Seafloor topography affects the sediment gravity flows that interact with it. Understanding this interaction is critical for accurate predictions of sediment distribution and paleogeographic or structural reconstructions of deep-water basins. The effects of seafloor topography can be seen from the bed scale, through facies transitions toward intra-basinal slopes, to the basin scale, where onlap patterns reveal the spatial evolution of deep-water systems. Basin-margin onlap patterns are typically attributed to allogenic factors, such as sediment supply signals or subsidence rates, with few studies emphasizing the importance of predictable spatio-temporal autogenic flow evolution. This study aims to assess the autogenic controls on onlap by documenting onlap styles in the confined Eocene-to-Oligocene deep-marine Annot Basin of SE France. Measured sections, coupled with architectural observations, mapping, and paleogeographical interpretations, are used to categorize onlap styles and place them within a generic stratigraphic model. These observations are compared with a simple numerical model. The integrated stratigraphic model predicts that during progradation of a deep-water system into a confined basin successive onlap terminations will be partially controlled by the effect of increasing flow concentration. Initially thin-bedded low-density turbidites of the distal lobe fringe are deposited and drape basinal topography. As the system progrades these beds become overlain by hybrid beds and other deposits of higher-concentration flows developed in the proximal lobe fringe. This transition is therefore marked by intra-formational onlap against the underlying and lower-concentration lobe fringe that drapes the topography. Continued progradation results in deposition of lower-concentration deposits in the lobe off-axis, resulting in either further intra-formational onlap against the lobe fringe or onlap directly against the hemipelagic basin margin. Basinal relief is gradually reduced as axial and higher-volume flows become more prevalent during progradation, causing the basin to become a bypass zone for sediment routed down-dip. This study presents an autogenic mechanism for generating complex onlap trends without the need to invoke allogenic processes. This has implications for sequence-stratigraphic interpretations, basin subsidence history, and forward modeling of confined deep-water basins.

Deep-water lobes are amongst the largest sedimentary bodies on Earth and comprise terrigenous sediment shed from the adjacent continental shelf and slope (e.g., Piper et al. 1999; Talling et al. 2007; Prélat et al. 2010; Clare et al. 2014). They offer a record of Earth's climate and sediment transport history, form valuable hydrocarbon reservoirs, aquifers, and are sites of mineral accumulation (e.g., Weimer and Link 1991; Hodgson 2009; Sømme et al. 2011; Bell et al. 2018a). Sediment-gravity-flow evolution across unconfined deep-water settings results in a fairly uniform radial spreading and deceleration of flows, and development of elongate to lobate sedimentary bodies with predictable facies transitions (e.g., Baas 2004; Hodgson et al. 2009; Spychala et al. 2017).

Sediment gravity flows encountering seafloor topography in confined-basin settings form a range of onlap geometries that are often associated with complicated sedimentary facies (Fig. 1) (e.g., Kneller 1991; Haughton 1994; Wynn et al. 2000; McCaffrey and Kneller 2001; Amy et al. 2004; Gervais et al. 2004; Lomas and Joseph 2004; Puigdefàbregas et al. 2004; Smith 2004a, 2004b; Smith and Joseph 2004; Gardiner 2006; Mayall et al. 2010; Tinterri et al. 2016; Hansen et al. 2019). The effect of seafloor topography on sediment gravity flows, their deposits, and onlap styles has been studied through outcrop data (e.g., Kneller et al. 1991; Sinclair 1994; Bakke et al. 2013), subsurface data (e.g., Prather et al. 1998, 2012; Covault and Romans 2009; Bakke et al. 2013), and numerical models (e.g., Smith 2004b; Kubo 2004; Gardiner 2006; Sylvester et al. 2015) and physical models (e.g., Kneller 1995; Kneller and McCaffrey; 1995; Amy et al. 2004; Brunt et al. 2004; Kubo 2004). Seafloor topography is generated by a variety of geological processes, such as: pre-depositional tectonic deformation (e.g., Jackson et al. 2008; Kilhams et al. 2012), syn-depositional tectonic deformation (e.g., Wilson et al. 1992; Haughton 2000; Grecula et al. 2003; Hodgson and Haughton 2004; Tomasso and Sinclair 2004; Kane et al. 2010; Salles et al. 2014), mass transport deposit relief (e.g. Armitage et al. 2009; Ortiz-Karpf et al. 2015, 2016; Kneller et al. 2016; Soutter et al. 2018) and salt diapirism (e.g., Hodgson et al. 1992; Kane et al. 2012; Prather et al. 2012; Oluboyo et al. 2014; Doughty-Jones et al. 2017). Improved prediction of the distribution of sediment gravity flow deposits around seafloor topography is therefore critical for both paleogeographic reconstructions (e.g., Pinter et al. 2017) and stratigraphic hydrocarbon or CO2 trap risking (e.g., McCaffrey and Kneller 2001).

The onlap geometry (3D shape of an event bed or sequence of related event beds at pinch-out) and facies (internal sedimentary characteristics of an event bed at pinch-out) (Fig. 1A), herein termed onlap termination, are controlled by: 1) flow magnitude, duration, velocity, thickness, concentration, and sediment composition; 2) the gradient and incidence angle of the counter-slope; and 3) seafloor composition and induration. Typically, high-concentration flows and steep counter-slopes cause abrupt terminations, whereas low-concentration flows and shallow counter-slopes cause draped terminations (Fig. 1) (Smith and Joseph 2004; Bakke et al. 2013). Flows with a high mud content may also be more prone to varying degrees of rheological transformation approaching counter-slopes, resulting in complicated facies distributions at confining basin margins (Fig. 1) (Barker et al. 2008; Patacci et al. 2014; Southern et al. 2015).

Recent field-based studies on the spatial and temporal evolution of unconfined submarine lobes have used the longitudinal evolution of flows and their associated facies to establish criteria for differentiating lobe sub-environments at the bed scale (e.g., Prélat et al. 2009; Grundvag et al. 2014; Spychala et al. 2017). The applicability of these facies associations to confined lobes and the complex system-scale stacking patterns that they may produce is only recently being investigated (e.g., Marini et al. 2015; Spychala et al. 2015, 2017; Bell et al. 2018b; Fonnesu et al. 2018; Liu et al. 2018). Previous flow-dependant onlap models mainly focused on end-member geometries (e.g., McCaffrey and Kneller 2001; Smith 2004b; Smith and Joseph 2004). As yet there is no generic model to account for how the wide variety of deposits resulting from the longitudinal evolution of sediment gravity flows will manifest themselves at onlap surfaces through the fill of a confined basin.

This study uses the well-constrained Cenozoic Annot Basin of SE France to integrate bed-scale and basin-scale onlap observations into a generically applicable depositional model. The aims of this study are to: i) reappraise the Annot Basin stratigraphy with respect to specific deep-water sub-environments, with particular emphasis on the poorly studied eastern exposures of the basin, ii) document lateral facies changes within beds approaching the basin margin and relate these facies changes to longitudinal flow evolution, iii) assess how the longitudinal evolution of confined flows impacts onlap geometry and stacking patterns, and iv) integrate these observations into a generic model for the evolution of onlap in deep-water basins.

Basin Structure

The 160-km-long and 80-km-wide (Clark and Stanbrook 2001) Cenozoic foreland basin of the western Alps formed due to SW-directed collision of the Adria and European plates, and subsequent loading of the European plate by the Alpine orogenic wedge (Figs. 2, 3) (e.g., Ford et al. 1999). This orogenic deformation is represented in the foreland-basin stratigraphy by a progressive younging to the southwest (e.g., Ford et al. 1999; Du Fornel et al. 2004). Sediment deposition in the basin was affected by complicated basinal topography (e.g., Joseph and Lomas 2004), which formed due to Late Cretaceous northward-directed Pyrenean compression that was subsequently overprinted by Cenozoic SW-directed Alpine compression (Fig. 3) (e.g., Apps et al. 2004). This resulted in NW–SE-oriented synclinal sub-basins with E–W anticlinal sills. The synclines are interpreted as the surface expression of underlying Alpine thrust-fault-propagation folds (Fig. 3) (e.g., Elliott et al. 1985; Apps 1987; Ravenne et al. 1987).

The Annot sub-basin, herein termed the Annot Basin, is one of the exhumed synclinal Cenozoic depocenters in the Alpine foreland basin, representing the proximal end of the deep-marine Annot–Grand Coyer–Chalufy chain of sub-basins (Figs. 2, 3) (see Joseph and Lomas 2004, for review). The Annot Basin is bounded to the south by the SW–NE Rouaine Fault Zone, which acted as an entry point for sediment gravity flows into the basin (e.g., Salles et al. 2014) (Figs. 2, 3). The SW–NE Braux normal fault is related to this fault zone and created local bathymetric relief during the late Eocene (e.g., Tomasso and Sinclair 2004) (Figs 3, 4). The western and eastern margins of the basin are defined by fault-propagation anticlines created by Mesozoic blind thrusts (Figs. 3, 4) (e.g., Apps 1987). The eastern margin is formed by the Melina anticline, or “kink zone” (Apps 1987), and the western margin by the Puy du Rent anticline (Fig. 3; 4B). These anticlines were developing during deposition of the Annot turbidites, causing syn-depositional rotation of the basin depocenter towards the west (Fig. 3) (e.g., Salles et al. 2014).

E–W-oriented Pyreneo-Provençal structures also affect the Annot Basin structure, with the northern extent of the basin formed by the Aurent anticline (Fig. 3). This structure forms a gently southward-dipping terminal slope (e.g., McCaffrey and Kneller 2004). More minor basin-floor relief may have been formed by the E–W Fugeret anticline, which lies between the Rouiane fault zone and the Aurent anticline (Fig. 3) (Salles et al. 2014). These E–W-oriented structures show little evidence of major syn-depositional movement due to compressional deformation being dominantly driven by the SW-directed Alpine orogeny during the Eocene and Oligocene.

Stratigraphic Evolution

The Annot Basin has the same transgressive Cenozoic foreland-basin stratigraphy as is seen across the western Alps (Figs. 2, 3) (e.g., Sinclair 1997), with Oligocene shallow-marine limestones of the Calcaires Nummulitique overlain by deep-marine marls of the Marnes Bleues. The Marnes Bleues records the deepening of the basin, with foraminifera suggesting water depths of ∼ 100 m at the base of the succession, to ∼ 800 m by the end of deposition (Mougin 1978). Supply of siliciclastic sediment began abruptly in the late Eocene (35.2 Ma) as the Corsica–Sardinia massifs were uplifted via subduction-related back-thrusting to the south (Fig. 2) (e.g., Stanley and Mutti 1968; Apps 1987). This resulted in a depositional shift from the marls of the underlying Marnes Bleues, into south-to-north dispersing clastic sediment gravity flows of the Grès d'Annot. An upwards coarsening trend in the Grès d'Annot suggests progradation of the clastic system, most-likely related to fan-delta advance (Fig. 3) (e.g., Puigdefàbregas et al. 2004).

During early clastic deposition the Annot Basin was located on the western side of a distal submarine fan extending over the foreland basin, with flows entering the basin from syncline-bounded fan deltas to the south (Fig. 2) (e.g., Stanley 1980; Sinclair 2000; Joseph and Lomas 2004) and being dispersed northwards through relay ramps in the Rouaine Fault Zone (Fig. 3) (Joseph and Lomas 2004; Salles et al. 2014). This early deposition is represented in the Annot Basin by low-density turbidites, often referred to as the Marnes Brunes Inferiérures (e.g., Stanbrook and Clark 2004), which form the distal equivalent of the Grès d'Annot. The lowermost Grès d'Annot member, termed Le Ray (Puigdefàbregas 2004) or the A Member (Du Fornel et al. 2004), onlaps both the underlying Marnes Brunes and Marnes Bleues slope (Fig. 3). Early flows were confined by the Braux Fault to the west, a combination of the Braux Fault and the Fugeret anticline to the north, and the Melina anticline to the northeast (Salles et al. 2014). The sediment entry point shifted throughout deposition of the Grès d'Annot, with a more easterly late Eocene entry point suggested for these early flows, as evidenced by an up-stratigraphy rotation of paleocurrents from NE- to NW-directed (McCaffrey and Kneller 2004; Salles et al. 2014). Alternatively, this may be an apparent repositioning of the sediment entry point as the basin depocenter itself migrated gradually westward due to Alpine compression (Salles et al. 2011, 2014). This deformation is believed to have been continuous throughout the fill of the basin; however, the apparent rate of deformation is suppressed in the Grès d'Annot due to the higher depositional rates associated with gravity-flow deposition (Apps et al. 2004). The remaining Grès d'Annot members were confined by the major Melina (east), Puy du Rent (west), and Aurent anticlines (north) as the topography of the Braux Fault and Fugeret Anticline was healed relatively early in the Oligocene (Fig. 3) (Salles et al. 2014).

The basin gradually filled throughout the early Oligocene, with contemporaneous deposition occurring in the parallel Grand Coyer sub-basin to the northeast (Salles et al. 2014). Once the basin was largely filled, the Aurent Anticline ceased to terminally confine flows, and flows bypassed the Annot Basin into the Grand Coyer and Chalufy sub-basins (Fig. 2) (e.g., Apps 2004; Salles et al. 2014). Few channel fills are seen within the Grès d'Annot succession, and the depositional architecture is interpreted as being predominantly sheet-like (Apps 1987).

The dataset comprises 50 (581 m cumulatively) sedimentary logs collected along sections predominantly oriented oblique to depositional-dip along the eastern margin of the Annot Basin (Figs. 3, 4A, 5). Logs were collected at 1:10 scale, and individual beds were walked out at outcrop (Figs. 6, 7). Higher resolution 1:2 scale logs were collected from some beds approaching onlap. Logs within thin-bedded facies were collected at a 1:5 scale to allow accurate representation of their thicknesses and structures. Samples of individual facies and individual beds were collected in order to quantitatively constrain lateral grain-size and matrix changes. 104 3D paleocurrent measurements were collected (Fig. 5), with 2D paleocurrent measurements qualitatively noted.

Margin Correlation

Sedimentological contacts between the discrete members of the Grès d'Annot from the geological maps of Puigdefàbregas et al. (2004), Du Fornel et al. (2004), and Salles et al. (2014) were ground-truthed and compared to observations made by this study (Fig. 3). The observations of stratigraphic contacts made during this study most closely agree with those made by Puigdefàbregas et al. (2004); therefore their geological map was used for placement of sedimentary logs within members and for the intra-member correlation of sedimentary logs (Figs. 5, 6). This allows facies transitions across the basin to be assessed both spatially and temporally. Where member boundaries are unclear due to the resolution of the geological map the top of individual members is defined by either abrupt facies changes, commonly an abrupt coarsening and thickening of event beds, or lateral relationships and correlations (Fig. 5). An attempt has been made to reconcile the nomenclature used by Puigdefàbregas et al. (2004) and Salles et al. (2014) to enable comparisons to be made between the two (Fig. 3).

Detailed correlations of the Le Ray member are based on the identification of key surfaces, such as onlap surfaces, and walking out of individual beds (Fig. 6). Where beds could not be walked out, units were correlated based on the methodology of Prélat et al. (2009) for the identification of the hierarchical elements that builds lobes, e.g., beds, lobe elements, and lobes.

Facies associations (FA) have been interpreted based on the dominant lithofacies (LF) and depositional features of a given succession (Table 1). The dominant lithofacies has been described and interpreted in each facies association in order to justify their placement in that sub-environment. The lobe sub-environments of Spychala et al. (2017) are used because they best fit the observations made in this study (Figs. 8, 9). The onlap geometry of each lithofacies, and therefore the inferred onlap geometry of the facies association in which that lithofacies dominantly occurs, is summarized in Table 1. Facies associations are presented in the following section from proximal to distal positions on the lobe, in descending stratigraphic order and, broadly, in descending order of thickness.

FA 1 Lobe-Axis


Facies association 1 is composed dominantly of one lithofacies: thick-bedded (0.5–2 m) sandstones (LF 1A) (Fig. 10A, B, E) with thin-bedded (0.01–0.5 m) sandstones (LF 1B) and medium-bedded sandstones (0.1–0.8 m) (LF 3) commonly associated. LF 1B is composed of medium- to cobble-grade (most typically coarse-grained), poorly sorted, massive sandstones (Fig. 10B) and is less prevalent than LF 1A. Individual beds have erosional bases, often with groove, flute, and tool marks, and irregular tops. Beds are often lenticular, thickening and thinning from < 10 cm before pinching out over tens of meters. The beds often occur within successions of medium-bedded sandstones (LF 2) below packages of LF 1.

The thick-bedded LF 1A sandstones are medium-grained to granular sandstones with bed bases that are flat (at exposure scale) or erosional (Figs. 7, 11B). Flat bases are most prominent when overlying mudstones, while erosional bases are most commonly expressed as amalgamation surfaces (Fig. 9B, C, D). Both bedding-parallel and erosional bases are associated with decimeters of deformation in the underlying beds. This deformation obscures the primary depositional characteristics of the underlying beds, making interpretation of the deformed beds difficult. Bed bases with steeper dips than the bed top are observed at pinch-out of these beds (Fig. 11B), with bed bases steepening towards the pinch-out. Amalgamation surfaces are identified by mudstone-clast-rich rugose horizons, abrupt grain-size breaks, and truncated trace-fossil burrows. Bed tops are flat and exhibit little depositional relief. Beds are typically structureless, with rarely preserved faint parallel lamination and tractional structures, and ungraded, though some beds show weak normal grading (Figs. 7, 9B, C, D).


The presence of erosional bed bases with tool marks and grooves indicates that LF 1A beds were deposited by high-concentration flows that initially carried large clasts and other detritus at the base of the flow. Superimposed flute marks, normal grading, and tractional structures are indicative of evolution to a less concentrated, more turbulent flow. The rarity of tractional structures and commonly massive and poorly sorted beds, however, indicates that these beds were deposited rapidly (sensuLowe 1982). This is most likely due to a reduction in flow capacity (sensuHiscott 1994), preventing the formation of grading and the preservation of bedforms. These turbidites are therefore interpreted as high-density turbidites (sensuLowe 1982). The presence of deformation structures beneath these flows has been attributed to high shear stresses acting on the seabed (e.g., Clark and Stanbrook 2001; Puigdefàbregas et al. 2004). These laterally extensive (up to 1 km where outcrop allows) high-density turbidites, which are commonly amalgamated, are interpreted to be analogous to lobe axis deposits observed elsewhere (e.g., Prélat et al. 2009).

The coarse grain size and thin-bedded nature of LF 1B suggests that these beds were deposited as a coarse-grained lag in a bypass-dominated part of the system. The lenticular geometry suggests that either i) flow energy, and therefore bypass potential, was not homogeneous laterally within the flow or ii) erosion by later flows or waxing of the flow (Kane et al. 2009), scoured the bed top (Fig. 10B). Because this lithofacies is confined stratigraphically to sequences underlying thick-bedded and amalgamated sandstones of similar grain sizes, they are inferred to be laterally related. These lags are interpreted to represent a mostly bypassing equivalent of the thick-bedded sandstones in the lobe axis.

FA 2: Lobe Off-Axis


Facies association 2 is composed primarily of normally graded medium-bedded (0.1–0.8 m) fine- to medium-grained sandstones (LF 2) with flat to slightly erosional bed bases and flat bed tops (Figs. 7, 10A, 12). Flutes and grooves are often seen at bed bases. Ripples at bed tops commonly show opposing paleoflow directions to those measured from flutes and grooves on individual event beds (Fig. 5). Beds pinch out abruptly towards the basin margin, and often can be traced away from the onlap surface to a parent thick-bedded sandstone (Fig. 7). In the uppermost stratigraphy of the basin, LF 2 commonly has highly irregular bed tops with abundant tractional structures, such as climbing ripples and convolute lamination (Figs. 10E, 12). These beds are termed LF 2B.


The presence of flutes, normal grading, and tractional structures indicate that the LF 2 beds were deposited by waning, turbulent flows which were able to rework the aggrading deposit (Bouma 1962). This suggests that these flows were more dilute than the parent flows of the thick-bedded sandstones (LF 1). These medium-bedded sandstones are therefore interpreted as medium-density turbidites and are differentiated from low-density turbidites by bed thicknesses being greater than 10 cm and a coarser grain size. These medium-density turbidites also have a thicker massive interval at their base compared to low-density turbidites. Opposing paleocurrent directions within event beds is characteristic of flows encountering topography (e.g., Kneller et al. 1991), indicating that many of these beds were deposited close to a basin margin. Coarser, denser parts of the flow may be more strongly influenced by topography than the upper more dilute part of the same flow (Bakke et al. 2013).

The finer, better sorted and thinner-bedded nature of this lithofacies compared to thick-bedded sandstones indicates that these beds were deposited beyond the axis of the lobe (off-axis) (Prélat et al. 2009; Bell et al. 2018a) (Fig. 8). LF 2B is interpreted as representing flows that were deposited close to the basin margin or possibly within slump scars on the basin margin. Bypassing flows deflected by the margin or trapped in scars caused complicated oscillatory-flow patterns to develop which deformed the aggrading deposits (Pickering and Hiscott 1985; Tinterri et al. 2016; Cunha et al. 2017). An example of this can be seen at Tête de Ruch, where this facies dominates an ∼ 10-m-thick scar fill (Puigdefàbregas et al. 2004) and is overlain by thin beds showing simple uniformly directed ripples and plane-parallel lamination (Figs. 5, 11C). These thin beds were deposited over the scar fill where flows were relatively unconfined, allowing more uniform and waning flow deposition to dominate.

FA 3: Proximal Fringe


Facies association 3 is dominated by medium- to thick- bedded (0.1–1.2 m) bipartite or tripartite event beds (LF 3) (Figs. 9, 11C, D, G). These beds are composed of a lower division of medium- to thick-bedded sandstone (LF 1 and LF 2) with an irregular top which is loaded and/or eroded into by an overlying argillaceous sandstone (Fig. 10D). The middle division is an argillaceous sandstone, poorly sorted and often deformed, which appears as a sheared fabric (Fig. 10D). The argillaceous sandstone can either contain clasts of the underlying sandstone or be clast-poor. Organic matter (< 70 cm in length) may be present in this bed, with organic-rich sandstones typically thicker than the more argillaceous sandstones. Elongate organic matter is usually aligned with its long axis approximately parallel to paleoflow recorded on the bed base (Fig. 10C).

Argillaceous sandstones, which are rich in cleaner sandstone clasts, commonly occur where the lower, cleaner sandstone is coarser-grained; thus most sandstone clasts seen tend be coarser grained than the argillaceous matrix. This deposit can show variable sand-to-mud ratios. Higher sand contents within this middle division are associated with increased prevalence of lamination, with centimeter-scale layering (mud-concentrated laminae) sometimes evident, and higher mud contents associated with more sheared and poorly sorted deposits. Both of these divisions may contain coarse-sand to granule-size quartz grains supported within the matrix.

Overlying this argillaceous sandstone, medium-bedded, often muddy sandstone may occur, although it is sometimes difficult to assess whether this sandstone is part of the underlying event bed or represents a separate event bed (Fig. 10D). This sandstone has an irregular base and can show loading and foundering into the underlying argillaceous sandstone (Fig. 10D). The bed top is typically flat or mounded. Approaching basin margins these beds can be seen to transition laterally from thick-bedded sandstones (Fig. 13A, B).

The middle division of the bed may pinch out between the underlying and overlying sandstones, which then amalgamate at the onlap surface (Fig. 13C), forming a complicated and often muddy pinch-out (Fig. 14). Stratigraphically this lithofacies dominantly occurs following the thin-bed-dominated sequence and underlying the medium- to thick-bedded sandstones (Fig. 9).


The basal and upper sandstones within these bipartite or tripartite event beds are interpreted as either high-density or medium-density turbidites. The massive, poorly sorted, and mud-rich nature of the division encased by these turbidites is interpreted as being caused by en-masse deposition of a laminar and cohesive flow (e.g., Middleton and Hampton 1973; Mulder and Alexander 2001). This bed division is therefore interpreted as a debrite. The irregular contact seen between these divisions has been attributed to complex short-wavelength soft-sediment deformation and erosion (e.g., Fonnesu et al. 2015). Where the overlying turbidite is relatively thick it is difficult to differentiate between a debritic or deformed (see LF 1 basal deformation) origin for the middle division, particularly when the overlying turbidite has foundered into the underlying division (e.g., Fonnesu et al. 2015). Where the middle sandstone division is slightly cleaner, with lamination and/or layering, the sandstone is interpreted as having been deposited by flows transitional between laminar and turbulent and therefore termed a transitional-flow deposit (Baas et al. 2011; Kane and Pontén 2012). These tripartite event beds therefore contain deposits of both turbulent and laminar flow regimes and are subsequently interpreted as hybrid beds (sensuHaughton et al. 2003, 2009).

It is suggested that the more organic-rich linked debrites, with decimeter clasts of terrestrial debris, are derived from flows which originated from events in the hinterland that carried significant amounts of terrestrial debris into the marine environment (see also Hodgson 2009). These beds are therefore deposits from particularly high-magnitude flows; this may explain their greater average thickness compared with the more argillaceous hybrid-beds. It is also possible that the organic material was staged for significant periods of time on the shelf, making terrestrial debris a poor indicator of flow magnitude. The close association of terrestrial debris and coarse grain sizes in these relatively distal environments, however, indicates that these beds were the result of high-magnitude or “outsize” flows that were capable of significant substrate erosion (Fonnesu et al. 2018). Incorporation and disaggregation of this eroded substrate will have primed these flows for rheological transformation (Kane et al. 2017; Fonnesu et al. 2018). LF 1 can be seen to transform into these hybrid beds over tens of meters approaching the slope, further indicating that these beds were highly erosive and prone to rheological transformation, even in proximal positions, due to forced deceleration against the basin margin (e.g., Patacci et al. 2014). The presence of large amounts of erodible and muddy draping substrate on the basin margin will aid in short-length-scale flow transformation in these high-magnitude flows (Fig. 14) (e.g., Fonnesu et al. 2018).

It is possible that the debritic division of these beds represents high-concentration turbidity currents hitting the counter-slope, causing intra-basinal slope instability and failure. The turbidity current will then have aggraded around this failure, represented depositionally by a sandwiched debrite (e.g., McCaffrey and Kneller 2001). It is difficult to differentiate between a flow-transformation or slope-failure origin for the co-genetic debrite at outcrop; however, the presence of large organic clasts within an identified debritic division may favor a flow-transformation origin.

Because these beds dominantly underlie the thicker-bedded and more sandy turbidites of FA 1 and 2, they are interpreted as being depositionally adjacent (sensuWalther 1894) (Figs. 8, 9). An abundance of hybrid beds indicates a more distal lobe sub-environment compared with the axis and off-axis deposits of FA 1 and FA 2 (e.g., Hodgson 2009; Jackson et al. 2009; Kane et al. 2017; Fonnesu et al. 2018). This sub-environment is termed the lobe fringe. FA 3 is therefore interpreted as being analogous to lobe-fringe deposition seen in unconfined systems (e.g., Spychala et al. 2017). The lobe fringe can be further subdivided into a lateral and frontal fringe, with the lateral fringe having a lower proportion of hybrid beds compared to the frontal fringe (Spychala et al. 2017). In confined settings this definition is complicated because flow deceleration against lateral slopes causes flow transformations and subsequent enrichment of hybrid beds in the lateral fringe (Figs. 8, 14, 16, 17). This relationship is also evident in the Late Cretaceous Britannia Sandstones of the North Sea, where flows underwent rheological transformation against a lateral slope (Lowe and Guy 2000; Barker et al. 2008). This study thus uses the general term proximal fringe because the frontal and lateral fringe are expected to be similar due to the counter-slope causing a prevalence of hybrid beds in both settings (Figs. 8, 14, 17).

It should also be noted that particularly erosive flows in the lobe axis and off-axis also generate hybrid beds at the onlap surface (e.g., Patacci et al. 2014; Fonnesu et al. 2018). This can make lobe-sub-environment interpretation challenging when adjacent to a steep basin margin because much of the succession may be margin-affected over short length scales (tens of meters) (Fig. 14) and therefore not represent their primary lobe-scale paleogeographic position (e.g., Southern et al. 2015). This is enhanced in tectonically active basins where flow types can be highly variable (e.g., Mutti et al. 2009). Facies back-stripping may therefore need to be attempted to assess the lobe-scale sub-environment (Fig. 16). These short-length-scale margin effects and attempts at back-stripping are summarized in Figures 16 and 17.

FA 4: Distal Fringe


Facies association 4 is dominated by thin-bedded (0.01–0.1 m) siltstones to fine-grained sandstones (LF 4) that form laterally continuous event beds (Figs. 7, 11A, 13). Parallel and convolute lamination is common (Fig. 10F). Beds are normally graded, with ripples common on bed tops (Fig. 10F). Ripples may show multiple and opposing paleocurrent directions within single beds. Beds tend to pinch out over tens to hundreds of meters towards the basin margin (Fig. 15), with amalgamation of event beds sometimes observed towards the onlap surface, causing local bed thickness increases in a regional thinning trend. Slumping and folding of beds (LF 5) is evident in this FA, particularly immediately underlying the abrupt transition to dominantly medium- and thick-bedded sandstone sequences (Fig. 10G). Fold-axis measurements indicate that deformation was both oblique and perpendicular to paleo-slopes. Stratigraphically, these beds immediately overlie the Marnes Bleues, forming a distinct sequence that significantly drapes the Marnes Bleues slope (11A). This lithofacies becomes less common up stratigraphy (Fig. 9).


The fine grain-size, abundance of tractional structures, and dominance of thin beds within LF 4 indicates deposition from dilute, low-density turbidity currents (sensuLowe 1982). The narrow grain-size range (dominantly silt) of these slope-draping beds is numerically predicted because of the quadratic decrease in settling velocity of silt and mud and the consequent increase in likelihood of flow inflation far above the initial flow depth in silt- and mud-rich flows (Dorrell et al. 2019). Because these beds are able to drape the existing relief they are suggested to have healed much of the initial relief present on the Marnes Bleues slopes (Figs. 7, 11A). Slumping in these successions (LF 5) is interpreted to represent slope failure on the steep basin margins (Fig. 10G). Progressive deformation and seismicity along the Alpine thrust front is suggested as the primary reason for slope failure (Fig. 2). Failure scars may contribute to some of the heterogeneity seen within the overlying medium- and thick-bedded sandstones by creating a rugose topography high on the slope (Fig. 11C).

The lateral continuity, fine grain-size and thin-bedded character of the thin beds is consistent with both levee deposition and lobe-fringe deposition. Because of the few long-lived channels identified and the low stratigraphic position of these beds in a prograding lobe (Fig. 9), they are not interpreted to be levee deposits, but are instead attributed to the distal fringe of a submarine lobe on the basin floor (e.g., Boulesteix et al. in press, Boulesteix et al. 2019) (Fig. 8), which caused termination of Marnes Bleues carbonate deposition. This is supported by the published paleogeographic position of the lowermost Grès d'Annot (Fig. 2) (Apps et al. 2004; Joseph and Lomas 2004; Salles et al. 2014). The relative lack of hybrid beds in the LF 5 sequence also supports a distal lobe fringe interpretation (Figs. 8, 9A, 13) (Spychala et al. 2017).

Le Ray

Confined basins have previously been associated with sheet-like deposition, where incoming flows are entirely confined by the basin margins, resulting in tabular stratigraphy with little or no autogenic compensational stacking (e.g., Sinclair and Tomasso 2002). Numerical studies (Dorrell et al. 2018), subsurface studies (e.g., Beaubouef and Friedmann 2000), and outcrop studies (e.g., Spychala et al. 2016; Bell et al. 2018b; Liu et al. 2018), however, have shown the stacking-pattern complexity that may arise in basins that display variable degrees of confinement. This study uses data from the lowermost member of the Grès d'Annot, the Le Ray member (Figs. 5, 6), to build on these studies. Le Ray (Puigdefàbregas et al. 2004), or member A (Du Fornel et al. 2004), was fully confined by the basin margins during its deposition (Callec 2004; Salles et al. 2014).

Unit 1.—

This unit comprises a sequence of thin-bedded low-density turbidites (LF 4) overlain abruptly by medium- and thick-bedded high-density turbidites (LF 1 and 2). These thick-bedded sandstones correlate with a thick sequence of thin-bedded turbidites ∼ 800 m to the NW adjacent to the Braux Fault footwall (Figs 3, 6). This transition is caused either by flow over-spill across the paleobathymetry of the Braux Fault or draping of the lower-density part of the flow up the Braux Fault topography. Correlation of this unit toward the NE is hindered by lack of exposure; however, it is possible that the distal Unit 1 correlates with some of the early Unit 2B deposition (Fig. 6).

Unit 2.—

Proximal Unit 2 is characterized by a thick 33 m amalgamated sandstone body interpreted as being deposited by high-concentration turbidity currents. Indicators of erosion, such as scours and decimeter-scale mud-clasts, indicate that this was a region of significant bypass (e.g., Stevenson et al. 2015) and is interpreted to represent the lobe axis (FA 1) through Le Ray deposition. Down-dip this unit has been subdivided into Unit 2A and 2B based on facies and stacking. Unit 2B is composed of thin-bedded turbidites that drape the frontal confinement of the early Annot Basin and stack aggradationally (Fig. 11A). This confinement was caused by a combination of the basin closure due to the NW–SE Melina anticline, the E–W Fuguret anticline, and the NE–SW Braux Fault (Fig. 3). Unit 2A represents an abrupt transition into thicker-bedded sandstones that pinch out against the underlying Unit 2B. These sandstones can be subdivided into various lobes based on the correlation of intervening packages comprising low-density turbidites (LF 4) overlain by hybrid beds (LF 3), medium-density turbidites (LF 2), and amalgamated high-density turbidites (LF 1), which are interpreted to represent the sequential stacking of distal-fringe (FA 4), proximal-fringe (FA 3), off-axis (FA 2), and axis deposits (FA 1). These lobes represent the depositional products of the bypassing flows from the proximal thick sandstone body, forming the lobe axis.

The axis of each of these lobes steps farther into the basin, suggesting allogenic progradation through increasing sediment flux from the uplifting Corsica–Sardinia hinterland (e.g., Apps et al. 1987). This pattern is not uniform, however, and a degree of compensational stacking is clearly visible, with the axes of successive lobes (represented by amalgamated high-density turbidites) overlying the fringes of underlying lobes, which represented lows on the seafloor (e.g. Deptuck et al. 2008; Prélat et al. 2009). These stacked lobes could also represent lobe “fingers” that were focused between the lows of the previous axial deposits (e.g., Groenenberg et al. 2010). In this case the apparent basinward stepping of the Le Ray lobes may be autogenically-driven, as flows are focused between the lows and build passively into the basin. This explanation may operate in tandem with allogenic progradation, particularly during early deposition, due to the prevalence of these finger-like geometries in the basal stratigraphy of lobe complexes (e.g., Prélat et al. 2009; Groenenberg et al. 2010). These indicators of compensational stacking are at odds with the prevailing suggestion that basins generically described as “confined” can be assumed to have a sheet-like architecture, and fit with the recent work indicating that lobes deposited within basins with varying degrees of confinement are characterized by more complicated stratigraphic relationships (e.g., Marini et al. 2015; Bell et al. 2018b; Liu et al. 2018).

The presence of thick and coarse sandstones in distal positions has also been described in subsurface datasets of submarine lobes and lobe elements and has been attributed to flow stripping of the upper dilute parts of flows over the basin's confining topography, leaving behind the coarser parts of the flow (e.g., Marini et al. 2016; Jobe et al. 2017). This process may also contribute to the preservation of the thick and abruptly terminating sandstones within Unit 2A of the Le Ray member.

Unit 3.—

This unit has been differentiated because it cannot be reliably correlated to another unit. It is possible that Unit 3 represents the distal expression of the upper parts of Unit 2, however the lack of exposure between the proximal and distal parts of the upper Le Ray prevents reliable correlation.

Alternative Interpretations.—

The observed intra-formational onlap against the low-density fringe (Unit 2B) at the Col du Fa locality is likely caused by the inferred proximity of the basin margin. An alternative explanation for this stratigraphic relationship may be the presence of a large erosional feature, such as a scour, at the Col du Fa margin. Such scours are interpreted higher in the basin stratigraphy at Tête du Ruch (e.g., Puigdefàbregas et al. 2004), and show the same intra-formational onlap relationship (Fig. 11C). Decameter-thick scour fills are also suggested to be present in the Grès d'Annot at Peira Cava (Lee et al. 2004). The scour would have created accommodation space to be filled by the incoming flows, resulting in the onlap geometries described by this study. Exposure limits further analysis of this problem, however, a scour interpretation is not believed to affect the underlying principles of this study as the low-density turbidites still drape the scour and are onlapped by the later higher-density turbidites.

Another explanation for the observed intra-formational onlap may be that the low-density fringe (Unit 2B) was tilted (e.g., Salles et al. 2014) and subsequently onlapped by higher-density flows when the basin was relatively static or depositional rates were higher, resulting in the wedge-shaped geometry observed (Apps et al. 2004). Unit 2B may therefore represent the distal extents of early Le Ray (i.e., Unit 1), while the onlapping Unit 2A is either more proximal and late-stage Le Ray or early La Barre (i.e., Unit 3) (Figs. 3, 5). Differential compaction between the basin center and the basin margin may have also acted to enhance the tilting effect (Sinclair 1994). The inability to walk out individual units between outcrop localities again makes further analysis of this problem difficult. If tectonism or differential compaction is the reason for the observed relationship, then run-up of the low-density turbidites that characterize Unit 2B would have acted to exaggerate the intra-formational onlap that was caused by tilting, and the underlying principles of this study are maintained. Unit 1 would therefore be analogous to the ponded aprons identified in intra-slope basins of the Gulf of Mexico, while Unit 2 would be similar to a perched apron (Prather et al. 2017).

Turbidity-Current Run-Up and Onlap Geometry

Sediment gravity flows are able to deposit high on confining slopes through flow run-up and/or inflation (e.g., Muck and Underwood 1990; Lamb et al. 2008; Dorrell et al. 2019). The distance a turbidity current runs up topography is termed the run-up height (H) and, in its simplest form (following Straub et al. 2008), can be represented by,
where h = flow thickness, U = flow velocity, forumla = bulk density of the flow (composed of quartz at 2650 kg m–3), forumla = density of the ambient water (seawater at 1020 kg m–3), and g = acceleration due to gravity (9.81 m s–1). In order to assign a single value for sediment concentration there is an assumption that there is no density stratification within the flow. A 30-m-deep channel-form at Chambre du Roi (Sinclair 2000) has been used as the basis for estimating minimum flow height. This is slightly arbitrary but serves the purpose of this thought-experiment. In order to correct for channel-related superelevation, this flow height has been multiplied by 1.3 (see Mohrig and Buttles 2007), giving a h value of 39 m. This height represents the minimum height of the flow at the lobe apex, disregards compaction, and is assumed constant. It should be noted that due to a lack of flow-height proxies preserved at outcrop, this value is used purely to demonstrate the underlying principles of the model, i.e., the model does not attempt to fully reconstruct the turbidity currents entering the Annot Basin. The depth-averaged flow velocities used range from 5 m s–1 to 0.5 m s–1. The upper limit of 5 m s–1 is derived from the maximum flow velocities (4–6 m s–1) reached by sand-rich flows in the Monterey Canyon (Symons et al. 2017), while the lower limit of 0.5 m s–1 is derived from average measurements of finer-grained flows in the Congo Canyon (Azpiroz-Zabala et al. 2017). The Annot system is sand-rich; therefore the faster Monterey flows are likely the most analogous, at least close to the input. Flow velocity will decrease with decreasing concentration, making the use of a constant velocity problematic. This is suggested to be less important for small and confined basins, such as the Annot Basin, where flows may be prevented from significant velocity decay between the axis and pinch-out. If velocities do decay significantly, then plotting single velocity values through the height of an individual flow will be required; for example, the 5 ms–1 and 10% concentration basal part of a hypothetical flow will run-up 9.2 m above the flow height, while its 1 ms–1 and 0.1% concentration tail will run-up 32 m above the flow height. This decameter scale difference may explain the ∼ 15 m of slope drape at Col du Fa (Figs. 6, 8). This parameter is further complicated by the possible presence of low-velocity dense basal layers within highly-concentrated stratified flows (e.g., Stevenson et al. 2018). Turbulence is suppressed in these basal layers; this reduces velocity and run-up heights, resulting in the increased likelihood of an abrupt pinch-out of the sand-rich basal layer and bypass of the upper and low-concentration parts of the flow. The wedged high-density turbidites identified at onlap by this study (Fig. 11B) may be the depositional products of such basal layers.

By varying flow concentration in equation (1), a clear trend is developed, with lower sediment concentrations resulting in greater run-up heights (Fig. 18A). In prograding lobe systems turbidity currents with lower sediment concentrations (well below 10%), forming low- or medium-density turbidites, are typically found in the distal or basal stratigraphy and flows with higher sediment concentrations (> 10%), forming high-density turbidites, are typically found in the proximal or upper stratigraphy (e.g., Hodgson et al. 2006, 2009; this study). This therefore suggests that earliest turbulent flows into a receiving basin should have the greatest run-up heights, assuming all of these flows are of similar thickness, with run-up heights decreasing through time as flows become more concentrated (Fig. 18A). It should be noted that the effect of increasing concentration through time will be counteracted by increasing velocities, as discussed previously. Suspended-sediment concentrations of > ∼ 8% have been shown to suppress the generation of current ripples and cause transformation from turbulent to transitional flow, where flows decelerate sufficiently, forming transitional or hybrid-flow deposits (Baas et al. 2011). This concentration ( > ∼ 8%) has therefore informed the placement of hybrid beds along the x axis of the run-up height trend (Fig. 18A), giving hybrid beds a similar run-up potential as high-density turbidites. It should be noted that Baas et al. (2011) emphasize that this value is dependent on flow velocity, grain size, and sediment composition, and that the dimensionless Reynolds number is a much better predictor of flow phase. The results of this analysis fit with facies-dependant thinning rates compiled from 18 outcrop studies by Tőkés and Patacci (2018), with hybrid-beds having 1.3 to 2.8 times higher thinning rates than turbidites. Data collected in this study also support the compiled data of Tőkés and Patacci (2018), with thinner-bedded turbidites having lower thinning rates than thicker-bedded turbidites and consequently draping topography (Figs. 6, 7, 14). Similar thinning trends have been reported from levee sandstones (DeVries and Lindholm 1994).

Lateral vs. Frontal Onlap

To assess the run-up variation between frontal and lateral onlap the velocity of the modeled flow (at 5% concentration) was varied according to the incidence angle of the flow with the slope. It is assumed that the flow velocity will be at its maximum (5 m s–1 in this case) in the principal direction of travel, or an angle of incidence of 90° with respect to the slope, and that the flow velocity will fall to 0 ms–1 perpendicular to its axis at an angle of incidence of 0°, i.e., when the flow is running perpendicular to the topography. In reality there will still be some lateral velocity; however, for the purposes of this simple analysis it is assumed that this is negligible. The fastest flows occur at an angle of incidence of 0° or perpendicular to the slope (Fig. 18B). These flows therefore run farther up the counter-slope. Deposits of these flows would pinch out higher up the frontal slope than the lateral slope (Fig. 18B, C). In the 5 ms–1 case, for example, the difference in run-up height between the frontal and lateral part of the flow is 17 m. The difference will be increased in lower-density flows and reduced in slower flows.

Stratigraphic Evolution of Onlap

Based on the similarity between facies transitions and associations seen in this confined basin and those in unconfined or weakly confined submarine lobes (Figs. 8, 9) (e.g., Hodgson 2009; Spychala et al. 2016, 2017), and the onlap termination styles shown by this study to be produced by the parent flows of these facies (Figs. 16, 17), a predictable stratigraphic evolution of onlap at confined basin margins is proposed (Fig. 19).

Distal Fringe.—

Initially, onlap terminations will be characterized by draping of the slope as low-density turbidites of the lobe fringe are deposited (Figs. 17, 19). This low-density turbidite drape is likely to be composed predominantly of silt or mud, because fine-grained flows are much more capable of flow inflation and deposition high on the slope (Dorrell et al. 2018). It is also suggested that much of the poorly-exposed hemipelagic sediment in deep-marine basins is composed of millimeter-scale and centimeter-scale event beds (Fig. 10F) (Boulesteix et al. in press, Boulesteix et al. 2019), and consequently represents the distal lobe fringe. Lobe-fringe deposition is therefore likely to be more widespread than previously appreciated (Boulesteix et al. 2019b; Spychala et al. 2019) and results in the healing of substantial amounts of basinal topography, forming a dominantly aggradational sequence of thin beds on the basin margin (Figs. 11A, 17, 19).

Proximal Fringe.—

Hybrid beds and low-density turbidites of the proximal lobe fringe are then deposited into the basin as the system progrades (Fig. 19). These hybrid flows are more concentrated and will therefore have lower flow efficiencies when they encounter the slope, so will be unable to deposit as far up the slope as the underlying low-density turbidites of the distal lobe fringe (Figs. 17, 18A). This will cause abrupt intra-formational onlap of these higher-concentration flows against the underlying lobe-fringe deposits (Fig. 19). In the Annot Basin this is represented by proximal-fringe deposits wedging out against the underlying distal-fringe deposits (Figs. 11A, 14). The abruptness and complex 3D geometry of these terminations is enhanced by the combined potential for hybrid-bed development through long-run-out cohesive flow transformation (e.g., Haughton et al. 2009), slope-induced flow transformation (e.g., Barker et al. 2008; Patacci et al. 2014; Bell et al. 2018b), and flow-induced slope failures (e.g., McCaffrey and Kneller 2001) (Fig. 16). This depositional pattern will be seen in cross sections as a progressive migration of termination points towards the basin center (offlap) (Fig. 19) or a reduction in distance between successive onlapping termination points towards the basin margin. Sylvester et al. (2015) generated similar onlap trends using a geometric approach with subsidence and sediment supply as the variables.


As progradation continues, these hybrid-bed-prone proximal-fringe deposits will become overlain by deposits of more proximal flows which have not decelerated to the same degree and hence are more turbulent and of lower concentration, but sand-rich (Figs. 17, 19). The off-axis deposits will be able to drape the slope more effectively than the underlying hybrid beds owing to their lower sediment concentrations (Fig. 19). This will result in progressive termination of their deposits higher up on the slope and either intra-formational onlap against the now thinner veneer of the low-density fringe or onlap directly against the hemipelagic basin margin (Figs. 17, 19, 20, 21).


As higher-concentration flows begin to dominate, intra-formational onlap may occur against the underlying fringe or off-axis deposits that were able to run up the hemipelagic slope (Fig. 19). The highly concentrated basal layers of these flows will be preserved as abruptly onlapping high-density turbidites, with the low-density tail of the flow bypassed down-dip. These axial flows will also be more erosive and able to incorporate mud-rich substrate, resulting in an increasing likelihood of intra-formational onlap through short-length-scale rheological flow transformation and consequent higher thinning rates adjacent to the basin margin (Fig. 14). The scours formed by these erosive events close to onlap will promote further autocyclic modulation of stacking patterns adjacent to the basin margin (e.g., Eggenhuisen et al. 2011). This relationship will also be exaggerated in coarser-grained systems because higher concentration flows will be less able to deposit farther up the slope, particularly at lateral margins if these high-concentration flows are narrower (Al-Jaidi et al. 2004).

A critical point in the basin fill will then be reached when these deposits heal the confining topography sufficiently to allow bypass of the coarser components of flows over the confining slope, forming a stepped, instead of a ponded, basin (Prather et al. 2017). Until this level is reached the finer-grained parts of the flows were stripped, thus increasing the sand proportion in the up-dip basin (Prather et al. 2017). This transition is represented toward the sill of the Annot Basin, where proximal lobe-axis deposits bypass down-dip to the Grand Coyer minibasin (Fig. 2), where this stratigraphic evolution repeats in the next confined depocenter.

In the case that the topography is sufficient so as not to be healed by the underlying deposits, the lobe-axis deposits will continue to onlap against the underlying deposits until the accommodation is healed sufficiently to allow the axial deposits to onlap against the hemipelagic basin margin or completely fill the basinal relief and behave as essentially unconfined deposits, resulting in downlapping terminations (Fig. 19). Apparent unconfinement in a vertical section may also occur if allogenic progradation (e.g., increasing sediment flux) does not keep pace with topographic healing. This will result in the axial sandstones rarely onlapping directly against the constantly retreating basin margin and instead depositing away from the slope and on the basin floor (e.g., Kneller and McCaffrey 1999).

Applicability and Limitations

Syn-Depositional Deformation.—

This general stratigraphic evolution applies most readily to static basins with little syn-depositional deformation, where onlap trends are not modified by changes in subsidence or depocenter migration (Figs. 19, 21). The Annot Basin was subject to syn-depositional movement of the depocenter (e.g., Salles et al. 2014); however stratigraphic units of the basin fill show the same conceptual evolution as hypothesized for a static basin (Fig. 6), with the only difference being that the onlap is exaggerated on the tilted eastern exposures because of the increasing slope angle during deposition (Fig. 11A) (Salles et al. 2014). In salt-deformed basins, where syn-depositional subsidence can be rapid, this onlap evolution will be also be exaggerated. The general model for onlap evolution (Fig. 19) may therefore broadly apply to actively deforming basins (Fig. 21). It may be difficult, however, to differentiate between allogenic onlap patterns caused by syn-depositional subsidence or sediment flux (e.g., Sylvester et al. 2015) and those caused by autogenic variations in flow properties (Fig. 21). Care therefore needs to be taken when reconstructing tectonic histories from onlap trends alone, especially in datasets without lithological control.

Onlap Incidence Angle.—

Frontal onlap causes greater turbidity current run-up than lateral onlap (Fig. 18B, C). This in turn makes the model presented here more applicable for dip sections against a confining slope. This is further reinforced by the Le Ray dip-oblique correlation (Fig. 6), which is interpreted to record the stratigraphic evolution of onlap described. The lack of a significant thin-bedded slope drape at the lateral onlap of Grès d'Annot at Chalufy (e.g., Puigdefàbregas et al. 2004; Smith and Joseph 2004) may be an example of the importance of incidence angle, with run-up decreased against the lateral slope (Fig. 18B, C).

A minor thin-bedded slope drape does exist at Chalufy, however, and intra-formational onlap does occur against these thin beds (e.g., Bakke et al. 2013). This relationship may occur against lateral margins because flows become increasingly elongate with increasing sediment concentrations (Al-Jaidi et al. 2004). In strike sections this will result in a similar onlap termination pattern, with thicker and lower-concentration flows of the lobe fringe depositing higher on the lateral slope than higher-concentration flows of the lobe axis, which are elongated in the dip direction and more prone to bypass down-dip. This relationship may therefore apply to both strike and dip exposures; however, greater understanding of frontal vs. lateral onlap in exposed or subsurface deep-water systems is required for further analysis.

The lack of a substantial thin-bedded drape at Chalufy may also be caused by its relatively distal position compared to the Annot Basin (Fig. 2), resulting in lower-velocity and lower-concentration flows being unable to maintain enough energy during their longer passage down the slope to deposit significant thicknesses of sediment high on the confining slopes (Fig. 20).

Hierarchical Scale.—

The onlap patterns observed in this study are seen mainly on the scale of tens of meters, similar to typical lobe or possibly lobe-complex dimensions (Prélat et al. 2009). Above the spatio-temporal scale of lobe complexes the morphology of the depositional element is less likely to be the result of autogenic processes acting on individual events and more likely to be controlled by allogenic factors, such as the interplay of basin subsidence and sediment supply (see stratigraphic interval scale of Sheets et al. 2002; Jobe et al. 2017) (Fig. 21). It should be noted that in confined basins established morphometric ranges for unconfined depositional element thicknesses break down due to variable degrees of confinement across systems (Prélat et al. 2010). The lobe-scale applicability of the model is therefore purely a hierarchical observation because lobe thicknesses will vary across systems.

Stratigraphic Position.—

Lobe progradation is unlikely to be constant; fluctuations in sediment supply, relative sea level, or changes in routing may result in small-scale backstepping. This backstepping creates discrete boundaries between successive members in the Annot Basin (Callec 2004; Euzen et al. 2004) (Figs. 3, 5). These fluctuations will result in a non-uniform evolution of flow concentration during progradation, however on the basin-scale the overall trend of increasing flow concentration and subsequent onlap style will be maintained, with higher-density flows becoming more prevalent in the basin through progradation. These observations make the model most applicable to the early fill of confined deep-water systems, where confinement is greatest and distal thin-bedded turbidites will be most prevalent.

Low-density and thin-bedded turbidites also tend to form thinner successions than high-density and thick-bedded turbidites. This is because sediment volume and sedimentation rates are greater in higher-density flows and because differential compaction more heavily affects finer-grained successions. This has implications for the evolution of onlap because the distal lobe fringe will be thinner than the lobe axis. As a result, the intra-formational onlap against the underlying fringe will be much more difficult to detect through time as the draping low-density turbidites gradually become thinner as the basin fills, eventually resulting in the higher-density flows onlapping directly against the hemipelagic basin margin (Fig. 17). This relationship is seen in the Le Ray member (Fig. 6), where the thin-bedded fringe is gradually surmounted by later flows.

Other Variables.—

This conceptual model aims to describe in simple terms the effect that autogenic flow evolution may have on onlap patterns in confined basins. Variables such as a waxing–waning sediment supply, hemipelagic aggradation rates, and the dominant grain size of the system are not discussed fully; however they will act to alter the autogenic processes that affect onlap and should be explored in future studies.

Understanding flow interaction with, and bed termination against, confining topography is critical for reconstructing the structural and sedimentological evolution of deep-water basins. This study presents a review of onlap styles in deep-water settings based on detailed field investigations and compares these results against those from a simple numerical model. Onlap terminations are shown to evolve in a predictable way through the progradation of a submarine lobe succession, with different lobe sub-environments identified at the basin margin through the migration of successive termination points and facies trends.

Initially termination points migrate towards the basin margins as low-density turbidites significantly drape the inherited basinal topography. Progressively higher-magnitude flows with greater sediment concentrations of the hybrid-bed-rich proximal fan fringe onlap these underlying deposits, causing the development of an intra-formational onlap surface that is characterized by either a basinward shift in termination points or a reduced distance between successive termination points towards the basin margin. Hybrid beds are also shown to constitute significant thicknesses of the proximal fringe in confined systems through long run-out transformation, slope-induced transformation, and intra-basinal slope instability. Progradation of the lobe off-axis over the proximal fringe will cause further intra-formational onlap as the lower-concentration off-axis deposits drape the slope. Onlap against the fringe drape will continue until it is surmounted and onlap occurs directly against the hemipelagic basin margin. Intra-formational onlap may also occur in the lobe axis through abrupt onlap of these high-concentration deposits against the underlying lower-concentration fringe or off-axis deposits.

These sedimentological processes act to modulate the allogenic signals present in the Grès d'Annot of the Annot Basin, with onlap patterns controlled by both allogenic processes, through tectonic deformation and increasing sediment supply, and autogenic processes, through the interaction between longitudinal flow evolution and confining basin margins. Depositional hierarchy is also shown to be important when interpreting onlap patterns, with autogenic processes more important at the lobe scale and allogenic processes more important at the lobe-complex scale.

Soutter is funded by Natural Environment Research Council (NERC) grant number NE/M00578X/1. Constructive reviews by B. Prather, Y. Spychala, and M. Mayall helped clarify both the content and presentation of the manuscript, and are gratefully acknowledged. Associate Editor S. Hubbard and Corresponding Editor J. Southard are thanked for editorial handling of the manuscript and additional comments. Stereonet 10 is acknowledged for data plotting.

This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (, which permits unrestricted use, distribution and reproduction in any medium, provided the original work is appropriately credited.