belt (top left), erosional channel belt (top middle) and non-aggradational meandering channel belt (top right) in the studyarea. The RMS volumes shown by the maps are indicated in the corresponding vertical sections below. The high-amplitude reflections (HARs) at channel bends are thought torepresent channel-bend mounds.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3414

of deposits resembling channel-bend mounds is not scale-limitedand occurs in channels ranging from laboratory scale to anynatural scale (Janocko et al., 2013).

The outcrop shows a meandering channel belt in the lower partof a submarine valley-fill complex. The belt LAP consists ofconglomerate-sandstone couplets inclined at w6� towards thelast-stage channel and downlapping an erosional, originally hori-zontal base of the channel belt. The last-stage channel-fill consistsof conglomerate-sandstone couplets that show aggradationallateral accretion, with the conglomerate divisions thinning andsandstone divisions thickening in the updip direction. In thelowermost couplet, the parallel-stratified conglomeratic divisionforms a mound with an irregular convex-upward top and with thestrata changing laterally their attitude from paralleling the LAPbedding at the inner bank to gently rising against the outer bank.The conglomerate bed truncates the underlying beds, whichsuggests that the depositing flow had initially broadened theconduit by eroding its both banks. The overlying sandstone divi-sion has a sub-horizontal top and an uneven thickness compen-sating for the morphological irregularity of the conglomerate top.The conglomerate clast imbrication indicates sediment transportobliquely towards the outer bank (Janocko and Nemec, in press),which suggests that the flow helicoid at the channel bend wasrising against the outer bank. This evidence implies that the

rotation pattern of flows depositing channel-bend mounds is moresimilar to the rotation of currents producing outer-bank bars (seeNakajima et al., 2009) than to the rotation of currents formingpoint bars.

4.3. Mass-transport deposits (MTDs)

Non-turbiditic mass-transport deposits, attributed to suchprocesses as slides, slumps and debris flows, occur at various scalesin submarine channels and valleys. Slide blocks from valley wallsare among the largest features, with an areal extent reaching1500 m2 and thicknesses up to 120 m. In seismic attribute maps,they are recognizable as low-amplitude, elongate to crescent-shaped features associated with scallop-shaped scars at channelor valley margins (Fig. 8A). In seismic cross-sections, slide blocksshow rotational bases, stepped tops and undisturbed, parallel low-amplitude internal reflections (Fig. 8B, E).

Slump deposits appear in attributemaps as circular or crescentichigh-amplitude patches (Fig. 8C). In cross-sections, they typicallyshow curved rotational bases and internal transparent pattern oflow-amplitude discontinuous reflections (Fig. 8D). They are oftenlaterally more extensive than slide blocks, with areas of up to2000 m2 and thicknesses reaching 50 m. Deposits attributed tolarge debris flows may or may not be associated with slump scars

Figure 7. (A) Outcrop section of a meandering channel belt in the Rosario Fm., Mexico, showing a conglomeratic channel-bend mound (indicated in red) at the base of the last-stagechannel-fill. (B) Close-up view of the mound, showing how the plane-parallel stratification of the mound changes laterally its attitude from paralleling the LAP bedding on the rightto sloping gently away from the outer bank on the left. Note the erosional base of the overlying conglomerate-sandstone couplet truncating the LAP beds to the right. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 15

and are typically spread across the entire width of the channel orvalley (Fig. 8F). In both plan view and cross-section, they appear astransparent to chaotic seismic facies with a sharp, often erosionalbase and irregular top. In the study area, the MTDs are generallyrecognized in valleys, rather than in channels, which may be due tothe limited seismic resolution or to a natural scarcity of channel-bank collapses, as compared to the gravitational instability ofsteep valley walls.

MTDs have been reported from deep-water channels and valleys(e.g., Droz and Bellaiche, 1985; O’Connell et al., 1995; Morris andNormark, 2000; Deptuck et al., 2003; Samuel et al., 2003; Heiniöand Davies, 2007; Armitage et al., 2009), and are considered to bean important element affecting the evolution of submarineconduits and their hydrocarbon storage potential (Prather, 2003). Inthe present case, the local collapses probably played a major role inthe development of deep-water incised valleys from erosionalchannels. The emplacement of MTD may cause flow retardationand enhance deposition in the thalweg zone (Peakall et al., 2000;Nakajima et al., 2009) or may cause flow avulsions (Fig. 8C),whereas the uneven top relief of slump and slide bodies may pondturbidity currents or entrap channels (Fig. 8B; Faulkenberry, 2004).

The internal character of MTDs varies, depending upon thecollapsing sediment facies of the channel or valley wall and theintensity of shear deformation involved. As end-members, slideblocks are relatively coherent and internally intact, whereas debris-flow bodies are strongly homogenized by pervasive shear. TheMTDs derived from channel-bank levée collapses tend to be sand-prone (Kane and Hodgson, 2011), but for this reason also have a lowpreservation potential in an active channel. The more cohesive,mud-prone MTDs will provide abundant mudclasts and affect therheological properties of subsequent currents, while possiblyaffecting also the physiographic development of the channel(Hodgson, 2009).

4.4. Levées

Levée deposits are the largest and most extensive sand-pronearchitectural element of sinuous channel belts. They are thusimportant as an exploration target and element of reservoir char-acterization. They occur at various scales, and their morphologyand facies help to shed light on the character of flows conveyed bythe channel. Levées occurring in isolated unconfined channel belts,in unconfined and confined channel-belt complexes and at valleymargins (Fig. 9) are described separately below.

In isolated unconfined channel belts, levées are typically two ormore reflections thick and characterized by a gull wing-shapedcross-section (Fig. 9A). The reflections downlap the underlyingdeposits and range from low-amplitude discontinuous to high-amplitude continuous. In all cases, the levée reflections are ofhigher amplitude than the reflections of the associated channel-fill.The height of the outer-bank levées (40e60 m) generally exceedsthe height of the inner-bank levées (30e50 m) along the entirestudied course of a channel. The top surface of these levées istypically smooth and their thickness decreases in an exponential orhyperbolic (power-law) manner with distance from the channel(Fig. 9E). Some of the large outer-bank levées have an undulating,wavy top downstream of channel bends (Fig. 9A), which is attrib-uted to the formation of overbank sediment waves. More irregular,jagged tops may be due to synsedimentary gravitational faultsdipping away from the channel. The continuity of reflections fromthe levée and channel-fill varies considerably. In most cases, thereflections in the upper part of levée can be traced to the last-stagechannel-fill, but those in the lower part tend to terminate at thechannel bank. In seismic attribute maps, the levées in unconfinedsolitary channel belts appear as areas of a high-amplitude signaldeclining away from the channel. The levée width tends to beinversely proportional to the channel-belt gradient.

Figure 8. Seismic evidence of mass-transport deposits (MTDs) in the valley-fill complexes in the study area. The seismic maps (A, C) and corresponding vertical sections (B, D) showevidence of slides and slumps. Slide, slump and possible debris-flow deposits are shown also in the two other sections (E, F).

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3416

In unconfined or weakly-confined channel-belt complexes,levées are typically stacked in a compensational manner and showcomplex geometries (Fig. 9B). In simple cases, the levées showcontinuous, onlapping high-amplitude reflections and an expo-nential decrease of thickness away from the channel (Fig. 9F). Inmore intricate cases, where younger levées extend over an aban-doned levéed channel belt, the overlapping levée shows an irreg-ular, concave- to convex-upward top and continuous, similarlyundulated high-amplitude reflections (Fig. 9B). The levée reliefvaries from less than 20 m to more than 100 m. There is rarelya continuity of reflections from the levée to channel-fill, exceptwhere an abandoned channel segment was buried by levées ofadjacent active channel. In seismic maps, the levées are asym-metrical and increase in areal extent with a decrease of channel-belt gradient.

In confined channel-belt complexes, levéed channel belts tendto show systematic aggradation, rather than compensational

stacking (Fig. 9C). The confined mode of channel-belt developmenttypically results in simpler levée geometry than observed inunconfined complexes. Two types of levées, internal and external,are associated with confined channel-belt complexes (Kane andHodgson, 2011). The external levées in cross-sections are typifiedby downlapping, high-amplitude continuous reflections and locallyshow small cross-cutting crevasses (Fig. 9C). The continuity ofexternal levée reflections to the valley-fill is poor in erosionally-confined channel-belt complexes, but good in levée-confinedcomplexes (Fig. 9C). The thickness of proximal levée varies from 50to 120 m, and the levée lateral tapering trend is best approximatedby an exponential function (Fig. 9G). In attribute maps, the externallevées are recognizably asymmetrical, with scoop-shaped inden-tations (collapse scars) at the inner margin. The internal levées, inturn, are gull wing-shaped and 20e50 m thick, characterized bycontinuous to discontinuous low-amplitude reflections (Fig. 9C)downlapping the substrate and often also onlapping the

Figure 9. Seismic evidence of levées in the study area. (A) Example seismic sections showing levées in unconfined isolated channel belts, where the levées downstream of sharp bends locally have wavy tops attributed to the occurrenceof sediment waves. (B) Example seismic sections showing levées in unconfined multi-storey channel-belts stacked in a compensational manner. (C, D) Example seismic sections showing levées in channel belts confined by valley reliefor external levées, with well-core photographs of internal levée facies. The diagrams to the rights show the lateral thinning trend of levées in isolated (E) and multi-storey channel belts (F), and of the external (G) and internal levées (H)in submarine valleys in the study area.

M.Janocko

etal./

Marine

andPetroleum

Geology

41(2013)

7e34

17

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3418

confinement margin. Their lateral thinning trend is best describedby an exponential function (Fig. 9H). In attribute maps, the internallevées show up as uniform or patchy low-amplitude zones paral-leling the sinuous planform of the associated channel. A 50-m coresample from the internal levées in one of the channel complexes(Fig. 9D) shows a fining- and thinning-upward trend and a sand-stone net/gross of 19%. The deposits are thinly bedded (1e30 cm),planar parallel-stratified to ripple cross-laminated sandstonesinterlayered with mudstones.

The deposition of levées in deep-water channel belts is attrib-uted to the overspill of turbidity currents conveyed by the channel.A turbidity current spills out of the channel because it is eithervolumetrically too large for the channel capacity or in a hydraulicdisequilibrium with the channel geometry. The channel-formingcurrents are thought to be considerably thicker than the actualchannel depth and hence inevitably spilling out, but also an inter-vening smaller current may run up on the outer bank at channelbends and spill over until the flow volume critically declines andinertia drive dissipates (Straub et al., 2008). Currents may spill outexcessively in response to the local plugging of channel by MTDsand will also increasingly spill over as the last-stage channel beginsto be filled with sediment prior to abandonment.

The outer-bank levée in most of the studied channel belts ishigher than the inner-bank levée, which suggests that the flowsuperelevation and outer-bank run-up at channel bends playeda significant role (Straub et al., 2008, 2011; Amos et al., 2010). Theexponential outward-thinning trend of levées is consistent withsimilar observations from levéed channels in many other subma-rine systems (Skene et al., 2002; Skene and Piper, 2005). Kane et al.(2010) have attributed the exponential trend to the spill-out ofsustained (long-duration), quasi-steady high-competence flows,which is in agreement with the hydraulic conditions for theformation of overbank sediment waves found on levées (Fig. 9A;Nakajima and Satoh, 2001; Cartigny et al., 2010).

The poor continuity of levée reflections to channel-fill in iso-lated channel belts and unconfined channel-levée complexessuggests that the denser, channel-confined parts of flows werecommonly bypassing the bends and eroding the outer bank, eventhough these sinuous channels generally show little or no lateralmigration. The notion of erosional bypass is supported by thedifferential aggradation of the channel floor and levées.

In contrast, the continuity of levée reflections to channel-fillappears to be good in the aggradational channel belts of erosion-ally- or levée-confined channel-belt complexes. The channels insuch settings were apparently conveying net-depositional flows,with the channel-floor aggradation keeping pace with the levéebuild-up.

The external levées are thought to formwhen the overbank flowfrom a channel inside the valley spills out beyond the valleymargins (Kane and Hodgson, 2011). These levées are unlikely to beconnected with the parental channel-fill because of the negligiblecapacity of valley walls to store sediment and the valley-margintendency for mass wasting.

4.5. Last-stage channel-fills

The last-stage channel-fills, as an element heralding channel-belt abandonment, are generally deposited by flows differingfrom those which formed and shaped the channel belt (Clark andPickering, 1996; Kneller, 2003; Wynn et al., 2007) and hence aregenetically unrelated to the belt’s other, earlier-formed architec-tural elements. However, the channel-fills themselves are animportant element, because they vary from sand- to mud-prone,may constitute a major part of channel belt and determine theconnectivity of the channel belt’s other architectural elements.

In seismic cross-sections, the last-stage channel-fills show high-to low-amplitude continuous horizontal reflections (Fig. 10A)indicative of vertical accretion. If the channel-fill is dominated bysand, its seismic signature tends to be a convex-upward “hat”(Fig. 10B, C) attributed to the differential compaction of the axialsandy fill and adjacent muddier deposits (Posamentier, 2003). Thehat-form signature typifies channel-fills that are at or below theseismic tuning thickness, but may also occur at larger channelscales and/or higher seismic resolution (e.g., see Fig. 10B). In attri-bute maps, the last-stage channel-fills show up as either low- tohigh-amplitude continuous threads or high-amplitude discontin-uous patches.

Last-stage channel-fills have been documented and catego-rized by several authors (e.g., May et al., 1983; Mutti andNormark, 1987; Shanmugam and Moiola, 1988; Cook et al.,1994). Factors that determine channel-fill facies include theparameters of last-stage flows, the pre-existing channel topog-raphy and the location of the channel in the submarine envi-ronment and in respect to coeval active channels. The key flowparameters are the ratio of flow thickness to channel depth andthe duration and sediment load of the flows. The channel depthand sinuosity will determine the thickness and spatial distribu-tion of the infill deposits, with possible slides and slumps relatedto the channel bank steepness. Some channels or their cut-offsegments are abandoned abruptly and then gradually filled byspill-out flows from adjacent active channels, slope-derivedminor “wild” flows and/or hemipelagic sedimentation. Such“passive” channel-fills may thus be highly heterogeneous orvirtually mud-dominated and their facies are difficult to predict,as they will bear virtually no genetic relationship to the wholepreceding development of the channel belt.

The sedimentary facies and internal architecture of last-stagechannel-fills are generally difficult to recognize from seismicimagery due to its insufficient resolution, but can possibly beinferred on the basis of an understanding of particular channel-beltdevelopment combined with outcrop- analogue studies. Fouroutcrop examples (Fig. 10DeJ) have been selected to illustratevariation in the last-stage fills of meandering and erosional sinuouschannels, with sandstone net/gross ranging from 0 to 100%. Theexamples show four different styles of channel-fill architecture:vertical aggradation, aggradational lateral accretion, single-bedplugging and a polygenic infill.

Last-stage channels filled by simple vertical aggradation(Fig. 10D, E) are found in both erosional and meandering channelbelts and typically consist of margin-onlapping tabular beds witha thinning upward trend. Internal scours abound in larger channels,with the individual beds commonly thinning or pinching outtowards the channel margins. The channel-fill may be sandy andpossibly gravel-bearing or be heterolithic, composed of sand-mudcouplets. The deposits are thus either products of bypassing “over-sized” non-incising flows or products of waning flows that wereconsiderably “undersized” with respect to the channel hydraulicgeometry and fully dissipatedwithin the channel. The demise of thechannel occurs because none of these flow varieties can possiblykeep the channel active, while depositing sediment in it.

Last-stage channels filled by aggradational lateral accretion aretypical of meandering channel belts (Figs. 5B and 10F, G; see Wynnet al., 2007, case 3 in Fig. 18; Dykstra and Kneller, 2009, Fig. 12).These channel-fills may range from fully sandy and possibly grav-elly tomud-rich heterolithic, but typically consist of planar parallel-stratified to ripple cross-laminated sandstone beds, some witha massive division and dewatering structures. The lateral accretionsuggests that the depositional stacking pattern of turbidites wassimilar as during the formation and growth of the channel pointbars, but the flow magnitude had apparently decreased to render

Figure 10. Seismic evidence and outcrop analogues of last-stage channel-fills. (A) Aggradational fill of an erosional channel with levée, seen as moderate-amplitude horizontalreflections. (B) Sand-prone fill of an erosional channel with complex internal geometry and a convex-up “hat” top due to differential compaction. (C) Sand-prone fills of levéedchannels at seismic tuning thickness, showing convex-up “hat” tops. (D) Aggradational, thinly-bedded heterolithic fill of a meandering channel in the Ross Fm., Rehy Cliffs, Ireland.(E) Aggradational fill of a small erosional channel, comprising amalgamated planar parallel-stratified sandstone beds, in the Messenger Fm., Waikiekie South Beach cliff, NewZealand. (F) Laterally-aggraded fill of a meandering channel, composed of parallel-stratified conglomerate-sandstone couplets, in the Rosario Fm., Pelican Point, Mexico. (G)Laterally-aggraded fill of a meandering channel, comprising massive and parallel-stratified sandstone beds with occasional mudclast lags, in the Mt. Messenger Fm., WaikiekieSouth Beach, New Zealand’s North Island. (H) Single-bed plug of a meandering channel, composed of sand-rich mudclast conglomerate, in the Rosario Fm., San Fernando Canyon,Mexico. (I) Single-unit plug of a small erosional channel, composed of laminated mudstone, in the Ross Fm., Kilbaha Cliffs, Ireland. (J) Composite channel-fill with a basal package ofvertically-aggraded conglomerate beds thinning against outer bank, deeply re-scoured and overlain by bank-derived fine-grained slump deposit, whose uneven relief was smoothedand buried by a new aggradational package of conglomerate beds; example from the Rosario Fm., San Fernando Canyon, Mexico.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 19

the flows fully depositional, incapable of eroding the outer bankand maintaining channel migration.

Last-stage channels plugged by a single deposit are typicallyshallow conduits finalizing the development of some erosional andsome meandering channel belts. The channel-fill deposit may bea normally-graded, massive and/or stratified sandstone ormudclastconglomerate, with its coarsest-grained part at the channel thal-weg, or may be mudstone (Fig. 10H, I). The coarse-grained deposit,whether a turbidite or a deposit of debris flow spawned by turbiditycurrent, is attributed to a flow that was grossly “oversized” withrespect to the channel capacity. The flow would likely be erosive in

the channel upper reaches, as indicated by the common occurrenceof mudclasts (Fig. 10H), and the erosional bulking of sedimentwould then render it highly depositional and essentially non-erosive upon its arrival in the lower reaches. Muddy channel-plugs are rare in the middle to lower parts of submarine slopes,but may be more common in the upper parts. Their origin isattributed to the accumulation of hemipelagic mud is an aban-doned channel or to an accidental emplacement of a local slope-derived mudflow.

Last-stage channels filled with polygenic deposits are some ofthe relatively deep conduits with low aspect ratios that typify

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3420

erosional channel belts. Such channel-fills show a complex archi-tecture and are often highly heterogeneous (Fig. 10J). Their basalsandy turbidites are commonly deformed by the emplacement ofa bank-derived fine-grained slump deposit (MTD), possiblymultiple, which is covered by mud-capped turbidites with unevenbases, draping and smoothing the irregular relief of the underlying

Figure 11. Seismic RMS attribute maps showing planform characteristics of sinuous channaggradational channel belts. (C) Erosional cut-and-fill channel belts. The corresponding ver

slump deposit. The infilling of the channel thus apparentlycommenced with erosive flows, undercutting the outer bank andcausing its collapses, which were followed by smaller, net-depositional flows waning within the channel. This kind ofchannel-fill may be limited to single bends and its occurrences arethus difficult to predict.

el belts in the study area. (A) Meandering non-aggradational channel belts. (B) Levéedtical sections AeA0 to KeK0 are shown in Figure 14.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 21

5. Channel-belt types

On the basis of their planform, cross-sectional geometry andrange of architectural elements, the deep-water sinuous channelbelts in the study area have been classified into four end-membercategories (Fig. 1AeC): (a) meandering channel belts, formed bylaterally migrating non-aggradational sinuous conduits andgenerally lacking levées; (b) levéed channel belts, formed byaggradational sinuous conduits showing little or no lateral migra-tion; (c) erosional channel belts, formed by the cut-and-fill ofa sinuous conduit with no significant lateral migration and onlyminor levées; and (d) hybrid channel belts, combining features ofthe other three categories.

5.1. Meandering channel belts

Meandering channel belts in seismic maps are characterized bya high-sinuosity (1.9e2.8) conduit, regular and smoothly-curvedmeander bends and evidence of bend cut-offs (Figs. 11A and 12).Characteristic feature are LAPs formed on the inner side of channelbends. The last-stage channels show up as continuous, high- to

Figure 12. Statistical comparison of the planform geometrical characteristics of the threeparameter specified on diagram vertical axis; the horizontal axis accommodates individua moderate width/depth ratio, low sinuosity and low bend amplitude and length; the meanhigh bend amplitude and length; and the erosional channels are characterized by intermechannels are formed by entrenchment of either a meandering or a levéed primary conduit

moderate-amplitude threads with high-amplitude patches at thebends representing channel-bend mounds. Levées and intra-channel MTDs are apparently lacking.

In seismic cross-sections, these channel belts are typified byLAPs that appear as shingled reflections dipping towards the last-stage channel thalweg (Fig. 13A, section AeA0). LAPs are notrecognizable in channel-belts thinner than the seismic tuningthickness, where they instead appear as a single, laterally extensivereflection of anomalously high amplitude (Fig. 13A, section BeB0).The last-stage channels have a width/depth ratio in the range of10e20 (Fig.12), higher than in the other channel-belt types, and thechannel-fill shows low-amplitude reflections indicating verticalaggradation or aggradational lateral accretion. Channel widthsreach 325 m and depths of 25 m. The inclination of channel outerand inner banks at the bend apices is in the range of 8e14� and5e10�, respectively. Core samples from meander belts showa fining-upward succession composed of amalgamated, crudelystratified, mudclast-bearing sandstone beds in the lower part andmainly ripple cross-laminated sandstone beds in the upper part(Fig. 4). Extraformational conglomerates and/or intraformationalmudclast conglomerates may occur in the basal part.

main types of sinuous channels in the study area. The mean values pertain to theal datasets; n ¼ number of data. Note that the levéed channels are characterized bydering channels are characterized by a high width/depth ratio, high sinuosity and alsodiate values of all these parameters. This evidence supports the notion that erosional.

Figure 13. Vertical seismic sections of meandering (A), levéed (B) and erosional channel belts (C) in the study area. The location of cross-section lines AeA0 to KeK0 is shown inFigure 11. For descriptive and interpretive comments, see text.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3422

The meandering channel belts occur typically at the base oflarge incised valley-fills, where they are directly overlain byaggradational levéed channel belts.

5.2. Levéed channel belts

Levéed channel belts are characterized by conduits with low tomoderate sinuosities (1.1e1.4) and irregular, occasionally sharpbends, and by laterally extensive levées appearing in maps asuniform to patchy high-amplitude zones on both sides of thechannel (Figs. 11B and 12). The channel-fill shows lower amplitudesthan the overbank deposits, although bright-amplitude patchesoccur, particularly at the apices of channel bends. These patchescorrespond to channel-bend mounds. LAPs are lacking in isolatedchannel belts, but occur as aggradational packages in confinedchannel-belt complexes. No channel bend cut-offs and channel-margin collapse scars or MTDs have been recognized.

In seismic cross-sections, the levéed channel belts show a char-acteristic gull-wing shape with high-amplitude signature of levéesand low-amplitude signature of aggradational channel-fill(Fig. 13B). The levée reflections lack continuity with those of thechannel-fill, except for the last-stage deposits. The base of thechannel belt is generally flat, but may be slightly uneven due tovariable depth of thalweg scour. The channel width/depth ratio is inthe range of 6e10 (Fig.12), with the channel widths reaching 900mand depths 100 m. Channel banks are mainly symmetrical, inclinedat 11e18�.

The levéed sinuous channel belts predominate in the study area,occurring as isolated features (Fig. 9A) or as components ofevolving channel-belt complexes (Fig. 9B, C).

5.3. Erosional channel belts

In seismic cross-sections, the erosional sinuous channel beltsare V- or U-shaped features with relatively high-amplitude reflec-tions, inset in deposits with low-amplitude horizontal reflections(Fig. 13C). The channel-belt thicknesses are up to 300 m, generallyexceeding those of the other channel-belt varieties. Channel width/depth ratio is in the range of 5e7 (Fig. 12). Channel banks aremainly symmetrical, inclined at 15e18�, and the outer bank tendsto be steeper in higher-sinuosity channels. The base of channel belthas a smooth V- or U-shape, but shows a stepped profile in beltswith significant phases of erosional rejuvenation, indicated byintra-channel palaeotopographic terraces. The last-stage channel-fill is characterized by continuous, high-amplitude horizontal toconvex-upward (hat-shaped) reflections in the basal part, up to50 m thick, and by low-amplitude, onlapping or convergingcontinuous reflections in the remaining higher part. LAPs arelacking and bank collapse scars or MTDs are rare. Likewise, levéesare generally lacking, though may occur in smaller, less incisedchannel belts.

In attribute maps, the erosional channel belts appear asmoderate- to high-amplitude sinuous threads with bright patchesat the bends interpreted as channel-bend mounds. The higher-

Figure 14. Examples of hybrid-type channel belts from the study area. (A) The most common hybrid variety are channel belts with an erosional thalweg and thick levées. (BeD)Uncommon hybrids are channel belts showing both LAPs and thick levées.

Table 1Characteristics of the three examples of submarine valley-fill complexes from the study area in offshore West Africa (see seismic images in Figs. 15e17). U e valley upperreaches; L e valley lower reaches.

Main characteristics Valley-fill complex I (Fig. 15) Valley-fill complex II (Fig. 16) Valley-fill complex III (Fig. 17)

Mappable length 50.2 km 19 km 40 kmDepth 220 m (U) to 100 m (L) 180 m (U) to 250 m (L) 250 m (U) to 220 m (L)Basal width 700 m (U) to 1400 m (L) 400 m (U) to 600 m (L) 1470 m (U) to 1940 m (L)Valley-wall inclination 32� (U) to 12� (L) 16� (U) to 18� (L) 22� (U) to 25� (L)Morphology of valley margins U-part: slump scars and associated

MTDs. L-part: external levées,slump scars

Erosional terraces with LAPs,slump scars and associated MTDs

External levées, slump scars,and associated MTDs

Channel-belt types and MTDoccurrences (numbered inascending order)

(1)Non-aggradational meanderingchannel belt(2)e(7)Levéed sinuous channel belts.

(1)Erosional cut-and-fill channel belt(2)MTDs(3)e(6)levéed sinuous channel belts

(1)Meandering belt with deeplyincised last-stage channel(2)Non-aggradational meanderingchannel belt and MTDs(3)Levéed sinuous channel belt(4)Non-aggradational meanderingchannel belt(5)Levéed sinuous channel belt(6)Non-aggradational meanderingchannel belt(7)Levéed sinuous channel belt

Upward change inchannel sinuosity

2.2 / 1.9 1.6 / 1.2 2.3 / 2.5 / 1.5 / 1.8 / 1.5 / 2.6 / 1.2

Upward change inchannel width/depth ratio

3.6 / 2.0 3.0 / 2.8 2.1 / 4.3 / 2.2 / 1.9 / 3.3 / 2.0 / 2.7

Upward change inchannel migration style

Lateral migration / aggradationallateral migration and lateralswitching / aggradation andlateral switching

Vertical aggradation/ aggradation andlateral switching

Lateral migration / aggradation andlateral switching / lateral migration/ vertical aggradation / aggradationand lateral switching / lateralmigration / aggradational lateralmigration /lateral migration/ aggradation and lateral switching

Upward change in channel-beltstacking pattern

U-part: orderly stacked / isolateddisorderly offsetL-part: orderly stacked / isolatedorderly offset

Orderly stacked / isolateddisorderly offset

Disorderly stacked / orderly stacked/ isolated disorderly offset / orderlystacked / disorderly stacked

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 23

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3424

amplitude thread is often bordered by lower-amplitude patchescorresponding to the upper part of the channel-fill (Fig. 11C).Channel sinuosity is in the range of 1e2, and channel bends havemoderate lengths and bend amplitudes (Fig. 12).

Figure 15. Valley-fill complex I (Table 1) as an example of the typical submarine valley-fill(top row) indicate the time-window slices and channel-belt storeys displayed as maps belowsalt-doming to the right, which resulted in formation of an extensive external levée on the ostratigraphic evolution of the valley-fill channel system, as the succession resembles closel

5.4. Hybrid channel belts

Channel belts that show a combination of the above-describedarchitectures and cannot readily be divided into single-type

s in the study area; for description, see text. The seismic section and its interpretation. Note that the valley-fill in this case developed under the influence of syndepositionalther side of the valley. However, the doming does not seem to have much affected they other cases where no such deformation was involved.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 25

segments are considered to be hybrid varieties. The mostcommon variety are channel belts with planform characteristicssimilar to those of levéed sinuous channels, but which arecommonly larger, characterized by a well-incised thalweg andless extensive levées (Fig. 14A). Other hybrid belts show variouscombinations of LAPs, channel-bend mounds and extensivelevées at the individual bends (Fig. 14BeD). The hybrid channelbelts are distinguishable in cross-sections, but are generally moredifficult to recognize in attribute maps due to the overlappingand cross-cutting relationships of the diverse architecturalelements.

6. The evolution of channel belts in submarine valleys

The study area affords more than twenty valley-fill complexes,with multi-storey and isolated channel belts either erosionallyconfined or erosionally- to levée-confined. The valley-fills vary inscale, architectural complexity and location on the submarineslope, but their vertical succession and stacking pattern of channel-belt types appear to be similar, though not without some notabledepartures. Three examples (Table 1) are described below toillustrate the development and observed variation of valley-fillcomplexes.

Valley-fill complex I (Table 1, Fig. 15) shares its internal archi-tecture with a vast majority of valley-fills in the study area. Its basalpart consists of a high-sinuosity meandering channel belt withLAPs, bend cut-offs and a high sandstone net/gross (Fig. 15, time

Figure 16. Valley-fill complex II (Table 1) as an example of a relatively uncommon variety ofits interpretation (top row) indicate the time-window slices displayed as maps below. Notewhole stratigraphic succession a two-storey valley-fill complex set.

windows 1 and 2). It is overlain by a series of aggradational, high-sinuosity levéed channel belts (Fig. 15, time window 3) whichlack LAPs, decrease in sinuosity and assume disorderly directionsincreasingly unrelated to the direction of underlying channel belts(“disorganized” stacking sensu McHargue et al., 2011). However,the channel belts at some of their down-stream bends apparentlyevolved in continuity with the underlying meander-belt loops,showing aggradational lateral accretion and consistent directionaltrend (Fig. 15). The channel belts in the upper part of the valley-fillcomplex (Fig. 15, time windows 3 and 4) are vertically offset andconfined by external levées. Theymimic the architectural pattern ofthe underlying levéed channel belts, but have wider and deeperthalweg zones, lower sinuosity and thicker gull wing-shapedlevées. They show bright-amplitude spots on the inner side ofsome bends, which may indicate modest point bars. The topmostpart of the valley-fill complex shows a belt of large, moderatelysinuous, mud-plugged levéed channel with a directional trendsimilar to that of the underlying channel belt (Fig. 15, timewindow 5).

Valley-fill complexes II and III (Table 1) are examples illus-trating less common stratigraphic architectures. Complex II(Fig. 16) is an incised valley-fill with an uneven, terraced basebearing some “hanging” erosional relics of meander belts. Thesuccession commenced with an erosional channel belt ofmoderate sinuosity (Fig. 16, time window 1) which was coveredby valley wall-derived MTDs and overlain further by a series ofmulti- to single-storey hybrid levéed channel belts. High-

submarine valley-fill in the study area; for description, see text. The seismic section andthe erosionally superimposed younger incised valley-fill complex, which renders the

Figure 17. Valley-fill complex III (Table 1) as an example of another uncommon variety of submarine valley-fill in the study area; for description, see text. The seismic section and itsinterpretation (top row) indicate the time-window slices and channel-belt storeys displayed as maps below.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3426

amplitude reflection wedges pinching out against valley wallsrepresent internal levées and can be observed in most cross-sections of the palaeovalley (Fig. 16, time window 3). The corre-sponding channel-fills had been removed by incision of theoverlying younger valley.

Complex III in its basal part shows a meandering channel beltwith many bend cut-offs and a last-stage conduit that wasslightly deepened by incision before being filled (Fig. 17, timewindow 1). The channel belt was buried by MTDs and overlainfurther by a moderately sinuous meandering channel-belt thatevolved into an aggradational levéed channel belt (Fig. 17, timewindow 2). The overlying succession shows a double alternation

of meandering and low-sinuosity levéed channel belts (timewindows 3 and 4). The valley-fill as a whole is highly heteroge-neous, differing from those formed by consistent, uninterruptedaggradation.

The main significance of valley-fill complexes II and III is intheir showing that a particular type of channel belt may occur atvarious levels of an evolving valley-fill and that the verticalsuccession of channel-belt types may not necessarily followa predictable pattern. As discussed further in the next section, theevolution of a valley-fill system may be perturbed by internal and/or external factors, which results in considerable stratigraphicvariation.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 27

7. Discussion

7.1. The diversity of sinuous channel belts

The analysis of seismic data from the offshore study area showsthat deep-water sinuous channel belts can be differentiated on thebasis of their planform, cross-sectional geometry and componentarchitectural elements. The tentative classification of channel beltssuggested in the study sheds some new light on the genetic rela-tionships among various sinuous channels.

7.1.1. Meandering channel beltsThese channel belts are typified by high-sinuosity, regular and

smoothly-curved meander loops and a conduit with low banks andhigh width/depth ratio (Figs. 11 and 12). Their most characteristicarchitectural element are LAPs, indicating that the channel evolvedfrom relatively straight and systematically increased its sinuosityby erosion and deposition in the lateral domain. The bases and topsof non-deformed LAPs are horizontal (Figs. 2, 5 and 13), as alsoevidenced elsewhere by seismic images (Abreu et al., 2003; Mayallet al., 2006; Kolla et al., 2007, Figs. 10e12; Labourdette, 2007;Nakajima et al., 2009) and outcrop sections (Campion et al.,2000; Abreu et al., 2003; Lien et al., 2003; Shultz et al., 2005;Beaubouef et al., 2007; Cronin et al., 2007; O’Byrne et al., 2007;Wynn et al., 2007; Dykstra and Kneller, 2009; Gamberi et al.,2013; Janocko and Nemec, in press), which suggests that theprocess of submarine channel meandering occurs when no signif-icant aggradation takes place. The lack of vertical thalweg migra-tion indicates that the system is dominated by equilibrium flowswith the rate of lateral deposition balanced by the rate of lateralerosion (Kneller, 2003). Channel-bend mounds are rare inmeandering channels, probably because the cut-bank erosion andminimal thalweg aggradation render their preservation potentialvery low (Nakajima et al., 2009).

Another striking characteristic of the meandering channelbelts in the present case is the apparent lack of recognizablelevées or other localized overbank deposits in seismic sectionsand attribute maps. However, there is no doubt that the chan-nelized flows were spilling out and spreading sediment in over-bank areas. Coarse-grained LAPs, some of them fullyconglomeratic (Janocko and Nemec, in press), are clear indicationthat the flow height must have grossly exceeded the channeldepth (Dykstra and Kneller, 2009). Overbank deposition is knownto be associated with meandering channel belts (Lien et al., 2003;Arnot et al., 2007; Janbu et al., 2007; Dykstra and Kneller, 2009;Janocko and Nemec, in press) and is a prerequisite for theformation of multi-storey, vertically-stacked meandering channelbelts (e.g., Dykstra and Kneller, 2009, Fig. 8). Nevertheless,seismically-recognizable levées have been reported mainly fromnon-meandering sinuous channels (e.g., Clark and Pickering,1996; Nakajima and Satoh, 2001; Fildani and Normark, 2002).The excessive spill-out of equilibrium flows in meanderingchannel belts (Dykstra and Kneller, 2009) apparently renders thethickness/width aspect of levées very low and their relief beyondthe seismic resolution.

7.1.2. Levéed sinuous channel beltsThese channel belts have prominent, gull wing-shaped levées

and an irregularly curved conduit with a low width/depth ratio,common sharp bends and channel-bendmounds at the bend apices(Figs. 11e13). Isolated channel belts are generally typified by flatbases (Figs. 9A and 14B), whereas those stacked vertically inchannel-belt complexes generally evolve from a pre-existingmeandering channel (Fig. 9B, C and section AeA0 in Fig. 15). Thevertical stacking indicates that the channel evolved by a concurrent

aggradation of its thalweg zone and levée, which makes itreasonable to infer that the formation of these channel belts is dueto some better confined and net-depositional flows. If the flow issufficiently confined by channel, its spill-out will be moderate andhence will dissipate within a relatively short distance from thechannel, resulting in levée build-up.

Some initial incision may be required for the inception of iso-lated channel belts (see Fildani et al., 2013), even though they seemto form on an apparently flat substrate to aggrade by thalweg andlevée accretion (Fig. 13B; see also Fonnesu, 2003; Gee andGawthorpe, 2007; Clark and Cartwright, 2009; Hubbard et al.,2009; Stevenson et al., 2013). Experiments by Rowland et al.(2010) for a wide range of flows failed to produce a purely depo-sitional self-confinement of turbidity current, which suggests thatsome incipient erosional confinement of flow may be needed toinstigate the formation of aggradational levéed channels. However,the relief of this basal “inception” scour may well be of sub-seismicscale and hence unrecognizable in seismic sections.

Levéed channel belts may be laterally offset due to slight lateralmigration in continuity with the underlying belt or to intermittentlateral switching (Figs. 15e17; see also McHargue et al., 2011). Thelatter mode may involve some incipient incision (see previousparagraph). Themode of lateral shifting depends probably upon themagnitude and velocity of channel-conveyed flows in addition tothe system’s net rate of aggradation and channel infilling. Channelmigration in continuity with the underlying belt will likely occurwhen the aggradational channel remains sufficiently deep toconfine a major part of flow volumes. According to Kane et al.(2008), flows that are more than five times thicker than thechannel depth tend to deposit sediment at the outer banks, therebypotentially straightening the channel or causing its avulsion. Flowsbelow this threshold tend to deposit sediment on the inner bank,forming aggradational LAPs in an aggrading channel belt.

Another prerequisite for the formation of aggradational pointbars may be the cohesiveness of outer bank. Channel bends withaggradational LAPs are generally tangential to the valley walls(Fig.15, profile AeA0; see also Posamentier et al., 2000; Posamentierand Kolla, 2003; Deptuck et al., 2007), which typically consist ofcompacted mud-rich deposits. Notably, aggradational LAPs in thestudy area occur in valley-confined channel-belt complexes, but aregenerally lacking in unconfined levéed channel belts.

Channel-belt stacking with intermittent lateral switching willoccurwhere the channel becomes pluggedwith sediment to a pointwhen it can no longer contain the flows and avulsion occurs. Thenew conduit may significantly deviate from the previous one,resulting in a disorderly directional stacking (McHargue et al., 2011).

Channel-bendmounds are a characteristic architectural elementof levéed channel belts. In seismic attribute maps, these depositsgenerally occur in the thalweg zone near the outer bank and aremost notable at the bend apices (Fig. 6), but may extend eitherupstream or downstream from the apex (Fig. 6, top middle;Fig. 18A, B) or occur at the outer- to inner-bank transition in a bendinflection zone (Fig. 6, top right; Fig. 18C). Deposition localizedbetween the bend apex and downstream inflection point (Fig. 18A)is attributed to relatively well-confined flows, with the flow run-upon the outer bank at sharp bends causing abrupt deceleration andloss of capacity (Kane et al., 2008; Straub et al., 2008; Nakajimaet al., 2009; Amos et al., 2010). It has been suggested (Piper andNormark, 1983; Pickering et al., 1989; Clark and Pickering, 1996;Peakall et al., 2000) that the intermittent decoupling of the upperpart of the flow by overspill will likely decelerate the channel-confined part of the flow and cause rapid deposition at anddirectly downstream of the bend apex.

Deposition at the outer bank between the bend apex andupstream inflection point (Fig. 18B) is attributed to flows with

Figure 18. Schematic diagrams showing variable location of channel-bend mounds insubmarine sinuous channels and its suggested causes (based on Janocko et al., 2013);for discussion, see text. (A) Mounds formed around and directly downstream of thebend apex. (B) Mounds formed between the bend apex and upstream inflection point.(C) Mounds form in the channel-bend inflection zone.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3428

a major overspill and its large part re-entering the channel at theadjacent bend, causing there flow separation and rapid sedimentdumping (Janocko et al., 2013). Deposition at the outer- to inner-bank transition in bend inflection zone (Fig. 18C) is attributed tohigh-discharge, poorly confined depositional flows that run down-valley across the channel bends and experience localized basaldeceleration when crossing bend inflection zones (Janocko et al.,2013).

7.1.3. Erosional channel beltsThese channel belts are formed by a simple cut-and-fill process,

with little or no evidence of levées and with the last-stage channel-fill typically mud-prone, composed of heterolithic deposits (Figs.10,11 and 13). Other characteristic features of erosional channel beltsinclude low to moderate sinuosity, sediment mounds at channelbends, low width/depth ratios and steep banks (Figs. 11e13). Influvial geomorphology, such conduits are referred to as laterallyinactive sinuous channels (Schumm, 1985; Nanson, 2010).

As discussed further below, the erosional channel belts in thestudy area probably evolved by incision of incipiently meanderingor levéed channels, whose primary depositional elements wereerased in the process by bank undercutting, slumping and erosion.Only the planform of the original conduit (sinuosity, bend shapeand amplitude) would be inherited by the deepened and enlargedchannel. The entrenchment was likely due to an allogenic factor,such as the growth of salt domes or continental slope tectonics,rather than to flow hydralics alone (see Sylvester et al., 2011), as

these channel belts occur in only some valley-fills, most of themaffected by halokinesis.

Once the incising channel reached an equilibrium profile, theerosion apparently ceased and sediment bypass prevailed.Although a channel at equilibrium would normally tend tomeander, the odd hydraulic geometry of a deep and narrow conduitwith cohesive banks and inherited sinuosity could not match flowhelicoid or be readily adjusted, thus preventing lateral migration(see Nanson, 2010; Janocko et al., 2013). Most flows were likely“undersized” with respect to the channel capacity, resulting in nomajor overspill and no significant levées. Sand-prone sedimentpatches at the bend apices resulted probably from localized depo-sition due to intermittent flow deceleration (Straub et al., 2008;Nakajima et al., 2009), rather than to flow stripping by overspill(cf. Piper and Normark, 1983).

7.1.4. Hybrid channel beltsThe formation of hybrid channel belts, combining features of

the three other categories, could theoretically be due to a wholerange of factors, such as occasional “outsized” flows or majorchange in flow discharges, flow avulsions, channel confluences,variable substrate properties, base-level changes, halokinesis orslope tectonics. For example, occasional “outsized” flows maydeposit channel-bend mounds in a meandering channel (Fig. 7;Nakajima et al., 2009), straightening its path and instigatingincision, or the growth of levées in a levéed channel belt mayeffectively increase the channel depth, leading to a better flowconfinement and conduit lateral migration (Kane et al., 2008). Asdiscussed further below, the hybrid channel belts in the study areaare apparently products of various failed channel transformationsof this kind.

7.1.5. Genetic relationships among channel-belt typesA cross-plot of the bend length and amplitude of sinuous

channels (Fig. 19A) reveals two different styles of bend expansion:one dominantly transverse and the other dominantly longitudinalwith respect to the channel-belt axis (see the inset sketches inFig. 19A). The continuum of planform variation from nearly straightto highly sinuous horseshoe-shaped channels is thought to repre-sent the evolutionary trend of meandering channels (Fig. 19B),because neither the levéed nor the erosional channels show anysignificant lateral expansion or sinuosity change in their develop-ment. The two-tier trend (Fig. 19A) implies that a developingmeandering channel initially increases mainly its bend amplitudeby transverse expansion, before reaching a threshold above whichthe channel length and sinuosity increase by longitudinal expan-sion. This means also that a meandering channel, if not perturbedand transformed, should invariably reach a mature or supermatureplanform, which is consistent with the evidence from all measuredlast-stage meandering channels and with their LAPs signifyingconsiderable lateral migration.

Consequently, the levéed channels are thought to have evolvedfrom immature or submature incipient meandering conduits, per-turbed by system net aggradation, whereas the laterally inactiveerosional channels are thought to have evolved by incision of eithersome moderately sinuous levéed channels or submature/maturemeandering channels, fromwhich they inherited their sinuosity. Asto the hybrid channel belts, their most common variety (Fig. 14A)can be attributed to a failed transformation of levéed channel intoerosional channel, whereas the less common varieties (Fig. 14BeD)can be attributed to a failed transformation of meandering channelinto an aggradational levéed channel. The failed transformationscan be attributed to a weak allogenic perturbation of the turbiditicsystem’s profile, probably by minor salt movements or slopefaulting.

Figure 19. (A) Cross-plot of the bend length vs. amplitude of sinuous channels (datasetsas in Fig.12), revealing two different styles of bend expansion: dominantly transverse ordominantly longitudinal with respect to the channel-belt axis, as shown by the insetplanform sketches. (B) Interpretation of the upper plot, discussed in the text.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 29

The general characteristics of the main channel-belt types in thestudied area of West African continental palaeoslope are summa-rized in Figure 20.

7.2. The development of submarine valley-fill complexes

The majority of valley-fill complexes in the study area showa comparable stratigraphic pattern of channel-belt evolution. Thebasal part of the succession consists of meandering channel beltswith a high sandstone net/gross (Figs. 15 and 20). They are typicallyoverlain by local MTDs, which may range from mud-prone, valleywall-derived slide and slump deposits to coarser-grained depositsof co-genetic debris flows spawned by large high-density turbiditycurrents. The middle to upper part of the succession consists ofmulti-storey to isolated levéed aggradational channel belts witha decreasing sinuosity and moderate to low sandstone net/gross.The valley-fill succession as a whole is fining upwards, increasinglyheterolithic and richer in mud.

A similar stratigraphic trend has been documented from confinedchannel-belt complexes in offshore Borneo (Posamentier et al., 2000;Posamentier and Kolla, 2003), offshore Gulf of Mexico (Posamentier,2003; Posamentier and Kolla, 2003; Posamentier et al., 2007),

offshore Nigeria (Posamentier and Kolla, 2003; Deptuck et al., 2007),offshore Angola and Congo (Labourdette and Bez, 2010), offshoreEgypt (Samuel et al., 2003, Fig. 7; Cross et al., 2009), the MagallanesBasin of Chile (Macauley and Hubbard, 2013) and the Delaware Basinof Texas, USA (Beaubouef et al., 2007). A net upward fining ofconfined channel-belt complexes, attributed to increasing aggrada-tion rate, has beenpostulated also by recentmodels (McHargue et al.,2011; Sylvester et al., 2011), although the issue of channel-belt typesis considered simplistically in these studies. The event-based modelof McHargue et al. (2011) takes no account of the effects of lateralaccretion and suggests that a greater lateral migration of channels,expected to be favoured by low sand/mud ratio, will likely occur atthe latest stages of channel-complex development. In the model ofSylvester et al. (2011), high net/gross belts of laterally migratingchannels are expected to be stacked disorderly in the basal part ofchannel-belt complex and to become more orderly stacked as therate of aggradation increases. This latter suggestionmatches roughlythe pattern recognized in the present study, although the modelassumes one channel-belt type with ever-present LAPs and nochange in sinuosity with time.

Somewhat different models have been derived from studies inoffshore Egypt (Samuel et al., 2003), offshore Gabon (Wonhamet al., 2000), offshore Brazil and Angola (Mayall and Stewart,2000; Mayall et al., 2006) and the Elazi�g Basin of eastern Turkey(Cronin et al., 2005, 2007). Mayall and Stewart (2000) and Mayallet al. (2006) have postulated a model that consists of a basalerosional surface overlain by coarse sediment lag 1e2 m thick,deposited in low-sinuosity channel belt, covered with MTDsderived either locally or from distant sites; the higher part consistsof laterally migrating aggradational channel belts with high sand-stone net/gross, whereas laterally migrating high-sinuosity levéedchannel belts with low net/gross occur at the top. In this model,submarine valleys are considered to be like giant channels scouredby powerful erosive flows, which is why the depositional succes-sion is suggested to commence with a channel-fill element (basallag), rather than a valley-floor channel belt.

Submarine valleys are similarly ascribed to large erosive flows inthe models of Wonham et al. (2000) and Samuel et al. (2003, end-member B in Fig.10), wheremeandering channel belts are expectedto occur in the upper part of channel-belt complex and aggrada-tional levéed channel belts to occur in the lower part, with anoverall upward fining due to decreasing sandstone net/gross.Samuel et al. (2003) also noted that the basal part of valley-fill maycontain locally-derived MTDs and occasionally bears thin sandsinterpreted to be relic deposits of the erosive flows that incised thevalley. A slightly different stratigraphic pattern is suggested in themodel of Cronin et al. (2005, 2007), based on outcrops of severalchannel-belt complexes, where basal conglomeratic deposits ofbraided channel belts are overlain by finer-grained belts oferosional channels and covered by the low net/gross meander beltsof small, high-sinuosity channels.

Despite their variation and interpretive discrepancies, subma-rine valley-fill complexes appear to have several main features incommon e such as:

� an erosional base marking deep incision of turbiditic system;� basal deposits indicative of considerable sediment bypass withlittle or no aggradation;

� common occurrences of MTDs in the lower part;� aggradational succession of levéed sinuous channel belts,multi-storey to isolated;

� non-aggradational meandering channel belts which may occurin the basal and/or top part of valley-fill or occasionally also inthe middle part;

� an overall upward fining with decreasing sandstone net/gross.

Figure 20. Summary of the main characteristics of submarine channel belts and incised valley-fills in the studied area of West African Miocene continental slope (for discussion, see text).

M.Janocko

etal./

Marine

andPetroleum

Geology

41(2013)

7e34

30

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 31

Submarine valley-fill complexes are clearly recording periods ofa deep incision, aggradation and abandonment of a particular flowroute of turbidity currents (Fig. 20). This development suggestsa continuum of changes in the system profile, from deep erosion toan incipient “graded” state with basal lag or meandering channelbelts, and further to aggradation with levéed channel belts and toeventual abandonment (Kneller, 2003). Most of the slope valley-fills in the study area and a majority of other reported cases showsuch a pattern of development, exemplified by valley-fill complex I(Table 1, Fig. 15).

The formation of a slope valley is thought to be due to eithera rapid incision of an erosional channel belt or a slow gradualincision of a meandering channel belt (Fig. 20). In the former case,the rapid incision will prevent lateral migration and result in a V-shaped valley with a thalweg sinuosity similar to that of theoriginal channel. In the latter case, the slow incision will onlygradually restrain lateral migration, resulting in a wide-floor, U-shaped low-sinuosity valley with common terraces and a remnant,high-sinuosity meandering channel (Figs. 16 and 20; see alsoSylvester et al., 2011). The onset of aggradation will produce multi-storey meander belts, so long as the aggradation rate is not toohigh (Dykstra and Kneller, 2009, Fig. 8). As the system’s equilib-rium profile begins to rise, the increased accommodation willenhance aggradation and result in levéed channel belts. Witha gradual increase in aggradation rate, the vertically stackedchannel belts will not only decrease their sand net/gross, but alsoassume a more orderly directional pattern (McHargue et al., 2011).The transition to abandonment phase will be marked by offset-stacked to isolated levéed channel belts (Figs. 15 and 16) andsheet-like deposits formed as crevasse splays or terminal splays/lobes (Gardner and Borer, 2000; Posamentier and Kolla, 2003;Cross et al., 2009).

Variation among valley-fills can be attributed to the effect ofexternal factors (e.g., halokinesis, slope tectonics) or to an autogenicforcing due to the evacuation of sediment to the valley terminusand related base-level change, possibly accompanied by a majoraccretion of muddy deposits on the adjoining slope. A significantbase-level rise will raise the potential equilibrium profile, wherebythe aggrading valley-fill system will reach the valley top withoutattaining a “graded” state and will shift outside the valley beforereaching an equilibrium. The valley-fill will then have isolatedlevéed sinuous channel belts at the top, as exemplified by thepresent study area (Figs. 15 and 16) and as seems to be the case inthe majority of reported valley-fill complexes (see referencesabove).

In the case of no major rise in base level or considerablesediment accretion in the valley neighbourhood, the aggradingvalley-fill system may incidentally reach its equilibrium profilenear the valley top, whereby non-aggradational meanderingchannel belts will form. Similar channel belts will form when theaggrading valley-fill system approaches the valley top and thespilling-out flows begin to spread sediment over a wider area,which may dramatically reduce the intra-valley effective aggra-dation rate. In other sporadic cases, the aggrading valley-fillsystem may reach a transient equilibrium profile much earlier,at a mid-depth height of the valley, before the resulting bypass ofsediment raises further the base level and instigates renewedaggradation (Fig. 17).

The occurrence of MTDs in valley-fills is common, but by nomeans universal (Fig. 20). The emplacement of MTDs generallytakes place at an early stage, when the valley has assumed itsmaximum relief (Fig. 20), but no valley-wall collapses mustnecessarily occur or they may be very local and their products mayalso tend to be removed by erosion. The preservation potential ofMTDs increases with the onset of valley-floor aggradation.

8. Conclusions

The study was focused on the interpretation of seismic sectionsand attribute maps of sinuous channel-belt complexes from thecontinental slope in offshore West Africa, calibrated to well-coresamples and supplemented with comparative outcrop cases fromother ancient submarine slopes. Fivemain architectural elements ofchannel-belts were recognized:

� Lateral-accretion packages (LAPs) e interpreted as depositsformed by the accretion of sediment at the inner bank ofchannel bend in association with the lateral migration ofchannel thalweg.

� High-amplitude reflection patches in the thalweg zone nearouter bank e interpreted as coarse sediment mounds resultingfrom the deceleration and/or overspill of turbidity currents atchannel bends.

� Levées e recognizable as wedge-shaped sand-prone ridges atchannel margins and attributed to the overspill of channel-conveyed flows.

� Mass-transport deposits (MTDs) e recognizable as odd-shaped“chaotic” units and ascribed to local slides, slumps or debrisflows derived from the channel banks or valley walls.

� Last-stage channel-fills e recognizable as a mud-prone fillheralding and recording the abandonment of a channel.

These elements occur in various combinations, but no singlechannel belt combines all of them, which suggests that someelements may be mutually exclusive. Four end-member categoriesof sinuous channel belts were distinguished on the basis of theirplanform and transverse geometry and the architectural elementsinvolved:

� Meandering channel beltse characterized by the occurrence ofLAPs, markedly erosional base, high-sinuosity conduit,horseshoe-shaped bends and unrecognizable levées. They areformed by the lateral migration of sinuous channel and aredominated by flows combining erosion and deposition in theirlateral domain, with the overbank flow in equilibriumwith thesubstrate gradient and no significant lateral dissipation at thechannel margins.

� Levéed channel belts e characterized by a slightly incised ordepositional base, prominent levées, low-sinuosity conduit,irregular parabola-shaped and often sharp bends and commonassociated mounds. Aggradational LAPs may occur in beltswhere aggradation was combined with significant lateralchannel migration. Levées indicate overbank flows that rapidlydissipated at the channel margins.

� Erosional channel belts e characterized by concave-upwardserosional bases, moderate bulk sinuosity, smoothly-curved tosharp bends, channel-bend mounds and a general lack ofrecognizable levées and LAPs. They are formed by entrench-ment of one of the previous channel-belt types and filled byvertical aggradation, with the overbank flows negligible ornon-dissipating at the channel margins.

� Hybrid channel belts e which are typically levéed and haveerosional bases, showing characteristics of more than one ofthe other channel types. Their origin is attributed to channelsthat underwent “mutation” due to external perturbationsinvolving either flow discharge or profile gradient.

Quantitative analysis indicates that meandering channels formwhen system is near its potential equilibrium profile. They evolvefrom nearly straight to highly sinuous by increasing first the bendamplitude and then the conduit length. Levéed channels are

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3432

thought to evolve from incipient meandering conduits perturbedby aggradation and erosional channels to evolve from eithermoderately sinuous levéed or highly sinuous meandering conduits,inheriting their sinuosity. Hybrid channels signify a failed channeltransformation.

The channel belts may occur isolated, but are commonly stackeduponone another, formingmulti-storeychannel-belt complexes thatare either unconfined or developed within incised valleys. Uncon-fined channel-belt complexes, made of levéed channel belts stackedvertically in an offset “compensational” manner, are relativelyuncommon in the study area. Confined channel-belt complexespredominate, with the confinement provided by valley relief andoften also by the valley external levées at the late stage of infilling.

The majority of valley-fill complexes show a development fromdeep erosion to a transient equilibrium state with the deposition ofcoarse lag or non-aggradational meandering channel belts, andfurther to aggradation with levéed channel belts and eventualabandonment. Non-aggradational meandering channel belts mayform in the basal and/or top part of valley-fill and sporadically inthe middle part. MTDs tend to be emplaced when the valleyassumes its maximum relief, but may not necessarily be present.The observed variation among valley-fills can be attributed toexternal factors (e.g., halokinesis, slope tectonics) or to an autogenicforcing related to the evacuation of sediment from the valley, base-level change and mud accretion on the adjoining slope.

Acknowledgements

This study was a part of the first author’s 3-year PhD researchproject funded by Statoil ASA. Statoil is also acknowledged for thepermission to publish the selected seismic and well data. The studyin its initial phase benefited from the stimulating discussions withAnna Pontén, Anjali Fernandes and Ben Kneller. The manuscriptwas critically reviewed by Rick Beaubouef and Bill Arnott, whoseinsightful comments are much appreciated by the authors.

References

Abreu, V., Sullivan, M., Pirmez, C., Mohrig, D., 2003. Lateral accretion packages(LAPs): an important reservoir element in deep water sinuous channels. Marineand Petroleum Geology 20, 631e648.

Amos, K.J., Peakall, J., Bradbury, P.W., Roberts, M., Keevil, G., Gupta, S., 2010. Theinfluence of bend amplitude and planform morphology on flow and sedimen-tation in submarine channels. Marine and Petroleum Geology 27, 1431e1447.

Armitage, D.A., Romans, B.W., Covault, J.A., Graham, S.A., 2009. The influence ofmass transport-deposit surface topography on the evolution of turbiditearchitecture: the Sierra Contreras, Tres Pasos Formation (Cretaceous), southernChile. Journal of Sedimentary Research 79, 287e301.

Arnot, M.J., King, P.R., Browne, G.H., Helle, K., 2007. Channelized, innermost, basin-floor-fan morphologies, Mount Messenger Formation Waikiekie South Beachand Inland, New Zealand. In: Nilsen, T., Shew, R., Steffens, G., Studlick, J. (Eds.),Atlas of Deep-Water Outcrops. AAPG Studies in Geology, vol. 56, pp. 249e256.

Arnott, R.W.C., 2007. Stratal architecture and origin of lateral accretion deposits(LADs) and conterminuous inner-bank levee deposits in a base-of-slope sinuouschannel, lower Isaac Formation (Neoproterozoic), East-Central BritishColumbia, Canada. Marine and Petroleum Geology 24, 515e528.

Babonneau, N., Savoye, B., Cremer, M., Bez, M., 2004. Multiple terraces within thedeep incised Zaire Valley (ZaïAngo Project): are they confined levees? In:Lomas, S.A., Joseph, P. (Eds.), Confined Turbidite Systems Geological Society ofLondon Special Publication, vol. 222, pp. 91e114.

Beaubouef, R.T., Rossen, C., Lovell, R.W.W., 2007. The Beacon Channel: a newlyrecognized architectural type in the Brushy Canyon Formation, Texas, USA. In:Nilsen, T., Shew, R., Steffens, G., Studlick, J. (Eds.), Atlas of Deep-Water Outcrops.AAPG Studies in Geology, vol. 56, pp. 432e443.

Bouma, A.H., 1962. Sedimentology of Some Flysch Deposits: a Graphic Approach toFacies Interpretations. Elsevier, Amsterdam, 168 pp.

Brice, J.C., 1974. Evolution of meander loops. Geological Society of America Bulletin85, 581e586.

Bridge, J.S., 2003. Rivers and Floodplains: Forms, Processes and SedimentaryRecord. Blackwell Publishing, Oxford, 491 pp.

Campion, K.M., Sprague, A.R., Mohrig, D., Lovell, R.W., Drzewiecki, P.A.,Sullivan, M.D., Ardill, J.A., Jensen, G.N., Sickafoose, D.K., 2000. Outcrop expres-sion of confined channel complexes. 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. SEPM Gulf Coast Section 20th Bob F. PerkinsResearch Conference, pp. 127e150.

Carlson, P.R., Bruns, T.R., Molnia, B.F., Schwab, W.C., 1982. Submarine valleys in thenortheastern Gulf of Alaska: characteristics and probable origin. MarineGeology 47, 217e242.

Cartigny, M.J.B., Postma, G., Van den Berg, J.H., Mastbergen, D.R., 2010.A comparative study of sediment waves and cyclic steps based on geometries,internal structures and numerical modeling. Marine Geology 280, 40e56.

Clark, I.R., Cartwright, J.A., 2009. Interactions between submarine channel systemsand deformation in deepwater fold belts: examples from the Levant Basin,Eastern Mediterranean sea. Marine and Petroleum Geology 26, 1465e1482.

Clark, J.D., Pickering, K.T., 1996. Architectural elements and growth patterns ofsubmarine channels: applications to hydrocarbon exploration. AAPG Bulletin80, 194e221.

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. SEPMGulf Coast Section 20th Bob F. Perkins Research Conference, pp. 241e251.

Collinson, J.D., Thompson, D.B., 1982. Sedimentary Structures. Allen and Unwin,London, 207 pp.

Cook, T.W., Bouma, A.H., Chapin, M.A., Zhu, H., 1994. Facies architecture andreservoir characterization of a submarine fan channel complex, JackforkFormation, Arkanas. In: Weimer, P., Bouma, A.H., Perkins, B.F. (Eds.), SubmarineFans and Turbidite Systems: Sequence Stratigraphy, Reservoir Architecture andProduction Characteristics. SEPM Gulf Coast Section Foundation 15th AnnualResearch Conference, pp. 69e81.

Cronin, B.T., Çelik, H., Hurst, A., Türkmen, I., 2005.Mud prone entrenched deep-waterslope channel complexes from the Eocene of eastern Turkey. In: Hodgson, D.M.,Flint, S.S. (Eds.), Submarine Slope Systems: Processes and Products. GeologicalSociety of London Special Publication, vol. 244, pp. 155e180.

Cronin, B.T., Çelik, H., Hurst, A., 2007. Sinuous channels in late stages of entrencheddeep-water channel complexes: Hasret Mountain Main Channel, Turkey. In:Nilsen, T., Shew, R., Steffens, G., Studlick, J. (Eds.), Atlas of Deep-Water Outcrops.AAPG Studies in Geology, vol. 56, pp. 368e372.

Cross, N.E., Cunningham, A., Cook, R.J., Taha, A., Esmaie, E., El Swidan, N., 2009. Three-dimensional seismic geomorphology of a deep-water slope-channel system: theSequoia field, offshore west Nile delta, Egypt. AAPG Bulletin 93, 1063e1086.

Dalrymple, R.W., Boyd, R., Zaitlin, B.A., 1994. Incised-Valley Systems: Origin andSedimentary Sequences. In: SEPM Special Publication, vol. 51, 391 pp.

Demyttenaere, R., Tromp, J.P., Ibrahim, A., Allman-Ward, P., Meckel, T., 2000. Theinfluence of bend amplitude and planform morphology on flow and sedimen-tation in submarine channels. 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 Reser-voirs of the World. SEPM Gulf Coast Section Foundation 20th Annual ResearchConference, pp. 304e317.

Deptuck, M.E., Steffens, G.S., Barton, M., Pirmez, C., 2003. Architecture and evolutionof upper fan channel-belts on the Niger Delta slope and in the Arabian Sea.Marine and Petroleum Geology 20, 649e676.

Deptuck, M.E., Sylvester, Z., Pirmez, C., O’Byrne, C., 2007. Migration-aggradationhistory and 3-D seismic geomorphology of submarine channels in the Pleisto-cene Benin-major Canyon, western Niger Delta slope. Marine and PetroleumGeology 23, 406e433.

Droz, L., Bellaiche, G., 1985. Rhone Deep-Sea fan: morphostructure and growthpattern. AAPG Bulletin 69, 460e479.

Dykstra, M., Kneller, B.C., 2009. Lateral accretion in a deep-marine channelcomplex: implications for channelized flow processes in turbidity currents.Sedimentology 56, 1411e1432.

Elliott, T., 2000. Depositional architecture of a sand-rich, channelized turbiditesystem: the Upper Carboniferous Ross Sandstone Formation, western Ireland.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. SEPMGulf Coast Section 20th Bob F. Perkins Research Conference, pp. 342e373.

Euzen, T., Eschard, R., Alboyuy, E., Deschamps, R., 2007. Reservoir architecture ofa turbidite channel complex in the Pab Formation, Pakistan. In: Nilsen, T.,Shew, R., Steffens, G., Studlick, J. (Eds.), Atlas of Deep-Water Outcrops. AAPGStudies in Geology, vol. 56, p. 20. CD-ROM.

Faulkenberry, L., 2004. High-resolution seismic architecture of upper slopesubmarine channel systems: Gulf of Mexico and offshore Nigeria. UnpublishedPh.D. thesis, University of Leeds, 258 pp.

Fildani, A., Normark, W.R., 2002. The stratigraphic evolution of Monterey Fan areaand growth patterns of related channel complexes. AAPG Bulletin 12, A54.Annual Meeting, Program.

Fildani, A., Hubbard, S.M., Covault, J.A., Maier, K.L., Romans, B.W., Traer, M.,Rowland, J.C., 2013. Erosion at inception of deep-sea channels. Marine andPetroleum Geology 41, 48e61.

Fonnesu, F., 2003. 3D seismic images of a low-sinuosity slope channel and relateddepositional lobe (West Africa deep-offshore). Marine and Petroleum Geology20, 615e629.

Gamberi, F., Rovere, M., Dykstra, M., Kane, I.A., Kneller, B.C., 2013. Integratingmodern seafloor and outcrop data in the analysis of slope channel architectureand fill. Marine and Petroleum Geology 41, 83e103.

Gardner, M.H., Borer, J.M., 2000. Submarine channel architecture along a slope tobasin profile, Brushy Canyon Formation, West Texas. In: Bouma, A.H., Stone, C.G.

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e34 33

(Eds.), Fine-grained Turbidite Systems. AAPG Memoir 72/SEPM Special Publi-cation, vol. 68, pp. 195e214.

Gee, M.J.R., Gawthorpe, R.L., 2007. Early evolution of submarine channels offshoreAngola revealed by three-dimensional seismic data. In: Davies, R.J.,Posamentier, H.W., Wood, L.J., Cartwright, J.A. (Eds.), Seismic Geomorphology:Applications to Hydrocarbon Exploration and Production. Geological Society ofLondon Special Publication, vol. 277, pp. 223e235.

Harms, J.C., Southard, J.B., Walker, R.G., 1982. Structures and Sequences in ClasticRocks. SEPM Short Course No. 9 Lecture Notes. Society of Economic Paleon-tologists and Mineralogists, Tulsa, 250 pp.

Heiniö, P., Davies, R.J., 2007. Knickpoint migration in submarine channels inresponse to fold growth, western Niger Delta. Marine and Petroleum Geology24, 434e449.

Hodgson, D.M., 2009. Distribution and origin of hybrid beds in sand-rich submarinefans of the Tanqua depocentre, Karoo Basin, South Africa. Marine and PetroleumGeology 26, 1940e1956.

Hubbard, S.M., De Ruig, M.J., Graham, S.A., 2009. Confined channel-levee complexdevelopment in an elongate depo-center: deep-water Tertiary strata of theAustrian Molasse Basin. Marine and Petroleum Geology 26, 85e112.

Janbu, N.E., Nemec, W., Kırman, E., Özaksoy, V., 2007. Facies anatomy of a channel-ized sand-rich turbiditic system: the Eocene Kusuri Formation in the SinopBasin, north-central Turkey. In: Nichols, G., Paola, C., Williams, E.A. (Eds.),Sedimentary Environments, Processes and Basins e A Tribute to Peter Friend.International Association of Sedimentologists Special Publication, vol. 38,pp. 457e517.

Janocko, M., Cartigny, M.J.B., Nemec, W., Hansen, E.W.M., 2013. Turbidity currenthydraulics and sediment deposition in erodible sinuous channels: laboratoryexperiments and numerical simulations. Marine and Petroleum Geology 41,222e249.

Janocko, M., Nemec, W. The facies architecture and formation of deep-water pointbars: an outcrop perspective. Sedimentology, in press.

Kane, I.A., Hodgson, D.M., 2011. Sedimentological criteria to differentiate submarinechannel levee subenvironments: exhumed examples from the Rosario Fm. (UpperCretaceous) of Baja California, Mexico, and the Fort Brown Fm. (Permian), KarooBasin, S. Africa. Marine and Petroleum Geology 28, 807e823.

Kane, I.A., McCaffrey, W.D., Peakall, J., 2008. Controls on sinuosity evolution withinsubmarine channels. Geology 36, 287e290.

Kane, I.A., Dykstra, M., Kneller, B.C., Tremblay, S., McCaffrey, W.D., 2009. Architec-ture of a coarse grained channel-levee system: the Rosario Formation, BajaCalifornia, Mexico. Sedimentology 56, 2207e2234.

Kane, I.A., McCaffrey, W.D., Peakall, J., Kneller, B.C., 2010. Submarine channel leveeshape and sediment waves from physical experiments. Sedimentary Geology223, 75e85.

Kneller, B.C., 2003. The influence of flow parameters on turbidite slope channelarchitecture. Marine and Petroleum Geology 20, 901e910.

Kolla, V., Bourges, P., Urruty, J.-M., Safa, P., 2001. Evolution of deep-water Tertiarysinuous channels offshore Angola (west Africa) and implications for reservoirarchitecture. AAPG Bulletin 85, 1373e1405.

Kolla, V., Posamentier, H.W., Wood, L.J., 2007. Deep-water and fluvial sinuouschannels: characteristics, similarities and dissimilarities, and modes of forma-tion. Marine and Petroleum Geology 24, 388e405.

Labourdette, R., 2007. Integrated three-dimensional modeling approach of stackedturbidite channels. AAPG Bulletin 91, 1603e1618.

Labourdette, R., Bez, M., 2010. Element migration in turbidite systems: random orsystematic depositional processes? AAPG Bulletin 94, 345e368.

Lien, T., Walker, R.G., Martinsen, O.J., 2003. Turbidites in the Upper carboniferousRoss Formation, western Ireland: reconstruction of a channel and spilloversystem. Sedimentology 50, 113e148.

Lowe, D.R., 1982. Sediment gravity flows. II. Depositional models with specialreference to the deposits of high-density turbidity currents. Journal of Sedi-mentary Research 52, 279e297.

Lowe, D.R., 1988. Suspended-load fallout rate as an independent variable in theanalysis of current structures. Sedimentology 35, 765e776.

Macauley, R.V., Hubbard, S.M., 2013. Slope channel sedimentary processes andstratigraphic stacking, Cretaceous Tres Pasos Formation slope system, ChileanPatagonia. Marine and Petroleum Geology 41, 146e162.

May, J.A., Warme, J.E., Slater, R.A., 1983. Role of submarine canyons on shelfbreakerosion and sedimentation: modern and ancient examples. In: Stanley, D.J.,Moore, G.T. (Eds.), The Shelfbreak: Critical Interface on Continental Margins.SEPM Special Publication, vol. 33, pp. 315e332.

Mayall, M., Stewart, I., 2000. The architecture of turbidite slope channels. 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. SEPMGulf Coast Section 20th Bob F. Perkins Research Conference, pp. 578e586.

Mayall, M., Jones, E., Casey, M., 2006. Turbidite channel reservoirs e key elements infacies prediction and effective development. Marine and Petroleum Geology 23,821e841.

McHargue, T., Pyrcz, M.J., Sullivan, M.D., Clark, J., Fildani, A., Romans, B., Covault, J.,Levy, M., Posamentier, H., Drinkwater, N., 2011. Architecture of turbidite channelsystem on the continental slope: patterns and predictions. Marine and Petro-leum Geology 28, 728e743.

Miall, A.D., 1985. Architectural-element analysis: a new method of facies analysisapplied to fluvial deposits. Earth-Science Reviews 22, 261e308.

Morris, W.R., Normark, W.R., 2000. Sedimentologic and geometric criteria forcomparing modern and ancient sandy turbidite elements. 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. SEPM Gulf CoastSection 20th Bob F. Perkins Research Conference, pp. 606e623.

Mutti, E., Normark, W.R., 1987. Comparing examples of modern and ancientturbidite systems: problems and concepts. In: Legget, J.K., Zuffa, G.G. (Eds.),Deep Water Clastic Deposits: Models and Case Histories. Graham and Trotman,London, pp. 1e38.

Nakajima, T., Peakall, J., McCaffrey,W.D., Paton, D.A., Thompson, P.J., 2009. Outerbankbars: a new intra-channel architectural element within sinuous submarine slopechannels. Journal of Sedimentary Research 79, 872e886.

Nakajima, T., Satoh, M., 2001. The formation of large mudwaves by turbiditycurrents on the levees of the Toyama deep-sea channel, Japan Sea. Sedimen-tology 48, 435e463.

Nanson, R.A., 2010. Flow fields in tightly curving meander bends of low width-depth ratio. Earth Surface Processes and Landforms 35, 119e135.

Nanson, G.C., Knighton, A.D., 1996. Anabranching rivers: their cause, character andclassification. Earth Surface Processes and Landforms 21, 217e239.

O’Byrne, C.J., Barton, M.D., Steffens, G.S., Pirmez, C., Buergisser, H., 2007. Architec-ture of a laterally migrating channel complex: Channel 4, Isaac Formation,Windermere Supergroup, Castle Creek North, British Columbia, Canada. In:Nilsen, T., Shew, R., Steffens, G., Studlick, J. (Eds.), Atlas of Deep-Water Outcrops.AAPG Studies in Geology, vol. 56, pp. 115e118.

O’Connell, S.,McHugh, C., Ryan,W.B.F.,1995. Unique fanmorphology in an entrenchedthalweg channel on the Rhône Fan. In: Pickering, K.T., Hiscott, R.N., Kenyon, N.H.,Ricci Lucchi, F., Smith, R.D.A. (Eds.), Atlas of Deep-Water Environments: Archi-tectural Style in Turbidite Systems. Chapman and Hall, London, pp. 80e83.

Peakall, J., McCaffrey, W.D., Kneller, B.C., 2000. A process model for the evolution,morphology, and architecture of sinuous submarine channels. Journal of Sedi-mentary Research 70, 434e448.

Peakall, J., Amos, K.J., Keevil, G.M., Bradbury, P.W., Gupta, S., 2007. Flow processesand sedimentation in submarine channel bends. Marine and PetroleumGeology 24, 470e486.

Phillips, S., 1987. Dipmeter interpretation of turbidite-channel reservoir sandstones,Indian Draw Field, New Mexico. In: Tillman, R.W., Weber, K.J. (Eds.), ReservoirSedimentology. SEPM Special Publication, vol. 40, pp. 113e128.

Pickering, K.T., Hiscott, R.N., Hein, F.J., 1989. Deep-marine Environments: ClasticSedimentation and Tectonics. Kluwer Academic Publishers, London, 424 pp.

Piper, D.J.W., Normark, W.R., 1983. Turbidite depositional patterns and flow charac-teristics, Navy submarine fan, California Borderland. Sedimentology 30, 681e694.

Pirmez, C., Beaubouef, R.T., Friedmann, S.J., Mohrig, D.C., 2000. Equilibrium Profileand base-level in submarine channels: examples from Late Pleistocene systemsand implications for the architecture of deepwater reservoirs. 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. SEPM Gulf CoastSection 20th Bob F. Perkins Research Conference, pp. 782e805.

Posamentier, H.W., 2003. Depositional elements associated with a basinfloorchannelelevel system: case study from the Gulf of Mexico. Marine and Petro-leum Geology 20, 667e690.

Posamentier, H.W., Kolla, V., 2003. Seismic geomorphology and stratigraphy ofdepositional elements in deep-water settings. Journal of Sedimentary Research73, 367e388.

Posamentier, H.W., Meizarwan, Wisman, P.S., Plawman, T., 2000. Deep waterdepositional systems e Ultra-deep Makassar Strait, Indonesia. 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. SEPM Gulf CoastSection 20th Bob F. Perkins Research Conference, pp. 806e816.

Posamentier, H.W., Davies, R.J., Cartwright, J.A., Wood, L.J., 2007. Seismic geomor-phology e an overview. In: Davies, R.J., Posamentier, H.W., Wood, L.J.,Cartwright, J.A. (Eds.), Seismic Geomorphology: Applications to HydrocarbonExploration and Production. Geological Society of London Special Publication,vol. 277, pp. 1e14.

Postma, G., Nemec, W., Kleinspehn, K., 1988. Large floating clasts in turbidites:a mechanism for their emplacement. Sedimentary Geology 58, 47e61.

Prather, B.E., 2003. Controls on reservoir distribution, architecture and stratigraphictrapping in slope settings. Marine and Petroleum Geology 20, 529e545.

Rowland, J.C., Hilley, G.E., Fildani, A., 2010. A test of initiation of submarineleveed channelsbydeposition alone. Journalof SedimentaryResearch80, 710e727.

Samuel, A., Kneller, B.C., Ralan, S., Sharp, A., Parsons, C., 2003. Prolific deep-marineslope channels of the Nile Delta, Egypt. AAPG Bulletin 87, 541e560.

Schumm, S.A., 1985. Patterns of alluvial rivers. Annual Reviews of Earth and Plan-etary Science 13, 5e27.

Shanmugam, G., Moiola, R.J., 1988. Submarine fans: characteristics, models, classi-fication, and reservoir potential. Earth-Science Reviews 24, 383e428.

Shultz, M.R., Fildani, A., Cope, T.D., Graham, S.A., 2005. Deposition and stratigraphicarchitecture of an outcropping ancient slope system: Tres Pasos Formation,Magallanes Basin, southern Chile. In: Hodgson, D.M., Flint, S.S. (Eds.), Subma-rine Slope Systems: Processes and Products. Geological Society of LondonSpecial Publication, vol. 244, pp. 27e50.

Skene, K.I., Piper, D.J.W., Hill, P.S., 2002. Quantitative analysis of variations indepositional sequence thickness from submarine channel levees. Sedimen-tology 49, 1411e1430.

Skene, K.I., Piper, D.J.W., 2005. Late Cenozoic evolution of Laurentian Fan: devel-opment of a glacially-fed submarine fan. Marine Geology 27, 67e92.

Sprague,A.R., Sullivan,M.D.,Campion,K.M., Jensen,G.N.,Goulding, F.J., Sickafoose,D.K.,Jenette, D.C., 2002. The physical stratigraphy of deep-water strata: a hierarchial

M. Janocko et al. / Marine and Petroleum Geology 41 (2013) 7e3434

approach to theanalysis of genetically related stratigraphic elements for improvedreservoir prediction. AAPG Annual Meeting Abstracts, Houston, pp. 10�13.

Stevenson, C.J., Talling, P.J., Wynn, R.B., Masson, D.G., Hunt, J.E., Frenz, M,Akhmetzhanhov, A., Cronin, B.T., 2013. The flows that left no trace: very large-volume turbidity currents that bypassed sediment through submarine channelswithout eroding the sea floor. Marine and Petroleum Geology 41, 186e205.

Straub, K.M., Mohrig, D., McElroy, B., Buttles, J., 2008. Interactions between turbiditycurrents and topography in aggrading sinuous submarine channels: a labora-tory study. GSA Bulletin 120, 368e385.

Straub, K.M., Mohrig, D., Buttles, J., McElroy, B., Pirmez, C., 2011. Quantifying the influ-ence of channel sinuosity on the depositional mechanics of channelized turbiditycurrents: a laboratory study. Marine and Petroleum Geology 28, 744e760.

Sylvester, Z., Pirmez, C., Cantelli, A., 2011. A model of submarine channel-leveeevolution based on channel trajectories: implications for stratigraphic archi-tecture. Marine and Petroleum Geology 28, 716e727.

Timbrell, G., 1993. Sandstone architecture of the Balder Formation depositionalsystem, UK Quadrant 9 and adjacent areas. In: Parker, J.R. (Ed.), PetroleumGeology of Northwest Europe. Proceedings of the 4th Conference, GeologicalSociety of London, pp. 107e121.

Wonham, J.P., Jayr, S., Mougamba, R., Chuilon, P., 2000. 3D sedimentary evolution ofa canyon fill (Lower Miocene-age) from the Mandorove Formation, offshoreGabon. Marine and Petroleum Geology 17, 175e197.

Wynn, R.B., Cronin, B.T., Peakall, J., 2007. Sinuous deep-water channels: genesis,geometry and architecture. Marine and Petroleum Geology 24, 341e387.