The Sombrero Escarpment (Figures 5A, 6) is oriented E-W with an average slope gradient of 14.7° and extends from 900 mbsl down to 6,000 mbsl. This escarpment is incised by numerous canyons (15–20 km long, up to 38 km for the longest, Figure 6), mostly linear and dipping northwards (Figure 6A) toward the Sombrero Basin. The shortest canyons present a 30 m-deep thalweg while the longest is up to 600 m-deep. For instance, canyons 1, 2, and 3 are representative of the erosional patterns on the Sombrero escarpment: Canyon 1 is a single and sinuous canyon presenting a continuous slope angle (∼14,7°, Figure 6C). Canyon 2 is linear and merges with two northward-oriented tributary canyons between 2000 mbsl and 4,100 mbsl (Figure 6C). Canyon 3 is more sinuous and the longest, (38 km), merging with three S-N linear canyons between 5,000 mbsl and 6,000 mbsl. In their upper part, most canyons are slope-confined and show no connection with feeding channels from the Anguilla Depression (Figure 5). The canyons are highlighted by an alternation of highly reflective echo-facies [a] and [b], contrasting with less reflective facies [c] and [d] (Figure 4B, Table 1) in between. Given the high seafloor reflectivity within the canyons, we interpreted them as major sediment transfer paths as observed in similar contexts (Counts et al., 2018). The amount of sediment transfer is however probably limited given the disconnection with the upper slope.

The northeastern slope of the Anguilla Bank, which extends from 100 mbsl down to 5,000 mbsl, is oriented NE-SW and shows a mean gradient value of 7.7° (Figure 5, Table 1). As observed along the Sombrero Escarpment, the northeastern slope of the Anguilla bank is incised by several canyons highlighted by reflective echo-facies [a] and [b] (Figure 4B, Table 1). Its northern tip, the Anguilla Spur (Figures 5A, 6A), includes two sectors: a steep one (11.8°) from 2000 mbsl down to 5,000 mbsl, and an extended, wide sector (∼550 km², 2.3°) from 5,000 mbsl to 6,000 mbsl.

In the Anguilla spur, seismic profiles reveal almost directly in the subsurface the acoustic basement (facies [α] in Figure 7A), with a limited number of discontinuous reflectors near the surface. Locally, a chaotic unit lies on top (enlargement in Figure 7A). We interpreted it as Mass Transport Deposit (MTD) (Weimer and Shipp, 2004), which attests to the occurrence of local destabilizations within the spur.

The Tintamarre Spur (Figures 5A, 8) is a SW-NE trending structure located along with the eastward extension of the Anguilla Bank. The spur is slightly asymmetric, revealing a steeper mean gradient on its northwestern flank (12°) compared to the southeastern one (9.7°), most likely related to the presence of normal faults on the western flank (Feuillet et al., 2002; and Figure 7A). The top of the spur reveals remarkable morphologies: two particularly planar and gently dipping surfaces (respectively ∼2° and 4.5°) (Figure 8, Table 1), with an area of 37 and ∼85 km², occur at ∼700 mbsl and at ∼1,600 mbsl. These features stand out from all other morphologies visible on the insular slopes, and we interpreted them as tilted strata. Given their location on the upper slope, at the edge of Anguilla Bank, these could be interpreted as subsided platforms (such structures have already been identified close to carbonate banks, Eberli and Ginsburg, 1987); we have named them paleo-platform PP1 and paleo-platform PP2 (Figure 8A).

On the northern part of the Tintamarre Spur, from 1,000 mbsl to 4,000 mbsl, numerous 10 to 18 km-long, northward-oriented canyons and/or gullies pour out into the Anguilla Valley (Figure 8A). These canyons/gullies are highlighted by the reflective echo-facies [a] and [b] while the rest of the spur is less reflective (echo-facies [c] to [e], Figure 4B, Table 1). Most canyons are slope-confined, with a spoon-shaped canyon head located at the edge of paleo-platform PP2 (Figure 8A). Similar headscarps also incise the edge of the upper paleo-platform PP1, with some gullies at its foot, streaming down to PP2. For instance, Canyons 4 and 5 incise both paleoplatforms via gullies in-between, while Canyon 6 is limited to PP2 (Figure 8B).

In the northeastern part of the Tintamarre Spur, from 4,000 to 6,000 mbsl, no canyon is identified. On the southeastern flank of the spur, only a few gullies and sinuous canyons occur (for instance Canyon 6 (Line 6, Figure 8B). Its slope gradient is locally steep between 1700 and 3,800 mbsl (14.5°), and decreases from 4,100 mbsl to 4,600 mbsl (11.9°). The canyon pours out in a relatively small slope basin located at 4,680 mbsl, (15 km², Table 1) called Tintamarre Spur Basin (Figures 7B, 8). Its reflectivity is very low (echo-facies [e], Figure 4B, Table 1). It is highlighted by the seismic facies [β] on seismic profiles (Figure 7B). This basin is confined by the height of its flanks.

The southeastern slope of the Anguilla Bank (oriented NW-SE, Figures 3, 5) is not imaged by multibeam data, but based on GEBCO Satellite bathymetry the shape of the 1000-m isobaths and some spoon-shaped (Canyons heads, Figure 5A) in the Anguilla Bank, which we interpreted as canyon heads, suggest the existence of SW-NE oriented canyons reaching the St-Barthélémy Valley.

The northern slope of the Antigua Bank is composed of a major SW- NE trending spur called Barbuda Spur, cut by the 1000 m-deep Barbuda Valley (Figures 5A, 9). The Barbuda Spur shows generally a low to very low reflective seafloor (echo-facies [d] and [e], Figure 4B, Table 1). With an average slope gradient of 5.7° from 120 mbsl to 5,000 mbsl, the SW-NE northern slope of Antigua Bank is also incised by northward oriented canyons, over 30 km long, 3 km wide, and ∼500 m deep. The canyons, highlighted by echo-facies [a] and [b] (Figure 4B, Table 1), seem either slope-confined or connected at their head by gullies themselves connected upwards to the Antigua Bank (Figure 9A). These gullies are located downstream from passes across the reefs of Barbuda (yellow dashed arrows, towards gullies, Figure 9E). Four main canyons are visible between 4,000 mbsl and 5,000 mbsl (light pink cross-section, Figures 9A,B) and present a “U-shape”. They result from the merger of other canyons between 1,500 mbsl and 3,500 mbsl. The 50-km-long canyon shown in line 7 (yellow line, Figures 9B,D) presents 400-m-high walls while the three other canyons have smaller walls, between 50 and 300 m (Figure 9B).

The Barbuda Valley has a depth of about 4,600 mbsl and is connected to the Antigua Bank through the main canyon (Canyon 8, Figure 9A) collecting sediment from several secondary canyons (Figure 9D). Between 2000 mbsl and 4,000 mbsl, where the secondary canyons spatially merge with the main one, the slope is steep (∼7.4°) then flattens (∼2.2°) below 4,000 mbsl (Figure 9D). The Barbuda Valley (Figures 5A, 9) is highlighted by highly reflective echo-facies [a] and [b] (Figure 4B, Table 1) and is partially filled with stratified deposits illustrated by the seismic facies [β] on seismic profiles (Figure 7B). Additionally, the valley is bounded by a normal fault on its eastern flank (Figure 7B).

With an average slope gradient of 6.4°, the NNW-SSE oriented Antigua northeastern slope is incised by numerous and very reflective (echo-facies [a], Figure 4B, Table 1) SW-NE oriented linear canyons which are also linked to gullies at the top of the slope (Figure 9). With a length of at least 20 km, these narrow (<1 km width) canyons present 80-m-high walls between 2000 mbsl and 3,000 mbsl (orange cross-section, Figure 9C) and 10-to-100-m-high walls between 3,000 mbsl and 4,000 mbsl (brown cross-section, Figure 9C). They pour out onto a 4,000 mbsl deep terrace (Figure 9). Canyon 9 is linear and its thalweg presents an average slope of 10.1° between 1,000 mbsl and 4,000 mbsl.

To summarize, compared to the eastern flank, the north-western flank of the spur is dominated by a greater number of canyons, which moreover appear more incised and more reflective, probably supporting a more active sediment transfer as observed in other studies (Iacono et al., 2008; Counts et al., 2018). The origins of these differences will be assessed in the discussion.

The Sombrero Basin, located at the toe of the Sombrero Escarpment, is 8 km wide and 86 km long, lying at ∼6,200 mbsl (Figure 6; Table 1). This structure is bounded by transtensive faults (Laurencin et al., 2017). The seafloor is flat, and poorly reflective, highlighted by echo-facies [d] (Figure 4B). However, in front of the numerous canyons, the seafloor is slightly more reflective but still represented by the same echo-facies. We interpreted the infill as being either fine-grained (as observed by Iacono et al., 2008), and/or the sediment transfers being inactive, apart from in front of the canyons where accumulation is limited.

Two valleys are identified between the spurs: the Anguilla Valley (Figures 5A, 10), delimited by the Tintamarre and Anguilla Spurs, and the St-Barthélémy Valley (Figures 5A, 10), delimited to the SE by the Barbuda Spur. These two valleys, filled with stratified deposits illustrated by the seismic facies [β] (Figures 7A,B), are bounded by normal faults (Feuillet et al., 2002). Their seafloor however presents a contrasting backscatter signature (Figure 10A).

The Anguilla Valley, from 5,100 mbsl to 5,350 mbsl, (Figures 5A, 10), is narrow (∼10 km). In its distal part (Figure 10B), the seafloor is poorly reflective along the Tintamarre spur, while the axis of the valley and its distal vicinity is highly reflective (Figure 10B). This highly reflective seafloor correlates with truncated reflectors (Figure 10C). We interpreted these features as the presence of an active sediment transfer and erosional turbidity currents (active currents Figure 10C), as attested by the backscatter signature of turbidites in other studies (Iacono et al., 2008). Five units separated by four unconformities are visible in this valley (Figures 10C,D). U1 to U3 present continuous reflectors; U4 is more chaotic and onlaps in its southwestern part while U5 appears very chaotic, with few strong and parallel reflectors. U1 and U4 share a similarity: an erosive depression is visible in U1 at the foot of the slope, and a similar buried feature is identified in U4 (Figure 10D).

Conversely, the St-Barthélémy Valley (Figures 5A, 7–10) is a broader submarine valley (∼28 km wide, 4,900–5,200 mbsl) presenting a low reflective seafloor (Figure 10A, echo-facies [d], Figure 4B; Table 1) compared to the Antigua valley. At the foot of the southeastern slopes off the Anguilla Bank, the seafloor is locally more reflective (echo-facies [a] to [c], Figure 4B, Table 1) describing a lobate shape (Lobate Structures, Figure 10). The infill of the valley is well stratified and shows at least 4 units (Ua to Ud) separated by three unconformities (A to C in Figure 7B). The most recent infill shows well-stratified reflections (Figure 7B), lying unconformably on top of truncated reflectors. We interpreted the sediment transfer within the valley as being less active and/or comprising finer-grained deposits than within the Anguilla valley (as observed in other studies (Iacono et al., 2008; Counts et al., 2018).

The presence of unconformities in both valleys suggests successive periods of sediment transfer (as attested by other studies, e.g. Laban et al., 1984) but also at least one subsidence phase, attested by the divergent geometry of the reflectors of U3, U4, and U5 in the Anguilla Valley. Drastic subsidence in the area has already been identified on MCS profiles (Boucard et al., 2021), suggesting the importance of tectonics in the Northern Lesser Antilles Segment.

In the St-Barthélémy Valley, erosional depressions (Figures 7B, 11) occur along the 5,000 m isobaths and are located at the foot of the northern slope of the Antigua Bank (Figure 11A). They are 540–1,200 m long and 16–20 m deep, where the slope break ranges from 4° to 7° (Figures 11B–E). In detail, sub-bottom profiles reveal a small, highly reflective depression located at the foot of the slope, with a mounded shape located immediately basinward from the depression (Figure 11B). The mound shows stratified acoustic facies with low amplitude continuous reflectors. A similar structure, visible in the Anguilla Valley (Figure 10), is 1,450 m long and 22 m deep, at a break in slope of ∼7°and corresponds to the erosive depression associated with U1 (Figure 10D). Due to their stratified composition with low amplitude continuous reflectors, corresponding to alternating fine- and coarse-grained material (e.g., Hassoun et al., 2009), we inferred that these features have a sedimentary origin. We interpreted these features as plunge pools, characterized by an erosional scour at the foot of a steep slope, and a depositional slope-break deposit downstream (Komar, 1971; Mulder and Alexander, 2001; Lee et al., 2002; Bourget et al., 2011; Migeon et al., 2012). Given their similarities, we interpreted the erosional and depositional features described at the slope break in Units U1 and U4 (Figure 10D) as active and buried plunge pools, respectively.

The northeastern slope of Antigua Bank (Figure 9) is the only area illustrating a direct spatial relation between reef passes, gullies, and canyons (Sedimentary system 1, Figure 13A, Table 2). The northeastern edge of Barbuda Island shows a narrow fringing reef disrupted by a series of passes visible in satellite photographs (Figure 9E). Relatively large gullies are identified on the upper slope in front of the passes (Figure 9A), likely acting as pathways for sediment flux. Such features are very similar to those observed off the Bahamas Banks (Mulder et al., 2017; Fauquembergue et al., 2018). They most likely result from sediment resuspension in the lagoon during cold fronts and hurricanes, leading to sediment export by density cascading and ebb currents (Wilson and Robert, 1995; Mulder et al., 2017). This direct connection between the Anguilla Bank and the slope implies a carbonate source due to the carbonated nature of the bank (Christman, 1953).

The orientation of the carbonate platforms and shelf edges with respect to major winds and oceanic currents plays an important role in sediment transfer down to deep basins (Gonzalez and Eberli, 1997; Rankey et al., 2006; Counts et al., 2018). In the study area, the Anguilla Valley is located in the continuity of canyons incising the northeastern slope of the Anguilla Bank (Figure 10), which represents a leeward (i.e. downwind) edge of the bank (Figure 3). The Anguilla Valley seafloor is highly reflective (Figure 10), likely indicative of coarse sediment and/or efficient sediment transport. Upstream, numerous bedforms, that we interpreted as submarine dunes are detectable along the western part of the Anguilla Bank (Figure 3). They suggest a westward migration of carbonate sands, in accordance with the main direction of the Caribbean oceanic current and winds (Richardson, 2005). The shelf edge is favorably oriented for sediment overspill caused by dominant currents, and the bedforms seem to migrate off-platform (Figure 3). They thus most likely constitute a source of sediment supply as documented in other regions (Gonzalez and Eberli, 1997; Jorry et al., 2016; Counts et al., 2018). The presence of numerous canyons located directly at the shelf break supports this interpretation (Figure 3). We therefore infer, by analogy, that sediment overspill is a major process for sediment delivery on the east of the Anguilla Bank (beige arrows, Figure 13A). This implies the remobilization of carbonate sediment and its transport from the Anguilla Bank to the slope (Figure 13A).

A similar configuration, with a shelf edge favorably oriented for overspill and canyons located downslope, exists off the northwest Antigua Bank (Sedimentary system 2; Figure 13A, Table 2). However, the Barbuda and Saint-Barthélémy valleys located downwards present a less reflective seafloor, likely transferring less or finer-grained sediment (Counts et al., 2018) compared to the Anguilla Valley. This fining is probably due to the greater distance (∼50 km, Figure 9) of the canyon heads from the shelf break, but also to the gentler slope gradient, which could favor deposition rather than transport.

The southeastern slope of the Anguilla Bank presents an opposite configuration, being located off a windward edge (i.e., facing the wind; Figures 3, 13A). In this sediment system, canyon heads have been identified on the platform edge (Figure 5A) and numerous highly reflective lobate patches are found in the downstream reaches of the St-Barthélémy Valley (Figure 10). Due to their high to medium reflectivity (Figure 10A), these features are interpreted as coarse-grained deposits, located close to the slope break. Given their lobate structure and based on other case studies, they could be interpreted as MTDs (Moscardelli and Wood, 2016) or submarine fans (Shanmugam, 2016).

Retrogressive erosion is a common process on submarine slopes in siliciclastic (Pratson et al., 1994; Pratson and Coakley, 1996), carbonate (Tournadour et al., 2017), or mixed sedimentary systems (Puga-Bernabéu et al., 2011). This process mainly affects the slopes disconnected from any input from the carbonate banks, for example on the Sombrero Escarpment or the Tintamarre Spur (Figures 6, 8).

The Sombrero Escarpment (Figure 6) is incised by numerous submarine canyons that are not directly connected to the Anguilla Bank. It is separated from the Anguilla Bank by the ∼60-km-wide Anguilla Depression located below the storm-wave base (300 mbsl), where waves generate no water motion (Peters and Loss, 2012 and references therein), preventing the entrainment of sediment particles down the canyons of the Sombrero Escarpment. The Sombrero Escarpment canyons are mostly slope-confined and disconnected from any feeding channel. Additionally, a ∼100 m-high bathymetric elevation at the edge of the Anguilla Depression (Figure 5C) most likely acts as a dam, preventing sediment overspill onto the slope. The canyons incise the seafloor composed of volcanic basement or mostly volcaniclastic deposits as attested by dredged rocks collected along the escarpment (Bouysse et al., 1985) supporting that the deposits transferred within this siliciclastic system derived from volcanic sources. The low reflectivity downstream of the canyons (Figure 4) suggests that the deposits are probably fine-grained, but also that the sediment transfer is limited.

On the Tintamarre Spur (Sedimentary system 4, Figure 13A, Table 2), two paleo-platforms are situated at 700 mbsl and 1,600 mbsl (Figure 8). Between the shallowest paleo-platform and the shelf edge, no gullies, no bedforms, and no areas of high reflectivity are identified, suggesting that there is no sediment transfer from the Anguilla Bank at the moment. Both paleo-platforms show a gently dipping surface (2.4–6.6°), with a sharp slope break (16.6–18.6°) affected by spoon-shaped head scarps with gullies and canyons downwards. Due to the absence of connection with the Anguilla Bank, these canyons were probably formed by retrogressive erosion.

Many canyons (and gullies) are clearly identified on the steep insular slopes of the NLAA (Figures 6, 8–11), but they are generally scarcely incised and slope-confined. This diffuse distribution leads to the presence of multiple but relatively small point sources along the insular slopes. The canyons feed the foot of the slope, resulting in a scattering of sediment fluxes, which, combined with the very low accumulation rates (from 1 to 3 cm/kyr, Reid et al., 1996), supports a sediment supply insufficient to build a channelized fan turbidite system (Yose and Heller, 1989; Nelson et al., 1999). Such scattered distribution is observed along and at the base of the entire steep and rough slope (slope gradient up to 25°, Seibert et al., 2020) down to Guadeloupe, while southwards from Guadeloupe, the smoother physiography shows well-identified single point sources, and associated distinct and long canyons and channels at their base (Seibert et al., 2020).

Additionally, the presence of a plunge pool associated with a slope-break deposit at the base of the slope in the Saint-Barthélémy (Figure 11) and Anguilla valleys (Figure 10) supports the occurrence of hydraulic jumps. Indeed, a slope-break deposit can be formed due to a significant change in slope gradient (Mulder and Alexander, 2001; Bourget et al., 2011; Migeon et al., 2012) close to a canyon mouth (Komar, 1971). The slope break that we observe ranges from 4° to 7°, and is consistent with other observations compatible with the formation of hydraulic jumps (Lee et al., 2002). One effect of hydraulic jumps is to disrupt the gravity flow, and thus maintain sediment in suspension and favor transport at the slope break (Gray et al., 2005). This process may contribute to lateral sediment dispersal towards distal sites and thus prevents a point source of sediment and channelization at the foot of the slope.

Some features of the NLAS support the importance of eustatism on carbonate sediment productivity and transfer: A) drowned reefs or platforms on the Anguilla Bank indicate past periods of exposure and/or drowning of the platform (Schlager, 1989, 1998; Leclerc et al., 2014; Leclerc and Feuillet, 2019; Carey et al., 2020), and B) unconformities between seismic units in the Anguilla Valley (Figures 10C,D) and the St-Barthélémy Valley (Figure 7B) highlight periods with variation in sediment transport.

In tropical carbonate turbidite systems, eustatic fluctuations have generally a high impact on sedimentation (Dunbar and Dickens, 2003; Tcherepanov et al., 2008). During sea-level highstands, carbonate productivity is usually enhanced (Reijmer et al., 1988; Schlager et al., 1994; Chabaud et al., 2016), providing high amounts of bioclasts and carbonate muds available for sediment transport. Conversely, during sea-level lowstands, production and off-bank transport are generally reduced while cementation, lithification, and induration of shallow-water carbonates are increased (Schlager et al., 1994; Chabaud et al., 2016), and thus prevent remobilization by currents and waves. Moreover, dissolution rather than erosion of carbonates prevails onshore (Chabaud et al., 2016), furthermore reducing sediment input during lowstands. Limited deposition of gravity flows may however also occur during lowstands, mainly at proximal locations of turbidite systems as documented for the Glorieuses islands (Jorry et al., 2020) and supported by higher seafloor reflectivity, suggesting some reworking on the upper slope.

Carbonate systems of the NLAS are therefore most likely active during sea-level highstands and less active during sea-level lowstands, such as in other mixed systems (Schlager et al., 1994; Jorry et al., 2008; McArthur et al., 2013). In contrast, the top of the siliciclastic system is too deep (∼900 mbsl) to be influenced by glacio-eustatic sea-level fluctuations as usually observed in siliciclastic systems, which are more active during sea-level lowstands (Posamentier et al., 1991; Vail et al., 1991; Hübscher et al., 1997).

Tectonics can play an important role in sediment transfer and destabilization, particularly along active margins, where vertical motions can deflect or dam transfer axis, modify the availability of sediment supply or drive the failure of over-steepened slopes (Laursen and Normark, 2002; Mountjoy et al., 2009; Michaud et al., 2011; Ratzov et al., 2012; Collot et al., 2019; Seibert et al., 2020; Claussmann et al., 2021a). Subsidence outlined by the presence of two tilted carbonate paleo-platforms at 700 mbsl and 1,600 mbsl (Figure 8) is consistent with the drastic subsidence of the forearc identified by Boucard et al., 2021. The present-day relatively deep position of these features, which were formed at a very shallow depth, attests to the important role of subsidence, probably accentuated along the Tintamarre Spur. For instance, at Les Saintes (offshore Guadeloupe), the subsidence of the Quaternary coral reef (Leclerc et al., 2014) is estimated at a rate of -0.3 to −0.45 mm/yr since the late Pleistocene (Leclerc and Feuillet, 2019), while in the north, this estimation reaches 340 m/Myr (0.34 mm/yr) in the deepest outer forearc since middle Miocene (Boucard et al., 2021). Possible causes of subsidence are debated, but include: A) a thermal origin caused by the cooling of the lithosphere related to a decline in volcanic activity and westward migration of the volcanic arc since the Miocene (Bouysse and Westercamp, 1990; Legendre et al., 2018); B) regional subsidence of the arc due to long-term plate motion along the subduction megathrust over several seismic cycles (Leclerc and Feuillet, 2019); or C) basal erosion of the upper plate by over pressured fluids related to a fully-hydrated subduction zone (Paulatto et al., 2017; Marcaillou et al., 2021).

Independently of the causes of tectonics in the area, two general effects are directly induced. The first is the opening of V-shaped valleys in map view (sensu Boucard et al., 2021), which created a complex morphology departing from a classical circular or elongated shape of the banks and shelf-edges; this promotes multiple edges that are either windward or leeward oriented, thus favoring, or not, carbonate transport off-bank. The second is the over-steepening of valley walls (or spurs, or flanks) bounded by normal faults. This last effect directly influences slope stability, the amount of retrogressive erosion, sediment production, and the formation of plunge pools due to hydraulic jumps. Indeed, the stability of plunge pools over time; demonstrated by the presence of a buried and tilted plunge pool along with unconformity U4, would have involved a slope-break angle of a few degrees to allow for a hydraulic jump (Komar, 1971). This break is probably maintained by fault activity, preventing sediment accumulation and progressive flattening at the foot of the slope (Bourget et al., 2011). The two combined effects of tectonics thus account for the long-term lack of point sources, scattered sediment delivery, and the mixing of the system.

The main particularity of the NLAS is the proximity of a subduction zone, implying possible seismically-triggered gravity-driven currents (see for example Goldfinger et al., 2003; Ratzov et al., 2015). In at least seven separate forearc sub-basins, thick transparent units (megabeds) are visible in sub-bottom profiles (Figure 12) and could correspond to homogenites or megaturbidites. Similar deposits are explained in the literature as resulting from surficial sediment remobilization (McHugh et al., 2016; McHugh et al., 2020), or from a seiche effect induced by seismic waves in a confined basin or in a lake (Chapron et al., 1999; Beck et al., 2007). Alternative explanations include the waning flow of a turbidity current triggered by an earthquake (Polonia et al., 2017) or the resuspension of shallow sediment due to a tsunami (San Pedro et al., 2017). The identification of MTDs (Figures 7A, 12D,H), which can be triggered by earthquakes (Moscardelli and Wood, 2016 and reference therein), confirms that seismicity is an important parameter considering sediment transfer in the study area. Ongoing detailed chronostratigraphy and dating of these deposits (Morena, 2020) should define the timing and spatial correlation across the margin, to infer a regional trigger, most likely caused by an earthquake as documented by Goldfinger et al. (2003).