Late Pleistocene-Holocene age and stratigraphy of the Currituck Slide Complex, U

.


Introduction
Considerable effort has been made recently to identify patterns in submarine landslide timing and to link these events to changes in local and global-scale environmental conditions (e.g., Lee, 2009;Urlaub et al., 2013;Brothers et al., 2013).The difficulty of dating submarine landslides has hampered efforts to identify statistically significant patterns and correlate them with driving mechanisms, such as rapid changes in sea level and global climate.Without a well-supported understanding of the conditions that lead to and trigger seafloor instability, our ability to assess future landslide and tsunami-generation potential remains limited.
The Currituck Slide Complex (CSC) offshore North Carolina, USA, is a large (total volume of ~160 km 3 ; Locat et al., 2009), likely retrogressive landslide at the site of a Pleistocene shelf-edge delta often associated with the Pleistocene and earlier pathways of mid-Atlantic coastal plain rivers such as the Roanoke and James.Mass transport deposits (MTD) from the terraced evacuation zone (Fig. 1) extend >250 km downslope.The morphology of the evacuation (scars) and proximal MTD regions of the CSC (Fig. 2) is composed of a series of 200-400 m high headwall scarps (upper and lower headwalls) separated by a shallowly seaward-dipping bench.Chaotic seafloor seaward of the lower headwall marks the beginning of thick MTDs sourced from both the CSC and adjacent slopes.The slope north of the CSC evacuation zone is cut by canyons and has not failed, while to the south a partially intact remnant of the pre-failure slope is observed (Fig. 2b).Details of the distal morphology of the MTD and the deeper stratigraphic framework of the CSC and surrounding area are described in Twichell et al. (2009) and Hill et al. (2017).
The large event volume, shallow headwall depth and proximity to the coast have made the CSC a benchmark source for modeling landslide-generated tsunami hazards to the U.S. east coast (e.g., Geist et al., 2009;Grilli et al., 2016).Yet, although the morphology, preconditioning scenarios, and possible mobility of the landslide have been explored in ever-increasing detail (Bunn and McGregor, 1980;McGregor, 1981;Prior et al., 1986;Locat et al., 2009;Hill et al., 2017), the age of the most recent events were only constrained by timing based on regional sedimentation rates (Bunn and McGregor, 1980;Prior et al., 1986).The broad range of estimated ages, from 25,900 to 46,800 yrs.BP (Prior et al., 1986) to the less specific "post-early Pleistocene" (Bunn and McGregor, 1980) have precluded evaluation of environmental drivers that triggered the landslide and were unsuitable for probabilistic hazard analysis purposes.Bunn and McGregor (1980) provided the first core-based analysis of the detailed shallow stratigraphy and sedimentology of the CSC excavation.The presence of clay balls, sand layers and friable clay layers within some of the cores indicated possible post failure deposition and slide surfaces shallower than seismic reflections profiles suggested.Issues with core siting and a lack of absolute age control could only point to a broadly Pleistocene failure age and made the identification of postfailure sedimentation, MTD facies, and unfailed stratigraphy difficult.
Here we describe the shallow stratigraphic composition of the evacuation and proximal deposition zones of the CSC and report the first 14 C radiocarbon supported age for the most recent large failure and one younger adjacent landslide.The coincidence of the failure timing with shifts in global climate and resultant increase in glacial meltwater through multiple eastern U.S. terrestrial river system pathways to the continental margin is explored as a driver for failure of the Currituck shelf-edge delta.

Data and methods
Core siting and interpretation of the surficial and shallow sub-surface morphology of the evacuation and proximal deposition zones of the CSC (Fig. 2) were carried out using a multibeam compilation dataset (Andrews et al., 2016) and Knudsen 320BR/3260 Chirp/sub-bottom profiles from several cruises (Chaytor, 2012;Shillington, 2014;NOAA, 2013;NOAA, 2014;Cantwell et al., 2021).Sub-bottom profiler data were used to create isopachs of hemipelagic sediment and "abovelayered" (hemipelagic plus unevacuated MTDs) thickness within the   J.D. Chaytor et al. upper scar by converting their time to depth assuming a P-wave velocity of 1500 m/s.The profiles' thicknesses were then expanded spatially using a 2nd order polynomial "linear trend" gridding method in ArcGIS Pro 2.8© (ESRI, 2022).Images of the seafloor and headwall scarps of the CSC were collected by a drift camera in 2012 on the R/V Hugh R. Sharp and from the D2 remotely operated vehicle deployed during NOAA Ship Okeanos Explorer expeditions EX1903L2 and EX2103.
Sixteen piston cores (up to 8 m long) were collected in 2012 using a jumbo piston corer deployed from the R/V Hugh R Sharp across the upper sections of the CSC and adjacent unfailed slope between water depths of 250 and 1900 m (Fig. 2).Sites were chosen to avoid areas of obvious reworking such as scarp degradation or anthropogenic disturbance.Sediment characterization of each core was performed via a combination of visual description, visible and x-ray imaging, physical properties logging (Gamma density, magnetic susceptibility, p-wave velocity, fractional porosity, line-scan imagery via Geotek multi-sensor core logger) targeted grain-size analysis, carbonate content, undrained shear strength, and water content (Boggess et al., 2023).Grain-size analysis was performed on 1-cm thick subsamples using the methods outlined in Chaytor et al. (2022).Carbonate content was determined using a modification of the loss on ignition method outlined by Dean (1974) that increased heating time at the 550 • C (3 h) and 950 • C (2 h) steps, with no correction made for clay mineral structural water loss.Undrained shear strength was measured in a consistent fashion using a combination of handheld vane shear (torque gauge mounted vane) and handheld torvane for stiffer sediments (Chaytor et al., 2022).A combination of biostratigraphic datums and radiocarbon dating were used to establish the chronology of sediment accumulation and the last major landslide event.The first appearance of Globorotalia menardii, which is thought to reflect the post-glacial intrusion of warmer waters to the North Atlantic or resumption of Atlantic overturning circulation ~4000 years after the onset of the Holocene (Ericson et al., 1961;Ericson and Wollin, 1968;Keigwin et al., 2005;Broecker and Pena, 2014) was used to rapidly determine the approximate extent of Holocene and Pleistocene sediments.A sample spacing of 20 cm was used for establishing the G. menardii first appearance datum.Accelerator mass spectrometry (AMS) 14 C radiocarbon dating was performed on planktonic foraminifera and shell fragments picked from several 1 cm-thick intervals in a subset of the collected cores at the National Ocean Sciences Accelerator Mass Spectrometry facility.Wherever possible, singleplanktonic species samples were analyzed.Low abundance of planktonic foraminifera below the upper 2-3 m of sediment in the cores necessitated the use of mixed-planktonic species samples for some analyses.To obtain a date from the shell debris layer at the base of HRS1209-6PC, a fragment of an unidentified bivalve species was used.Two-sigma calibrated ages (BP) were calculated using Calib 7.1 (http ://calib.org/calib/)and IntCal Marine13 (Reimer et al., 2013).No reservoir correction beyond the standard 400 years was applied.

Surficial morphology and shallow sub-surface structure
Because the broader-scale stratal architecture of the CSC has been extensively described by Hill et al. (2017), attention is focused here on the surficial and shallow buried morphology and structure of the CSC in the vicinity of the core sites (Figs. 2 and 3).The CSC is defined by three features: the low-relief, seaward dipping upper-scar platform (yellow in Fig. 2a), the hummocky debris-covered lower scar (orange in Fig. 2a), and the partially intact remnant slope section (Fig. 2a and b) on the southern flank of the CSC.

Upper scar
The upper scar platform is bordered to the west by the up to 275 m high, ~ 24.5 o (average) upper headwall, displaying several arcuate or scalloped forms and devoid of significant benches.The eastern edge of the upper scar platform is defined by the top of the morphologically complex lower scar headwall (Fig. 2).Canyonized upper slope borders the upper scar to the north and a largely intact block of upper slope separated from the shelf edge to the south.The average gradient of the upper-scar platform is 3.5 to 4.5 o (Fig. 3), with a steeper gradient (up to 8.0 o ) at the base of the upper headwall, decreasing to <3.0 o at the eastern edge (Fig. 3).Adjacent to the shelf-edge headwall indentation, the profile of the upper scar platform shifts in places from concave at the base of the upper headwall, to convex halfway across the platform (Fig. 3).
The generally low relief of the upper-scar platform is devoid of largescale current-generated features and gullying or other erosive processes, except for the elevated scar area adjacent to the shelf-edge headwall.The platform is mantled by hemipelagic sediments (Figs. 4 and 5) with a mean thickness of ~7.3 m.The hemipelagic sediments at the seafloor are slightly mounded and mixed via bioturbation (Fig. 6a).Small surficial debris lobes sourced from the upper headwall disrupt the sedimented slope in a few places and are visible on the bathymetry, sub-bottom profiles, and visual images (Fig. 6b).The thickness of the hemipelagic sediments is greatest closest to the upper headwall and decreases eastward and southward (Fig. 5), indicating that they are likely sourced from the upper headwall and/or the shelf.
A variable thickness layer of chaotic and discontinuous reflections between the base of the hemipelagic sediments and the top of the underlying continuously layered stratigraphy (slide plane in Fig. 4) is present across most of the upper scar platform.The chaotic and poorly resolved nature of this layer and its stratigraphic position suggest that it consists of material that was incompletely evacuated from the upper scar during failure or that was deposited shortly after failure and before the resumption of background sedimentation processes.The combined volume of the sediments deposited above the continuously layered stratigraphy in the upper scar is estimated to be ~2.67km 3 .

Lower scar
Blocky MTDs dominate the morphology of the lower scar, extending from the base of the lower headwall out to the lower continental slope (ten Brink et al., 2014).Although draped by hemipelagic sediments (Section 3.2.1), the chaotic nature of these the near-surface MTDs prevent clear imaging of the hemipelagic interval or the internal structure of the MTDs.Twichell et al. (2009) estimated the thickness of the MTDs from the most recent failure to be on average 20 m.The lower scar headwall is morphologically complex, formed by series of scarps and benches exposed at the face of the headwall and within u-shaped indentations that can extend over 3 km into the upper-scar platform (Fig. 2).The height of the lower headwall varies considerably, from as little as ~70 m in the south adjacent to a prominent step of intact stratigraphy covered by MTDs, to >400 m high along the central portion of the scarp.The overall gradient of the scarp is less than 15 o , but individual bench scarps can exceed 30 o .The exposed semi-lithified sediments along the low headwall are heavily bioeroded, fractured and indurated (Fig. 6c), with extensive development of debris aprons at the base of the steeper scarp sections (Fig. 6d).

Hemipelagic sediments
Sub-bottom profiler data shows that the thickness of the hemipelagic interval in the upper scar varies from >20 m below the headwall scarp to 4 m or less above the lower scar headwall (Fig. 5).Piston cores recovered between 3.50 m (8PC) and 7.3 m (2PC) of hemipelagic sediments in the upper scar (Figs. 7 and 8), within the range indicated by the sub-bottom profiler data.The thickness of the hemipelagic interval in the lower scar is difficult to establish on the sub-bottom profiler records, but core recoveries were between 2.60 (11PC) and 4.25 m (10PC), which is consistent with a decrease in sediment availability farther from the coast and the shelf break (Figs. 1 and 2).The entire 6.45 m length of 16PC, located on top of the intact slope block, is composed of hemipelagic material (Fig. 9) and used as a background sedimentation reference core.
The hemipelagic sediment is olive-green (Munsell: 10Y 4/1; Munsell Color, 2000), often intensely bioturbated clayey silt to silty clay, and forms a continuous drape over both the upper and lower scars, as well as the 'intact block' and the MTD channel to the south (Southern MTD in Fig. 2).In contrast, the hemipelagic drape above the upper headwall is significantly sandier than elsewhere across the CSC.Core 5PC above the upper headwall at ~345 m water depth contains of a mixture of siltysand, sandy clay-silt and poorly sorted interbedded sand and silty-clay layers.Core 1PC, located in a failure-generated shelf reentrant at ~530 m water depth, is composed entirely of a sandy clay-silt unit which is also present in 2PC, 3PC, and 5PC (Figs. 7 and 8).Across the upper and lower scar areas, the composition of these hemipelagic sediments is relatively uniform (Figs. 9 and 10): clay-rich silt (mean grain sizes of between 6 and 8 ϕ), with wet-bulk densities between 1.5 and 1.6 g/cm 3 , calcium carbonate content between 5 and 10%.Water content falls from ~60% at the tops of the cores, to ~40% at the base of the hemipelagic interval, consistent with compaction dewatering.Undrained shear strength values are relatively consistent within the A prominent bulge in the magnetic susceptibility within the hemipelagic interval is present in all cores within the upper and lower scar areas (Fig. 10) but absent in those cores above the headwall (1PC and 5PC) and on the top of the intact block (16PC).Grain size data that span these intervals show a marked increase in the sand size fraction (Fig. 11), although the mean grain size is largely unchanged.At a finer scale, the textural coarsening is driven by a shift of material from the 5 ϕ (coarse silt) to 4 ϕ (very fine sand) size classes.

Mass transport deposit facies
The hemipelagic interval conformably overlies one or more of the following facies types of MTD: MTD Facies I: Inhomogeneous mud-sand-gravel matrix-supported mud-clast conglomerate, which is characterized by polymictic mudclasts, some containing deformed strata, supported by a variable thickness chaotic, inhomogeneous mud-sand-gravel matrix.It is similar to Facies IVa of Tripsanas et al. (2008) where it is interpreted to be representative of shearing, plastic mixing and liquefaction (Fig. 12).
MTD Facies II: Hard-clast-supported mud-clast conglomerate composed of hard (S u 50-200 kPa) rounded to sub-angular polymictic mud clasts with a fine-grained matrix.Strongly deformed and horizontal layering are present throughout.It is similar to Facies IIIb of Tripsanas et al. (2008) interpreted to be from breakup of soft sediment mass.Fig. 12).
MTD Facies III: Poorly to very-poorly sorted, ungraded shell fragment-rich sand-gravel (Fig. 12).The bulk of this facies contains <5% silt and clay, with CaCO 3 contents as high as 45% where larger shell fragments are encountered.Some shell fragments had black surface patina, but no substantially intact shells were found that would facilitate identification.As MTD Facies III was only encountered in a single core, it may be a localized deposit sourced from a shallow-water shell bed.
MTD Facies IV: Normally graded sand-silt deposit/turbidite.In two cores, 7PC and -8PC (Fig. 8), a grey-green, clay-rich unit with abundant iron-sulfide staining was intersected below the MTD intervals.Undrained shear strength in this interval is 5 to 10 times higher than that of the hemipelagic interval (up to 200 kPa) and is similar to values measured in several of the hard mud-clasts of MTD Facies I (Figs. 10 and  12).Limited penetration of the cores into this unit prevents a more thorough investigation of the properties of the sediments below the MTDs and foraminiferal abundances were too low to recover sufficient material for radiocarbon dating.

Core geochronology
Calibrated AMS 14 C ages (Table 1) in the hemipelagic interval reflect relatively undisturbed Holocene and latest Pleistocene deposition.Differences in age between matched trigger and piston cores, suggest, however, that the piston cores may not recover the last few thousand years of deposition.Variations in the age of sediments in the hemipelagic interval across the suite of cores likely reflect both overall changes in local sedimentation processes as sea level rose and reached farther shoreward of the landslide in addition to enhanced and persistent bioturbation.G. menardii is consistently present in the upper ~1.75-2.25 m of all the cores.Its first appearance falls between calibrated ages of 7500 and 8300 yrs.BP and provides additional age control and correlation.Hemipelagic drape sediment samples within 50 cm of the contact with the underlying MTD facies have 2-sigma calibrated radiocarbon ages between 13,200-and 13,600-yrs BP in cores from both the upper and lower scars (Figs. 7 and 8).Samples taken closer to the hemipelagic-MTD contact contained very low abundances of foraminifera precluding more precise dating of the surface.Extending the minimum and maximum sedimentation rates derived from the lowest dated intervals of these cores (25 and 69 cm/kyrs.Respectively, Table 1) to the base hemipelagic-top MTD contact results in an age range between 13,835 and 16,020 yrs.BP.Datable material was absent, or its concentration was too low, in the MTD facies and in the homogeneous basal clay in every core.The only exception was 6PC, where the shell fragments in MTD Facies III had an uncalibrated age of >52,000 yrs. at a depth of 6.22 m (Fig. 8).The basal age of hemipelagic sediment, 20 cm above the MTD facies of the "South MTD" (core 14PC, Fig. 13), was found to be significantly younger than the main failure event with an age of ~5500 yrs.BP.This age is consistent with the presence of G. menardii throughout the entire hemipelagic interval of core 14PC.
Hemipelagic sediments comprise the entirety of core 16PC atop the unfailed slope block and appear to reflect continuous sedimentation over at least the last 17,000 years (Fig. 9).Near identical ages for sediments at core depths of 490 and 620 cm and a sedimentation rate of >150 cm/kyrs between core depths of 120 and 490 cm (Table 1, Fig. 9), may reflect either extensive sediment mixing, or perhaps a period of rapid sedimentation that coincides with increased sediment availability around the time of the CSC failure event (Section 4.2).

CSC failure timing
Only a single large-scale failure event between 13,835 and 16,020 years (BP) ago was captured in the shallow CSC sedimentary record, hence, a full picture of the complete failure history for the CSC remains elusive.Nevertheless, the well-correlated ages of hemipelagic drape between the upper and lower scars show that MTD deposition below the lower headwall scarp occurred at the same time as failure of the upper scar area, requiring either a bottom up or co-eval retrogressive failure process that mobilized a large volume of sediment and fundamentally reshaped the mid-Atlantic continental slope.The presence of smallerscale, younger landslide events such as the ~5500 yrs.BP "South" MTD and scarp degradation deposits at the base of the upper-headwall, shows that the region continues to be reshaped, albeit at a smaller scale.A combined upper and lower failure source with a volume of 160 km 3 (Locat et al., 2009;Geist et al., 2009) provides the most conservative constraint on landslide sediment volume for hazard analysis studies.The volume of unevacuated mass transport material indicated by the subbottom profiler data in the upper scar (2.67 km 3 ) does not significantly alter the combined CSC failure volume.Further work on fingerprinting the source of debris components within the MTD (e.g., clay and silt ball mineralogy, shell species habitat zones) deposited within the scars may reveal additional information on the failure sequence.
Previous age estimations for the timing of the CSC relied on sedimentation rates from dated cores up to 250 km to the north (Doyle et al., 1979;Prior et al., 1986), where sediment sources and oceanographic conditions are likely different.These spatially extrapolated rates  substantially underestimated the sedimentation rates and, as such, resulted in ages for the landslide that were as much as 30,000 years older than the current radiocarbon-based age and potentially much older than other dated landslides along the margin (Embley, 1980;Lee, 2009;ten Brink et al., 2014).However, even local sedimentation rates derived from 14 C and 210 Pb ages from the upper 250 cm of cores within the upper and lower CSC landslide scars (e.g., Table 1, this paper; Alperin et al., 2002;Thomas et al., 2002) can vary by an order of magnitude due to variability in sediment availability and transport from the adjacent shelf-slope and along slope (Gulf Stream driven) systems.Changes in sediment sourcing within the hemipelagic interval may be responsible for the sediment grain size change indicated by the magnetic susceptibility bulge (Figs. 10 and 11).Such local variability at or near the sediment/water interface highlights the difficulty in extrapolating sedimentation rates both along the margin and deeper in the sediment column.

Preconditioning and failure drivers
In the absence of modern observational evidence of the seafloor failure mechanisms and the processes that trigger them, it is widely assumed that these failures are primarily triggered by horizontal acceleration from earthquakes, which augment the static loading and gravitational forces acting on the continental slope (e.g., ten Brink et al., 2016).Although substantially less frequent than on active plate boundaries, moderate to large magnitude earthquakes frequently occur even on "passive" margins such as the U.S. mid-Atlantic.Events such as the M7.2 1929 Grand Banks earthquake have shown that along these margins, earthquakes large enough to destabilize a suitably located preconditioned seafloor (ten Brink et al., 2009) do occur and can result in significant coastal impacts via tsunami generation (e.g., Fine et al., 2005).Additionally, propagation of seismic waves over long distances for smaller magnitude earthquakes has recently been shown to be enhanced where suitable antecedent geology is favorable and seismic attenuation is low (e.g., U.S. east coast; Pollitz and Mooney, 2015), increasing the potential of smaller magnitude earthquakes to trigger slope failures (ten Brink et al., 2009).Over the last 20 years, 40 earthquakes with magnitudes >2.5 have been recorded within a 350 km radius of CSC, including an M4.6 on the continental rise on January 5, 2019, 160 km away from the headwall of the CSC (https://earthquake. usgs.gov/earthquakes/eventpage/us2000j4bf/executive;Fig. 1), highlighting that a seismic trigger for failure of the CSC is possible given favorable ground conditions.However, the long time, 13,835 to 16,020 years, since the last major failure event of the CSC and shifts in global climatic condition, continental shelf/shelf-edge oceanographic processes, and terrestrial sediment sources suggests that pore pressure, sediment composition, and grain size distribution of the sediments at that time may have been different from those at the present-day possibly enabling failure at lower earthquake magnitudes.
Extremely high sedimentation rates, especially at low-stand shelfedge deltas and associated continental slopes (e.g., Ducassou et al., 2009;Covault and Graham, 2010;Urlaub et al., 2013) can also lead to failure even in the absence of horizontal acceleration by earthquakes.Sediment consolidation decreases and pore pressure increases rapidly for sedimentation rates >500 cm/1000 yr., especially for clay-rich sediments (e.g., Morgenstern, 1967).Although unreasonably high in Fig. 11.Cumulative weight percent plots of the primary sediment textural components for HRS1209-6PC (A) and HRS1209-8PC (B) showing the change in grain size, which we interpret to be responsible for the magnetic susceptibility anomalies within the hemipelagic drape.most ocean settings over geologic time, these rates can occur in modern deltaic settings during flooding periods (Hart et al., 1998;Keller et al., 2016) and were perhaps present during drainage bursts of glacial lakes following the Last Glacial Maximum (LGM) (Ivanovic et al., 2017).
Although the presence of a Quaternary shelf-edge delta at the site of the CSC has been linked to the paleo-Roanoke and James rivers as the primary drainage source (Poag and Sevon, 1989;Twichell et al., 2009), climatic changes on these rivers' relatively small drainage basins may not have been sufficient to trigger failure, or even impact stability conditions of the shelf-edge delta in the vicinity of the CSC alone.The age range for failure of the CSC is coincident with significant shifts in northern hemisphere terrestrial and global oceanographic climate, including Heinrich Event 1, Melt water Pulse 1A (MWP-1A), and the Bølling-Allerød warming (Fig. 14).MWP-1A is a globally identified period of rapid sea-level rise of approximately 15 m, perhaps linked to Bølling-Allerød warming, which occurred between 14.7 and 14.3 ka (Clark et al., 2002;Deschamps et al., 2012).The southern Laurentide Ice Sheet (LIS) has been identified as a significant melt water source at this time, draining to the Gulf of Mexico via the Mississippi River (Aharon, 2003) and to the Atlantic Ocean via the Gulf of St. Lawrence (Clark et al., 2001) with sediment deposits marking the event.
Recently, the role of drainage flowing to the U.S. Atlantic coast, in carrying a significant portion of the sediment laden LGM and post-LGM southern LIS meltwater into the Atlantic Ocean has been evaluated (Carlson and Clark, 2012;Margold et al., 2015;Wickert, 2016;Ivanovic et al., 2017).Although the Hudson River has been identified as a primary meltwater pathway to the U.S. Atlantic margin due to its direct linkage to the southern LIS (Carlson and Clark, 2012), the Susquehanna/ Potomac (Chesapeake Bay) river system has been recognized by Wickert (2016) as another potentially significant U.S. Atlantic coast outlet for post-LGM meltwater delivery to the eastern Atlantic Ocean.The Susquehanna River, with similar headwaters to the Hudson River in the north Appalachian Mountains, has drained through Chesapeake Bay to the north of the CSC throughout the post-LGM period and may have also carried significant sediment volume at high rate to the mid-Atlantic continental shelf edge and slope.Wickert (2016) estimates that Susquehanna River discharge via a combination of meteoric and meltwater inputs may have been as much as 30 times greater following the LGM than the current 60-year average (1153 m 3 s − 1 ; Brodeur et al., 2019).Enhanced bedrock erosion due to the increased flow within the Susquehanna and adjacent Potomac drainage basins toward the end of the LGM and into the post-LGM period, ending at around 13-14 ka (Braun, 1989;Reusser et al., 2004), likely increased the suspended sediment load transported to Chesapeake Bay and the continental shelf edge (Fig. 15).
Periglacial conditions in the Appalachian Mountains also probably reached their maximum just prior to MWP-1A (Nelson et al., 2007).Periglacial incision rates in the southern Appalachian Mountains were an order of magnitude larger than present fluvial erosion and transport rates in the region (Braun, 1989;Hales et al., 2012).Increased discharge of sediment-laden river flow, in addition to periods of possible catastrophic flooding due to failure of Appalachian landslide dams (Nelson et al., 2007), may have increased the delivered sediment volume from smaller Appalachian sourced mid-Atlantic river systems (e.g., James and Roanoke rivers) directly to the continental slope during lowered sea level (Fig. 15).Although these regional and global climate and landform shifts are coincident with the last major CSC failure event, testing the sediment delivery and failure hypotheses is challenging.Remnants of the shelfedge delta and shelf/slope sediments accumulating prior to and during the post LGM period have been incorporated in the CSC MTDs and would be deeply buried or dispersed.
The significance of the ~5500 yrs.BP "South" MTD as a reflection of either local slope stability factors or changes in broader environmental conditions remains to be determined.Published records of morphologically well-defined and dated landslide events occurring during the relative environmental stability of the Late Holocene are limited (e.g., Laberg et al., 2003;Gracia et al., 2010), yielding little robust correlation to specific preconditioning factors or triggering mechanisms beyond seismic shaking.The recent identification of a ~ 1000-year recurrence interval of landslide activity on the statistically stable Laurentian Fan segment of the Scotian slope during the last 40,000 years (Normandeau et al., 2019) shows that areas of previous failure can remain susceptible to further seismically-induced landslide activity.The preservation of localized unstable conditions long after large-scale failure of adjacent slopes may prove to be a significant contributor to long-term slope modification through continued small-scale retrogressive erosion or periodic larger slope failure reactivation.Incorporation of these smaller, but still significant later events into geomorphological evaluations of landslide distributions (Chaytor et al., 2009;Urgeles and Camerlenghi, 2013) or under sampling of their younger ages, may have significant implications for marine geohazard assessments.

Conclusions
The timing of the most recent failure recorded within the sediments of the Currituck Slide Complex, one of the largest and most prominent submarine landslides along the U.S. Atlantic continental shelf and slope, has until now, been poorly constrained.AMS 14 C radiocarbon dating of hemipelagic sediments from above MTDs deposited during the last event was extrapolated to the slide plane using local sedimentation rate, and provide adjusted age of failure between 13,835 and 16,020 years BP.Thick Holocene and late Pleistocene hemipelagic sediments display morphologies indicative of plastic, shearing, and disintegrative behaviors of the pre-failure sediments, implying a complex or evolving failure process or marked variations in sediment rheologies.Continuous reshaping of the CSC scar area and adjacent continental margin is evidenced by smaller-scale landslides (e.g., ~ 5.5 ka"South" MTD), progressive degradation of scarps, and continental shelf and Gulf-Stream driven sediment deposition within the scar.
Timing of the most recent large-scale failure of the CSC is coincident

Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jason Chaytor reports financial support was provided by U.S.  (Brothers et al., 2020;Colman et al., 1990;Swift, 1976;Thieler et al., 2014).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Nuclear Regulatory Commission.Uri ten Brink reports financial support was provided by U.S. Nuclear Regulatory Commission.

Fig. 1 .
Fig. 1.Location of the study in relation to major river systems and the approximate position of the Laurentide Ice Sheet at the Last Glacial Maximum (21.7 ka yrs.calibrated) and at 15.5 ka yrs.(calibrated) (Dalton et al., 2020).The location of the January 15, 2019 earthquake is highlighted.Bathymetry data from Andrews et al. (2016).Land elevation and hydrography data from The National Map (U.S. Geological Survey, 2016; U.S. Geological Survey, 2017).

Fig. 2 .
Fig. 2. (A) High-resolution multibeam bathymetry of the Currituck Slide Complex (CSC) source area and proximal debris deposits, showing the location of collected and utilized cores, landslide scarps, and the distribution of surficial morphologies.The location of the sub-bottom profile shown in Fig. 4 is highlighted, as are the location of the core transects described in Figs.7 and 8. (B) section of migrated sparker multi-channel seismic reflection profile (Tiki2012-113c) across the intact slope block on the southern flank of the CSC.Data from Andrews et al. (2016).

Fig. 3 .
Fig. 3. Slope normal profiles showing variations in surface morphology of the CSC.Inset colour-shaded bathymetry map shows the locations of the profiles.

Fig. 4 .
Fig. 4. Uninterpreted (A) and interpreted (B) sections of sub-bottom profile (line EX1403L1-155) crossing the CSC from north to south across the upper scar area.See Fig. 2a for location).The profile shows a hemipelagic drape (yellow) above the indistinct/chaotic MTD layer and the underformed layered stratigraphy.Interruptions of the hemipelagic drape are due to small late-stage debris deposits likely due to scarp degradation.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5 .
Fig. 5. Hemipelagic isopach of the upper scar area in meters.Location of interpreted sub-bottom seismic profiles used to map the interval are show as dashed black lines.Pink dots are core locations.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 6 .
Fig. 6. (A) Image of the hummocky bioturbated hemipelagic sediments present on the upper scar seafloor surface, (B) debris at the base of the upper scar headwall illustrating the continued modification of the CSC morphologic elements (green lasers are 10 cm apart), (C) Near vertical lower scar headwall showing weakly defined layering and pervasive physical-and bio-erosion of the exposed semi-consolidated material, (D) Broken boulders and other debris near the base of the lower scar headwall.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 7 .
Fig. 7. Dated and correlated cores descriptions along a profile crossing the northern part of CSC.See Fig. 2 for location.Core descriptions are based on visual description, MSCL logs and discrete sampling.Correlations are based on the first appearance of Globorotalia menardii.Descriptions of the MTD facies are illustrated in Fig. 12.

Fig. 8 .
Fig. 8. Dated and correlated cores descriptions along a profile crossing the central part of CSC.See Fig. 2 for location.

Fig. 9 .
Fig. 9. Physical properties of core HRS1209-16PC/TC from the top of the intact block on the south side of the CSC.

Fig. 10 .
Fig. 10.Physical properties of cores HRS1209-6PC (A) and -8PC (B).The transitions from hemipelagic drape to MTD and the extremely stiff green silt/clay interpreted to be below the slide plane are marked by major changes in all measured properties.The magnetic susceptibility anomaly observed in the hemipelagic drape is indicated.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 12 .
Fig. 12. Core images X-ray, visual description (A and B only) and physical properties plots showing the characteristics of the four MTD facies discussed in the text.Red lines in Gamma Density/Mag.Susc.plots are magnetic susceptibility; blue lines are gamma density.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 13 .
Fig. 13.Visual description, full-length core image and dated intervals for core 14PC collected from the upslope portion of the "South" MTD as identified in Fig. 2.

Fig. 14 .
Fig. 14.Approximate timing of global and regional climatic, peak periods of river discharge and eustatic sea-level rise events during the late Pleistocene and Holocene with the timing of the last major failure at the CSC shown.YD -Younger Dryas, MWP-1A -Meltwater Pulse 1A, LGM -Last Glacial Maximum, H1 -Heinrich Event 1. Modified from Lambeck et al. (2014).

Fig. 15 .
Fig. 15.Location of the CSC in relation to identified paleoenvironmental features (red lines) of the U.S. mid-Atlantic shelf/slope and terrestrial drainages.Base map from USGS National Center for EROS (2003)(Brothers et al., 2020;Colman et al., 1990;Swift, 1976;Thieler et al., 2014).(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Table 1
AMS 14 C radiocarbon ages of foraminifera and bivalve samples from Currituck Slide Complex (CSC) cores.Sedimentation rates are calculated between dated intervals.