Submarine Mass Movements and their Consequences

Article 76 of the United Nations Convention on the Law of the Sea prescribes two approaches that a nation may employ to determine the extent of its’ legal continental shelf: (1) 60 nautical miles (M) seaward of the foot of the continental slope (FoS), or, (2) to a point seaward of the FoS where the sediment thickness is 1 % of the distance from the FoS. In both of these formulae, the “foot of the continental slope” is a critical metric. Article 76 defines the “foot of the continental slope” as: “

In the absence of evidence to the contrary, the foot of the continental slope shall be determined as the point of maximum change in the gradient at its base

”. Geomorphologic complexity or low gradients (<1°) of continental slopes rarely permit a ready determination of the maximum change in gradient; particularly at a position that a geologist might qualitatively recognize as the base-of-slope zone. Recognizing that submarine mass movement is a slope process that also influences the shape of the continental margin, several nations have successfully argued that the downslope termination of mass transport deposits assist in distinguishing the continental slope from the rise and abyssal plain. The Commission on the Limits of the Continental Shelf have now made recommendations for a number of coastal States with rift margins, transform margins and subduction margins where the extents of surficial mass transport deposits were used to help delineate the base of slope zone within which the foot of the continental slope is chosen.

Mass-transport deposits (MTDs) are important stratigraphic elements in many deepwater basins. MTDs have traditionally been identified as seals but can also act as migration pathways. Studying the character of deposits within a MTD from proximal to distal, in a framework of seismically identifiable morphologies provides a template for using seismic character to predict the petrophysical properties of such deposits. During failure and subsequent transport, MTDs are exposed to shear deformation and remolding that can enhance clay alignment and destroy large pore-throats thus creating potential seal quality facies. Deformation in the various MTD morpho-domains can be quantified by measuring the degree of clay-fabric alignment. In this study we investigate a MTD acting as the top-seal in the Jubilee gas field, Gulf of Mexico, by integrating 3D-seismic, core, and well-log data to characterize clay fabrics. X-ray-texture goniometry analysis was performed using core material from the top-seal MTD to determine the degree of clay fabric alignment. Final results indicate that samples have an anomalously high clay-fabric orientation not correlated with burial depth or diagenesis. We conclude that these zones with high clay-fabric alignment in the MTD are the result of shear deformation as the gravity flow moved downslope. Recognition of zones with enhanced microfabrics has important implications for shallow geohazards as well as sealing potential evaluation. This technique—although in its infancy—could be used to identify sealing MTD facies in core samples and outcrop studies.

Mass transport complexes (MTCs) are significant constituents of the post-evaporitic overburden in the Levant Basin, offshore Israel. Analysis of a new 3D seismic dataset offshore central Israel reveals that the Israel Slump Complex (ISC) consists of three stacked mass transport deposits (MTDs). The MTDs vary in lateral extent from between ~351 and 752 km

2

with thicknesses between ~190 and 325 m, accounting for a remobilised sediment volume of ~35–94 km

3

. Interestingly, each MTD is unique, exhibiting different geometries, internal architectures, and halokinetic-related imprints. We document a novel palm-like erosional morphology, blocky facies, arcuate facies, mounded structure, syn-depositional thrust systems, and a channelised geometry. These configurations indicate different transport distance, mechanics, and kinematic history for each MTD within the complex, and may suggest different trigger events. The results of this study shed light on the interplay of multiple MTDs within a greater MTC. It also provides new insights into the nature and formation of the ISC in the offshore area of central Israel, which arose from at least three short-timed mass wasting events during the Late Pliocene. Likewise it may serve as an analogue to understand the configurations of MTDs in basins with well-developed evaporite layers.

The interpretation of a new and extended 2D seismic database on the offshore Amazon Basin (Foz do Amazonas Basin) confirms the widespread presence of regional-scale mass-transport deposits (MTDs) that are important architectural elements of the Amazon Deep Sea Fan. These MTDs were deposited since the late Miocene and extend throughout an area of nearly 315,000 km

2

. They are grouped into three megaslide complexes: the northwestern Amapá Complex, the Central Amazon Fan Complex and the southeastern Pará-Maranhão Complex. Each complex has multiple stacked MTDs with various internal seismic facies that are indicative of large downslope modification and disruption of the original stratigraphy. The majority of the MTDs show chaotic or transparent internal seismic facies that we interpret as indicative of debris flow deposits. Although we cannot determine the exact triggering mechanism(s) for the various sediment failures, these events appear to be related to the gravitational compression of fold-and-thrust belts created by gravity-tectonic processes on the upper Amazon Fan and to structurally-induced mobilization of large blocks on the upper continental slope in response to overpressure along impermeable surfaces. These processes were apparently more active during the Pleistocene in response to increased sedimentation rates on the fan.

Mass transport deposits (MTDs) are formed by gravity driven processes whereby dominant transport direction is downslope. Here we use 2D seismic data from the deep water Taranaki Basin to describe two volumetrically extensive MTDs (MTD 1 and 2) that were emplaced within the Plio-Pleistocene succession. Key kinematic information, derived from the observed architectural relationships between the Aotea Seamount and MTD 2, suggest that this unit had a SW transport direction. This is in marked contrast to the NW transport direction derived for the underlying MTD 1. Given the geometry of these MTDs, we suggest that MTD 1 was triggered during the early stages of evolution of the Giant Foresets Formation when the system was prograding toward the north. On the other hand, the geometry of MTD 2 allowed us to infer that its headwall region was located toward the east (near the Northern Graben) in an area known to have been tectonically active at the time of deposition (c. 1.8 Ma to recent). These observations raise the likelihood of a tectonic trigger for MTD 2.

Based on high-resolution bathymetric records, sub-bottom profiles and sediment cores, we study postglacial mass transport deposits, slide scar complexes, cyclic steps and rockfall deposits as indicators of mass failures in the inner Hardangerfjorden system, western Norway. The stacked mass transport deposits show thicknesses of up to 4 m and witness that the inner Hardangerfjorden has been a site of repetitive mass failure events, potentially triggered by earthquakes related to glacioisostatic uplift. The cyclic steps, affecting an area of about 2 km

2

in the innermost fjord, have wave lengths of 40 m, heights of 5 m and are most likely related to fluvial sediment supply. Seven slide scar complexes, with stratigraphy-cutting scar heights of 6–34 m, are identified on the basin plain of the fjord. These are all associated with large depocenters along the fjord flanks, suggesting a link between locations of high sediment supply and mass failure. Although rapid deposition might, by itself, induce failures, an external trigger mechanism, such as an earthquake, should also be considered.

This study shows an onshore-offshore morpho-structural characterization of the El Golfo flank collapse and debris avalanche on El Hierro (Canary Islands, Spain). Erosive and depositional features have been identified based on: LIDAR topography and geology from water galleries (onshore); and high-resolution 3.5 kHz and multichannel seismic reflection profiles, and multibeam data (offshore). The onshore headwall scarp shows a non-continuous profile formed by two semi-circular amphitheatres and extends offshore by a smooth chute. The chute ends at about 3000–3200 m water depth in the distal depositional area. Multichannel seismic profiles show two major subunits of chaotic reflectors in the debris avalanche deposits. Results suggest that the El Golfo debris avalanche event likely took place in multiple stages. Consequently, we suggest that the multistaged nature of El Golfo debris avalanche greatly reduces the tsunamigenic potential of these flank collapses.

Newly collected multibeam and seismic data on the intra-slope Palmarola ridge show widespread pockmarks and landslide-related morphologies along its flanks. In detail, two main types of slope failures were identified: disintegrative-like and cohesive like landslides. The first type is characterized by a complex of small, nested scars affecting the steep and tectonically-controlled eastern flank of the ridge, suggesting a genesis related to retrogressive processes. The cohesive landslides affect the northern flank of the ridge and are characterized by larger scars, where material was not completely evacuated, and well-defined debris deposits at their base, with the development of pressure ridges. Tectonic activity and slope gradients represent the main controlling factors for the development of instabilities; moreover, we noted a relationship between pockmarks and landslide scars.

Over the last few years, canyons along the northern U.S. Atlantic continental margin have been the focus of intensive research examining canyon evolution, submarine geohazards, benthic ecology and deep-sea coral habitat. New high-resolution multibeam bathymetry and Remotely Operated Vehicle (ROV) dives in the major shelf-breaching and minor slope canyons, provided the opportunity to investigate the size of, and processes responsible for, canyon wall failures. The canyons cut through thick Late Cretaceous to Recent mixed siliciclastic and carbonate-rich lithologies which impart a primary control on the style of failures observed. Broad-scale canyon morphology across much of the margin can be correlated to the exposed lithology. Near vertical walls, sedimented benches, talus slopes, and canyon floor debris aprons were present in most canyons. The extent of these features depends on canyon wall cohesion and level of internal fracturing, and resistance to biological and chemical erosion. Evidence of brittle failure over different spatial and temporal scales, physical abrasion by downslope moving flows, and bioerosion, in the form of burrows and surficial scrape marks provide insight into the modification processes active in these canyons. The presence of sessile fauna, including long-lived, slow growing corals and sponges, on canyon walls, especially those affected by failure provide a critical, but as yet, poorly understood chronological record of geologic processes within these systems.

The Brazilian Continental Shelf Survey Programme (LEPLAC) identified the occurrence of large-scale mass-transport deposits on the southernmost limit of the Brazilian margin, based mainly on analyses of acoustic imagery. The mass-transport deposits, named the Chuí Megaslide Complex, comprise a stack of large translational slides that spread from the shelf break ~650 km downslope to ~4,900 m water depth, cutting into Pliocene-Quaternary sedimentary successions and strongly affecting both the margin morphology and regional depositional processes. The main headwall scarp is U-shaped, 400–500 m high, and extends

c

. 80 km downslope as a large elongated evacuated scar, 50–85 km wide. Outside this main failure scar, external scarps evidence a large area of erosion and faulted blocks, indicating ongoing retrogressive sediment disruption. Slide masses occur as a combination of variably deformed failed masses and debris flows, covering an area of

~

150,000 km

2

. Main preconditioning parameters and the possible triggering mechanism for the Chui Megaslide Complex are likely a combination of a series of causative factors such as slope failure structurally-induced by gravity tectonics and high sediment influx into the shelf-edge and upper slope during the Early Miocene-Quaternary.

An investigation conducted aboard the

RV Southern Surveyor

(

SS2013

-

V01

) in January 2013 offshore east Australia collected regional bathymetric data for the continental margin of southern Queensland between Noosa Heads in the south and Indian Head, Fraser Island in the north. This newly mapped area presents a particularly steep portion of continental slope (5–10°) that presents numerous submarine landslides, including two ‘whole-of-slope’ features (the Wide Bay Canyon, and Inskip Slides). The slope is also dissected by three large submarine canyons offshore northern Fraser Island, Wide Bay, and Noosa Heads (i.e. the Fraser Canyons, the Wide Bay Canyon and the Noosa Canyon). Dredge and core samples were collected from slide scars in the northern, central, and southern areas of the bathymetric survey area. The initial examination of the area’s bathymetry, the core and dredge sample sedimentology, and determination of biostratigraphic ages for these sediment samples indicates that the larger submarine slides present in this study area have probably been shed from the slope since the late Pliocene and that canyon incision is currently active on this portion of the slope. In one case, canyon incision is partly responsible for generating slides due to undercutting and removal of the toe of the slope. Slope sediments are dominantly comprised of hemipelagic muds but also include grain-flows and turbidites comprised of shelf-derived sands and upper slope sediment that have abraided the slope muds. The results confirm previous work that indicates that this margin is in an active phase of deconstruction dominated by mass failure.

Submarine landslides are one of the main marine geohazards worldwide. In order to better estimate their risk and develop mitigation measures, a better understanding of the failure mechanisms is needed. However, observing landslides in real time is near to impossible, hence careful study of both the failed sequence and the deposit is still the main source of information. Until recently, studies of the headwall scarps of submarine landslides were mainly based on shipboard acoustics and the descriptions of analogues on land. The increasing availability of Remotely Operated Vehicles (ROVs) now allows in-situ, close-up investigations in the marine environment. In this paper we present a novel methodology to obtain a detailed map of the headwall morphology of submarine landslides, including detail on vertical cliffs, overhanging strata and biological colonisation. Using a high-resolution multibeam system mounted on the front of a working-class ROV, rather than in a traditional downward looking configuration, we mapped three sections of a scallop-shaped headwall scarp which is part of the Rockall Bank Slide Complex, west of Ireland, as part of the SORBEH cruise funded by the Irish Government and the ERC CODEMAP project. The resulting 3D model provides insights in the build-up of the failed sequence, allowing advances in the understanding of rupture mechanisms. It can be combined with close-up video ground-truthing and carefully collected samples to create a complete picture of the headwall scarp.

Agadir Canyon is one of the largest submarine canyons in the World, supplying giant submarine sediment gravity flows to the Agadir Basin and the wider Moroccan Turbidite System. While the Moroccan Turbidite System is extremely well investigated, almost no data from the source region, i.e. the Agadir Canyon, are available. New acoustic and sedimentological data of the Agadir Canyon area were collected during RV Maria S. Merian Cruise 32 in autumn 2013. The data show a prominent headwall area around 200 km south of the head of Agadir Canyon. The failure occurred along a pronounced weak layer in a sediment wave field. The slab-type failure rapidly disintegrated and transformed into a debris flow, which entered Agadir Canyon at 2500 m water depth. Interestingly, the debris flow did not disintegrate into a turbidity current when it entered the canyon despite a significant increase in slope angle. Instead, the material was transported as debrite for at least another 200 km down the canyon. It is unlikely that this giant debris flow significantly contributed to the deposits in the wider Moroccan Turbidite System.

Multibeam bathymetry and chirp seismic reflection profiles collected using an autonomous underwater vehicle reveal the morphology and shallow seafloor structure of Tubeworm Slump on the flank of Monterey Canyon at an unprecedented resolution. The data show smaller subsidiary deformation above the headwall, on the headwall, within the sediment drape that covers the sole of the slide, and on the sidewall of Monterey Canyon below Tubeworm Slump. The AUV data indicate that the existing slump scar represents a composite of gravity-driven deformation generated by multiple failure events.

In this work, we sought to document how submarine mass-movements influence the submarine landscape and associated habitat distributions on the upper portion of the northern Ionian Margin (eastern Mediterranean Sea) between 200 m to greater than 1,000 m in water depth (w.d.). In this area, mass-wasting processes have created unique morphological forms that, in turn, have generated high diversity for edaphic and hydrogeologic conditions; and these areas are marked by the patchy occurrence of varying natural benthic habitats. Surficial or sub-surficial Mass-Transport Deposits (MTDs) were documented by seismic and high-resolution morpho-bathymetric data and displayed dense aggregation for detached blocks spread over 1,200 km

2

between 400 and 1,000 m in w.d.. Living Cold-Water Coral (CWC) communities populate the blocky region and form coral topped mounds. These habitats are important Vulnerable Marine Ecosystems (VMEs) that are exposed to human pressure in the deep sea. Through production of a detailed geomorphological map and an examination of published data on the extent and distribution of CWC communities in the area, we sought to document how comprehensive research into submarine slide topography should also take into account the peculiar characteristics of their biotopes.

Submarine geohazards threaten coastal communities and global economies. Submarine debris flows are the largest mass-wasting events observed on the Earth’s surface, comprising of up to 50 % of basin fill. Further insight can be gained into these important processes by understanding in-situ preconditioning factors that lead to slope destabilization. We examine two locations from the International Ocean Discovery Program data archive to determine how external effects on sediment properties compare between passive margins and active margins. We select representative passive margin (Amazon Fan) and active margin sites (Nankai Trough), and analyse peak shear strength, void ratio, and composition from the uppermost 100 m below seafloor. This depth corresponds to a depth range in which most submarine mass movements originate. However, it is not appropriate to directly compare shear strength and void ratio of samples from different settings due to differing stress histories, sedimentary composition, and consolidation properties. We focus on ideal locations on both margin types that have solely undergone one-dimensional burial, no diagenesis/cementation, and no unroofing. We find that active margin sediments exhibit an increase in shear strength when compared to their passive margin counterparts, while void ratio tends to be higher on active margins. We are currently conducting a focused lab program to better understand compositional effects and determine the intrinsic properties of each site to more definitively normalize the in-situ sediment profiles. Our results suggest a potential link between shear strength and margin seismicity.

A new small volume CPT calibration chamber with dynamically controlled boundary conditions has been built to improve the correlation between in-situ data and soil parameters. The sample volume in the new CPT calibration chamber has a diameter of 30 cm and a height of 54.5 cm. Therefore, it is possible to use reconstituted samples of limited quantity, e.g. from boreholes. The chamber is able to simulate large overburden stresses and overconsolidation ratios (OCR) up to 5 MPa. Horizontal, vertical and pore pressures are independently applied via syringe pumps while recording volume changes. All pressures are dynamically controlled and allowing stress, strain and mixed stress-strain stress boundary conditions BC1–BC5 to be enforced. The sample deformation is measured by circumferential laser triangulation sensors. In a first series of tests using Cuxhaven Sand, a 12 mm cone and BC1 conditions, CPT tip resistances reach steady state about a length of at least 10 cm. The corrected tip resistances and the inferred relative densities for Cuxhaven Sand differ substantially from previously established correlations, confirming the need for more advanced correction factors and relationships between CPT data and in-situ soil properties.

This study was initiated because underwater mass movements in Lake Mjøsa, Norway had caused utility pipeline breakages. Multibeam bathymetry, sub bottom profiler data and samples were acquired to allow morphological interpretations of the processes leading to the mass movements. The underwater slopes of the lake generally show gradients of 15–20°, but exceed 30° in places. The sediment thickness above acoustic basement interpreted from sub bottom profiler data show that the accumulation rate in the central lake basin is about 2 mm/year. Numerous channels are seen on the bathymetry as well as several slide scarps that are about 2 m high. The channels are interpreted to be caused by dense water cascading, probably induced by winter cooling. Calculations based on geotechnical tests of samples indicate that sediment layers in excess of 2 m have a factor of safety less than 1.5 on slopes above 30°. Triggering of slides may thus be spontaneous due to sedimentation, but may also be triggered by oversteepening due to erosional channeling.

Earthquake induced cyclic loading has the potential to destabilize submarine slopes either by liquefaction in coarse-grained deposits or by cyclic softening in cohesive sediments. Vibratory cone penetration tests (VCPTU) represent a new approach for the evaluation of cyclic softening in fine grained sediments. In the past, VPCTU were utilized to evaluate liquefaction potential of sands, but cyclic softening of fine-grained marine sediments has not yet been tested with VCPTU in situ. At the study site in Orkdalsfjord, mid Norway marine clayey silt deposits are interbedded with coarse silt and clay layers. Static and vibratory CPTU were performed down to 19 m penetration depth using the Geotechnical Offshore Seabed Tool (GOST) and in addition, two gravity cores were taken for cyclic triaxial testing and geotechnical index tests. From static and vibratory CPTU a number of coarse silt layers with a distinct drop in cyclic cone resistance were identified. Compared to surrounding finer sediments the coarse silt layers exhibited a higher potential for cyclic softening. This assumption is supported by cyclic triaxial tests on very coarse and surrounding medium-coarse silts, respectively, revealing a strong loss of cyclic shear strength in a controlled and documented stress-strain regime. This study highlights the potential for VCPTU as a promising tool to qualitatively evaluate the vulnerability of marine silts to cyclic softening. In combination with advanced laboratory tests these results are envisioned to help better identifying submarine slopes subjected to failure during earthquakes.

High-resolution geophysical data reveal the presence of several spatially-isolated, small-scale landslides along the gently dipping (~3–4°) upper slope off Vesterålen, Northern Norway. Dynamic slope stability analysis suggests that seismicity may be largely responsible for the occurrence of these slope failures. The landslides are clustered in two groups, with one group of parallel features with their headwalls in ~500 m water depths. The second group is found in ~800 m water depths.

We present first results of geotechnical

in situ

Cone Penetration Test (CPTU) data and TOPAS sub-bottom profiles collected during two cruises in summer 2013/2014. We obtained a total of six static CPTU profiles penetrating the top 20 m of soil. Three of these were taken across one of the landslide complexes (SL3) from the slide scar down to the depositional area. The other three are reference sites in the adjacent undisturbed areas.

The combination of geophysical and geotechnical data allows us to divide the well-stratified glacio-marine slope deposits into three different sediment-mechanical units, and reveals the occurrence of mechanically weaker zones (MWZ). These zones are interbedded by coarser layers with high values of cone tip resistance. The occurrence of sensitive fine-grained material may be responsible for the loss of strength in the deeper portion.

One-dimensional pseudo-static stability analysis attests that the Vesterålen slope is stable except for exceptionally large earthquakes, that induce a peak-ground acceleration (PGA) of 0.224 g or larger to the MWZ. The depth levels of the MWZ correspond well with the slide planes of the landslides.

We apply a novel micro-shear tester (μST) to investigate the shear flow behaviour of very small granular volumes (15 μl). We compare the results to standard ring shear devices which confirm that these small volumes are sufficient for granular flow analysis, and hence are helpful to identify the stable and unstable areas of very small granular quantities. Within the laboratory experiments we realised shear tests by using a torsional shear movement which was applied to a cohesive granular calcium carbonate. For the analysis of the shear behaviour the yield loci are used which were derived from the bulk shear and normal stresses. The comparison of the yield loci determined by μST and ring shear device shows only a minor deviation. A particle characterization of the calcium carbonate material according to the particle size distribution with a mean grain size of 5 μm using laser diffraction and determining cohesion forces on single particles with 34 nN using an atomic force microscope (AFM) is part of this study. Finally, we combine a X-ray computed tomography (XCT) with the μST to reveal changes in the microstructure and notice shear bands caused by the shearing process.

Wabush Lake is characterized by a nearly constant sediment input, resulting from the deposition of mine tailings. Five bathymetric surveys were conducted in this lake, in order to understand the sedimentation pattern over more than 12 years. Of the morphologies studied, submarine channel and knickpoints were traced out and documented. A physical model of Wabush Lake was also constructed in order to understand specific morphologies, such as the knickpoints.

As part of this study, it was observed that knickpoints migration could be explained, at least in the laboratory, by two mechanisms: landslide and erosion, and not only erosion as previously thought. This previous conclusion will be applied to some knickpoints found in Wabush Lake. Two cases are analysed: (1) a knickpoint where a tension crack is present and (2) another knickpoint that shows no sign of instability. It is found that, when an excess of pore water pressure is present due to rapid sedimentation, static liquefaction may occur at the head of a knickpoint leading to a localized slope failure.

Flow slides form a major threat to flood defences along coastlines and riverbanks in the Netherlands. Due to the uncertainties with respect to the process in combination with the severity of the consequences and costs for prevention measures, there is a need to improve existing models for prediction or occurrence of, and damage by, flow slides. One of the key questions to be answered is whether slope failure by a flow slide is caused by up-slope migrating breaches or by static liquefaction. Although fundamentally different mechanisms, both result in a flowing sand-water mixture or turbidity current that eventually redeposits on a gentle slope. Over the last decades numerical models have been developed for both mechanisms, based on flume experiments. Upscaling these experiments is complex, as scaling rules are different for the various processes involved. To evaluate the failure mechanism on a natural scale, validate numerical models and test new technology to monitor the occurrence of flow slides, a large, controlled field test was performed.

The test site was situated in the Westerschelde estuary, in the south-western part of the Netherlands (Fig. 24.1). Several flow slides of 10

5

–10

6

m

3

have occurred in this area in the past. In advance of the experiment, cone penetration tests and boreholes were performed on the test location. Pore water pressure sensors were installed in the sand. Triaxial tests and grain size distribution measurements were performed on collected soil samples.

The flow slides were initiated by means of steepening of the slope by dredging. Eventually several autonomously retrogressing flow slides were observed, running several hours over a maximum distance of about 100 m and resulting in a total displaced volume of several 10

3

m

3

of sand. During the test the evolution of the slope topography was monitored continuously by three multibeam survey vessels. This resulted in a full multibeam survey of the area almost every quarter of an hour and, assisted with other advanced instruments, enabling us to witness the early development of a flow slide.

Hovgaard Ridge is an 1800 m high bathymetric high in the Fram Strait, the only deep-water gateway between the Arctic Ocean and the other World’s oceans. The slopes of the ridge provide evidence of various types of sediment reworking, including (1) up to 12 km wide single and merged slide scars with maximum ~30 m high headwalls and some secondary escarpments; (2) maximum 3 km wide and 130 m deep slide scars with irregular internal morphology, partly narrowing towards the foot of the slope; (3) up to 130 m deep, 1.5 km wide and maximum 8 km long channels/gullies originating from areas of increasing slope angle at the margins of a plateau on top of the ridge. Most slide scars result presumably from retrogressive failure related to weak layers in contourites or ash. The most likely trigger mechanism is seismicity related to tectonic activity within the nearby mid-ocean fracture zone. Gully/channel formation is suggested to result from cascading water masses and/or from sediment gravity flows originating from failure at the slope break after winnowing on the plateau of the ridge.

Mass movements are a frequent sedimentary process in the marine realm, affecting both glaciated and non-glaciated continental margins. Here a 3D seismic data set from the North Sea Fan, NE Atlantic margin, is used to study internal architecture, external geometry and surface geomorphology of different types of buried sediment transport. We identify three mass transport deposits, at a depth of ca 100–1000 m below seabed, corresponding to the previously mapped Tampen (~130 ka) and M½re (~300 ka) slides and the Stad (~400 ka) Slide, identified in this study. These slides all eroded underlying sedimentary units and their surfaces include curvilinear ridges up to 20 m high and 10–15 km long. Locally, the slide surfaces also include rafted slide blocks, up to 200 m wide, 300 m long and 30 m high. We relate the curvilinear ridges and the slide blocks to submarine spreading. Intervals of glacigenic debris flows are identified between the mass transport deposits. These lens-shaped bodies are seen in plan to be flows that widen downslope and which are fed from 10 to 20 m deep, 50–200 m wide and >2 km long transport channels that extend from the paleo-shelf edge. Analyses of time slices suggest that such flows may have been operating on the North Sea Fan for the last ~1.1 million years.

The NE portion of the Gela Basin (Strait of Sicily) shows evidence of multiple mass wasting events of predominantly translational character. In this context, recent investigations implicate volcanoclastic layers as key stratigraphic surfaces acting as preferential planes of failure. We present an integrated analysis of a representative sedimentary transition from overlying homogeneous background sedimentation of silty clay to a volcanoclastic layer. A high-resolution CT scan and three drained direct-shear laboratory experiments from a 20 cm whole-round section (~28.2 mbsf) allow the delineation of the role of this volcanoclastic layer in the framework of slope stability and failure initiation. The mechanical results indicate a general strengthening of the material with increased volcanoclastic content. Tendency for failure is expected to be highest within the silty clay due to relatively lower shear strength and strain-weakening behaviour, which promotes progressive sediment failure. In contrast with recent findings, this suggests that volcanoclastic sediment would not act as a weak layer. However, the volcanoclastic layer exhibits significant mesoporosity (i.e., fracturing) and may therefore host large volumes of fluid. Temporarily undrained conditions, for example during seismic activity, could transiently elevate fluid pressures and thus reduce the material shear strength below that of the surrounding silty clay. Such a weak layer may preferentially form along the interface of fractured volcanoclastic material and relatively impermeable silty clay, where differences in material strengths are lowest.

The formerly glaciated continental margin off Norway has experienced a relatively large number of submarine landslides of varying sizes, volumes, and ages originating from contourites deposited on the continental slope. We review: (i) the origin and occurrence of weak layers involved, (ii) sediment disintegration and initial flow, and (iii) sediment run-out and resulting deposits. The following major knowledge gaps, critical for further progress in this field are identified: information on lithology and sediment properties of material recovered from below the depth of conventional coring, and in situ measurements of sediment physical properties at the depth of weak layers.

Lake Ohrid (LO), a transboundary lake shared by Macedonia and Albania on the Balkan Peninsula, is not only considered to be the oldest lake in Europe (~2 Ma) but has a long and continuous sedimentary history. An advantage at LO is the availability of hydroacoustic data sets of good quality covering the entire lake basin. The tectonically formed basin is filled with thick undisturbed sediments. However, the overall internal structure of LO is characterized by numerous faults, clinoform structures, and several Mass Transport Deposits (MTDs). By using a seismic chronology model (SCM) correlating seismic reflector packages with Marine Isotope Stages (MIS) we estimate the occurrence of the deepest MTD detected in the southern basin at the transition of MIS9 to MIS8 (~300 ka) defining the onset of the sliding history in LO that is still ongoing today. In general, MTDs are widespread within the basin but they do cluster at active faults. Two large MTDs occurred in the early MIS7 (~230 ka, ~220 ka) and after a quiesence period of about ~70 ka two additional large MTDs have been detected in the late penultimate glacial period MIS6 (~150 ka, 130 ka). MIS5 seemed to be another quiet period with respect to mass wasting. In the younger sedimentary history mass movement is a common process with several large and mid-sized deposits mapped at all stratigraphic levels. The youngest slide deposits are estimated to occur within the last 2,000 years. The main outcome of this paper is a model for the spatial and temporal distribution of mass wasting for Lake Ohrid.

The Pianosa Ridge forms the eastern flank of the Corsica Trough in the Northern Tyrrhenian Sea: it is the site of preferential accumulation of contourites and Mass Transport Deposits (MTDs). Along the Pianosa Ridge, 11 MTDs with a total volume of 6.5 km

3

were identified. These MTDs are distributed in three areas: (A) one small MTD associated to canyon flank destabilisation in the northern part of the study area; (B) six intermediate size MTDs in the central area; (C) four MTDs of larger size (up to 2.62 km

3

) to the south, including the Pianosa Slump, which is the most recent MTD in this area (aged at 42–50 kyr BP) and analysed in more detail. The main factor controlling the formation of MTDs in areas A and B seems to be steep slopes associated to erosion and heterogeneous sedimentation caused by bottom currents, respectively. In contrast, multiple factors may control slope instability in the zone where the largest MTDs took place (area C): the incision generated by contour currents, the presence of coarser layers in contourite drifts that may accumulate gas and the location of normal faults near the headwall.

East Lake, located at Cape Bounty (Melville Island, Canadian High Arctic), was mapped using a high-resolution swath bathymetric sonar and a 12 kHz sub-bottom profiler, allowing for the first time the imaging of widespread occurrence of mass movement deposits (MMDs) in a Canadian High Arctic Lake. Mass movements occurred mostly on steep slopes located away from deltaic sedimentation. The marine to lacustrine transition in the sediment favours the generation of mass movements where the underlying massive mud appears to act as a gliding surface for the overlying varved deposits. Based on acoustic stratigraphy, we have identified at least two distinct events that triggered failures in the lake during the last 2000 years. The synchronicity of multiple failures and their widespread distribution suggest a seismic origin that could be related to the nearby Gustaf-Lougheed Arch seismic zone. Further sedimentological investigations on the MMDs are however required to confirm their age and origin.

Two Pleistocene mass transport deposits (MTDs), with volumes of thousands of km

3

, have been identified from multi-channel seismic data in the abyssal plain at the front of the Barbados accretionary prism. Estimated sediment volumes for these MTDs are likely underestimated due to limited seismic coverage. In this work, we suggest that these MTDs are comparable in size to large submarine landslides as reported in the literature. These MTDs lie on the vicinity of two major oceanic ridges, the Barracuda Ridge and the Tiburon Rise. It is also suggested in this work that the MTDs come from seismicity associated with the formation of the Barracuda Ridge or the Barbados accretionary prism; however, triggering mechanisms involved in their formation remain uncertain. The present study discusses the potential causal factors accounting for the formation of these MTDs.

Gravity-induced sediment transport processes and associated deposits in deepwater settings have been investigated for decades. However, the role of deepwater bottom currents as a preconditioning factor for mass-transport deposits (MTDs), as well as their capacity to redistribute sediments, is poorly understood. MTDs form an important component of the stratigraphic column within continental margins, and these units are often found in association with current-generated sediment waves (CGSWs). Our analysis of geophysical data from the Gulf of Mexico (GOM) indicates that there are different types of stratigraphic associations between CGSWs and MTDs in this region. Comparison of our data with existing studies where the relationship between MTDs and CGSWs has also been reported seems to suggest that there might be causal mechanisms and pre-conditioning factors that could explain this coupling. In this study, we discuss these relationships, taking into account the geologic and paleo-environmental conditions under which these units were deposited. Two main factors appear to control the stratigraphic coupling of MTDs and CGSWs in the study area: (1) fine-grained CGSWs act as shear basal surfaces that apparently precondition the slope for mass wasting events, and (2) contour-following bottom currents erode the lowermost slope, destabilizing its base and increase the likelihood of mass-wasting events.

The French alpine foreland area has been struck by several earthquakes with magnitudes above 5 on Richter scale in recent history. In this paper we document the regional impact of historical and Holocene earthquakes based on the identification of mass wasting deposits in glacial lakes at different settings. Lake Le Bourget and Lake Paladru are situated at low elevations (respectively 231 m–492 m) and Lake Blanc Huez is located at 2500 m altitude. Through the integration of high-resolution acoustic profiles and accurately dated sediment samples from cores, recent coeval mass wasting deposits in each lake were correlated with nearby historical earthquakes, whereas coeval mass wasting deposits around 5200 cal BP and 9550 cal BP in these three lakes were correlated to regional earthquakes events.

A new study of the Israeli Mediterranean continental slope provides an understanding of the interaction between submarine landslides, fault scarps, and subsurface evaporites. Faults and landslides interact in the northern part of the studied continental slope where fault scarps rupture the seabed. In this area landslides are thought to be triggered by over-steepened fault-scarps and are observed to cover older fault scarps, or be cut by younger faults. These variable cross-cutting relationships indicate a multi-phase history in which landsliding and faulting both post-date and pre-date one another. Isopach maps of the Messinian evaporites further reveal that fault scarps are mainly found along a slope-parallel belt where the underlying salt layer is 150–500 m thick. We suggest that this rather thin sequence of the Messinian evaporites associated with faulting serves as a localized detachment zone for the overlaying strata. We argue that the multi-phase and spatially variable association of landslides and faults reveal a highly dynamic continental slope, which may be still active in the present day.

Lake Éternité, located between the Upper Saguenay Fjord and the St. Lawrence River has registered many submarine slides caused by at least one earthquake. Landslides are mostly rooted in the gyttja (Holocene sediments). Mapping of landslides revealed a total of 128 scars over an area of only 3.2 km

2

. A larger proportion of the landslide scars are located on the SE and NW facing slope which may support an epicentre location for the strongest earthquake (1663?), to the NW or NE of the lake. The preliminary numerical analysis of the site effects caused by topography on local preferred seismic amplification is not conclusive enough to support the observations made for landslides. Associating landslides to specific earthquakes will only be possible with further investigations, including coring of various features including rupture surfaces. The study also revealed interesting slide morphologies developed in homogeneous sediments, providing excellent examples for future modelling of similar events.

Large-scale landsliding is a common process in the Kumano Forearc Basin of the Nankai Trough accretionary prism. We use a 3D seismic data volume to map the seafloor reflection, which shows that there are two surficial landslides, one rotational slump ~3.4 km wide, 1.8 km long and 150 m thick and one disintegrative slide that has left a seafloor scar ~ >3.65 km wide, 2.6 km long and ~200 m deep. We see no evidence for any deposits related to the latter in our data, so the entire mass must have been transported as debris flows/turbidites outside the area covered by 3D data. The slump failures occurred along a bedding plane that dips ~5–7° landward, but the disintegrative landslide has a gently-dipping base and is associated with steep normal fault scarps. Several large subsurface mass-transport deposits (MTD)s are mapped in the 3D seismic data – all have slid along single landward-dipping bedding planes. Their bases range in depth from 140 to 700 m below sea floor (mbsf). The thickest MTD is ~6.5 km

2

× 155 m thick, encompassing a volume of ~1.0 km

3

. The three other large MTDs range from 0.3 to 0.6 km

3

in volume. The toes of the MTDs are imbricated, and the imbricate structure, as imaged in continuity displays, is aligned parallel to the slope. Many less extensive, thinner (<20 m thick) MTDs are also present in the Kumano Basin. Regional seismic-stratigraphy and age-constraints on MTD-correlative seismic reflections drilled at IODP drill Sites C0009 and C0002 reveal that four of the investigated MTDs are younger than 0.3–0.44 Ma, three are 0.44–0.9 Ma, and three others are between ~0.9 and 1.24 Ma.

This paper employs numerical modeling to investigate the ability of simplified procedures based on sliding block methodology to provide a reasonable characterization of the seismic displacements of a deepwater slope. Earthquake-induced permanent shear displacements obtained from dynamic finite-element analyses of a deepwater slope subjected to various input base excitations are presented and compared with the seismic displacements predicted by two relatively recent simplified procedures available in the literature. The numerical outcomes indicate that the simplified procedures may be overly conservative in evaluating the earthquake-induced permanent shear displacements along the sliding surface of deepwater slopes. Based on the limited displacement data set developed under the present study, correlations aiming at improving the predictive capability of the selected simplified procedures in respect to evaluation of seismic performance of deepwater slopes are provided.

Recently acquired 2-D seismic data from offshore North Carolina provides images of salt diapirs and landslides in the Carolina Trough that give insight into the interaction between slope sediments and intruding salt from below. The best example of this is the Cape Fear Slide Complex in which at least two salt diapirs are surrounded by the lower headwall of the slide. Here, we present seismic images that were collected for the Eastern North American Margin Community Seismic Experiment. We describe the morphology of the slide and diapirs in order to infer rates of salt rise. We have tentatively estimated a post-slide growth rates of 517 m per million years (m/Ma). However, as the analysis continues, it is possible this estimate will change. This research provides significant insight into the interplay of salt and slope failure processes in an ocean basin setting.

The potential of gas hydrate systems to play a role in submarine slope failure has been well-documented since the late 1970s. Several conceptual models exist for how the gas hydrate-free gas system might weaken submarine sediments, but there is no definitive evidence for gas hydrate-related processes being the primary cause of a particular submarine slope failure. We present a review of coincident gas hydrates and submarine slope instabilities on New Zealand’s active margins. The examples we show represent different failure modes in a range of slope environments, including the upper continental slope and tectonic ridges, with the common factor being that the base of gas hydrate stability approaches the seafloor in these regions. We synthesise several proposed sediment weakening mechanisms and draw comparisons to other global models for gas hydrate-related slope instability. This contribution highlights diverse influences that gas hydrate systems could have on submarine sediment strength, while acknowledging gaps in our understanding of the potential role of gas hydrates, free gas and fluid flow on slope stability.

Autonomous underwater vehicles have been used to characterize Eel Slump, a slide scar located south of Eel Canyon, California. The presence of a well developed dendritic network on the headwall with gullies tens of meters deep, thick sediment drape cover on the slide scar sole, and the absence of fresh surfaces on the scarp suggest that the mass failure(s) that produced this feature did not take place in the recent past. Thermogenic oil and gas emanating from a large mound in the sole of the slide scar were sampled with a remotely operated vehicle. Other distinctive morphologies observed from the seafloor of the slide scar indicate fluid seep has occurred at multiple sites within the slide scar sole.

Submarine spreading is a type of mass movement that involves the extension and fracturing of a thin surficial layer of sediment into coherent blocks and their finite displacement on a gently sloping slip surface. Its characteristic seafloor signature is a repetitive pattern of parallel ridges and troughs oriented perpendicular to the direction of mass movement. We map ~30 km

2

of submarine spreads on the upper slope of the Hikurangi margin, east of Poverty Bay, North Island, New Zealand, using multibeam echosounder and 2D multichannel seismic data. These data show that spreading occurs in thin, gently-dipping, parallel-bedded clay, silt and sandy sedimentary units deposited as lowstand clinoforms. More importantly, high-amplitude and reverse polarity seismic reflectors, which we interpret as evidence of shallow gas accumulations, occur extensively in the fine sediments of the upper continental slope, but are either significantly weaker or entirely absent where the spreads are located. We use this evidence to propose that shallow gas, through the generation of pore pressure, has played a key role in establishing the failure surface above which submarine spreading occurred. Additional dynamic changes in pore pressure could have been triggered by a drop in sea level during the Last Glacial Maximum and seismic loading.

Significant venting of methane gas has been observed on the upper continental slopes of the Arctic Ocean, coinciding with the landward limit of the gas hydrate stability zone. It has been inferred that the methane gas venting is related to the dissociation of methane gas hydrate induced by the unprecedented Arctic warming that has occurred over the last 30 years. Historically, the influence of hydrate dissociation on sediment stability was considered in terms of hydrate dissociation increasing pore pressures, reducing effective stress and therefore sediment strength, leading to slope failures along these over pressurized layers. Recent evidence has shown that gas hydrate readily forms in clay-rich sediments as fracture-filled near-vertical veins, which upon dissociation gives rise to sediments exhibiting high water content, high sediment compressibility and very low shear strength. Thus gas hydrate dissociation in fine-grained sediments may lead to significant slope instabilities, which at present is poorly understood. Slope stability analyses carried out considering the potential influence of hydrate dissociation induced by warming of the Arctic slope suggest that instabilities are likely, with failures ranging from surficial sloughing to deep-seated failures. This paper highlights the importance of understanding the sediment processes in clay soils, and how dissociation of gas hydrate can induce instabilities through sediment softening and generation of excess pore pressure.

Mass-transport deposits (MTDs) are well developed on the slope of the Eocene Sobrarbe delta (Ainsa Basin, Spanish Pyrenees) and are studied in order to improve the understanding of soft sediment deformation in MTDs. The five examples illustrate the distinction between the extensional zone and the compressional domain. The upslope domain is illustrated by a set of three stacked nummulite-rich slid layers. They are deformed by load structures suggesting density inversion and sinking of nummulite gravel down into the underlying silty-shaly material. Post-sliding deformation is evidenced from small-scale roll-overs, neptunian dykes and syn-sedimentary normal faults. Slow extensive deformation has continued in the substratum of the scar after the main sliding episode. Soft-sediment deformation is less obvious in large displaced blocks where deformation is limited at the periphery of the blocks. Imbricate thrusts, formed on the side of a sliding layer, illustrate structures related to displacement. Striations on the thrust planes indicate that sliding occurred along the strike of the imbricate slices. The km scale Castellazo outcrop shows slump folds resting over a basal debris flow.

The transition from Triassic to Jurassic strata along the passive northern margin of the Neotethys elucidates typical drowning successions of huge Triassic carbonate platforms. Drowning goes parallel with synsedimentary block tectonics that was initiated by rifting at the southern margin border. An intriguing case study is exposed along the saw-cut wall sections of quarries around the village of Adnet close to Salzburg. The deeply submerged, inherited relief of a drowned reef mound gave rise to pronounced Liassic facies differentiation, i.e. (1) deposition of grey spiculitic cherty limestone and marl beds in the former shelf basin, and (2) red nodular limestones, and red condensed limestones rich in ammonites and Fe-Mn crusts over the slope and top of the former reef mound. Faulting, tilting and submarine erosion of Hettangian drift deposits at the lower slope was followed by repeated down-slope gliding, shearing and multiple opening of fissures with different generations of sediment infill. Renewed tectonics from Late Pliensbachian to Middle Toarcian created deep reaching vertical fissures and triggered multiple mass flow events. At the upper and middle slope the so-called Adnet Scheck breccia, which is a special debrite deeply eroding and incising into well-bedded condensed hemipelagic limestone strata, was deposited. Further down-slope the Scheck breccia evolves into more matrix-rich nodular breccias. Basin sections reveal intercalations of mudflow deposits and were affected by various magnitudes of sliding and mass flow events forming complex mass transport deposits.

In this study we combine observations and analytical data from large-scale (10–100s of m-thick and 100 m

2

-extensive), siliciclastic and carbonate MTD/MTCs belonging to the Oligocene – Miocene foredeep and wedge-top successions of the Northern Apennines and the Paleocene – Eocene Friuli basin of the northwestern Dinarides (Italy and Slovenia), to discuss the deformation processes critical to the emplacement of submarine landslides. We focus on the identification of meso-scale structures, used as diagnostic kinematic indicators of local paleo-transport directions. These structures, represented by linear-planar and complex-shaped elements such as tabular shear zones and detached slump-type folds, are the product of ductile-plastic deformation developed at relatively low-confining pressure that involves water-saturated, un- to poorly-lithified sediments, along with liquefaction/fluidization processes. Their final appearance is thus mainly controlled by the mechanical-rheological behavior of deformed sediments, and eventually by tectonic fabrics inherited from deeper structural levels of deformation. Due to this parallelism these structures have been termed and classified accordingly. They reflect strain partitioning due to differential movements within the slide mass, which is in turn controlled by the overall landslide typology. Due to the parallelism with classified tectonic structures and structural associations, we have thus redefined and classified accordingly meso-scale kinematic indicators in ancient MTD/MTCs.

This study aims to resolve process-facies links at both bed and environmental scales for the channel lobe transition zone (CLTZ). Data comes from existing experimental and modern CLTZ studies and from new outcrop studies. The experiments show that the CLTZ architecture of supercritical turbidity currents is complex and different from their counterparts where flows are subcritical throughout. Supercritical CLTZ’s are characterised by erosive channels formed by supercritical turbidity currents, by offset stacked lobes deposited from subcritical turbidity currents and by hydraulic jump related mouth bar deposits and upslope onlapping backfill deposits at the down slope end of the transition zone. Erosive channels and backfill features can be resolved by high resolution seismic data, yet evidence for supercritical flow must come from facies analysis of core data. Outcrop examples of the CLTZ from the Tabernas submarine fan (SE Spain) and the Llorenç del Munt deep-water delta slope (N. Spain) are used to establish such links between seismic scale architecture and facies recognised in cores. The outcrops described here were mapped as transition zone, and show 100 m sized, spoon-shaped scours filled with sediment containing sandy to gravelly backsets up to 4 m in height. Their facies and architecture is indicative of deposition by hydraulic jumps, can be recognized from cores, and is a good proxy for further predicting CLTZ architecture constructed by supercritical turbidity currents.

A 3.4-m thick condensed section enriched in foraminifera formed the final detachment horizon of a retrogressive submarine landslide, in the Ursa Basin, northern Gulf of Mexico. The high concentration of foraminifera produces a high porosity (up to five porosity units) layer distinct from the background clay. Integrated Ocean Drilling Program Expedition 308 Site U1324 cored and logged this layer. We conducted a sedimentological analysis on 31 samples across this zone and the overlying and underlying background clay. CT images show that foraminifera are individuals dispersed within the clay, unbroken, and have retained a significant amount of intraskeletal void space. The assemblage is expected for this time interval in the Late Pleistocene (~24 kya). We interpret the layer is a result of a pause in terrigenous sedimentation. The condensed section was a preferred detachment horizon but only minimal sliding occurred before further movement ceased. One possible mechanism to explain this is the presence of foraminifera results in a dilational shear strengthening behavior, which arrested movement. Further work will be required to test this, however. On a broader scale, condensed sections with abundant microfossils, may play a key role in landslide mechanics because they can alter the shearing properties of the background material.

The preservation of large, relatively undeformed blocks is a characteristic feature of mass transport deposits (MTD). We examine a well-exposed succession at Cerro Bola in La Rioja Province, western Argentina, which comprises mid to late Carboniferous fluvio-deltaic sediments, turbidites and MTD’s. The main MTD, which is up to 180 m thick and crops out over 8 km, is characterized by allochthonous sandstone blocks that range in size from metres to 100s of metres in length, and are up to tens of metres in thickness. Blocks are preserved throughout the entire MTD, but are typically larger and much more abundant towards its base where they comprise up to ~30 % of the unit, and become progressively smaller and less frequent upward. Blocks were eroded from the underlying unlithified deltaic sands, and incorporated into the MTD during its transport and emplacement, resulting in local gouges and grooves in the substrate along the basal contact of the MTD. Sandstone blocks are interpreted to have undergone progressive abrasion and fragmentation as they rose through the MTD, thereby creating smaller blocks in the upper parts of the unit. We suggest that buoyancy-driven rise combined with the synchronous fragmentation of sandstone blocks that are entrained within a finer matrix, provides a mechanism for the observed distribution of blocks during overall downslope transport of the MTD.

This study presents the first version of a GIS catalogue of submarine landslides affecting the Spanish Continental Shelf pursuant UNCLOS (United Convention for the Law of the Sea) that comprise continental-type margins (Atlantic and Mediterranean) as well as hot-spot type volcanic islands and seamounts (Canary Archipelago). This first version compiles a total of 223 submarine landslides within an area of 1.5 million square kilometres. This catalogue, developed in a geographic information system, compiles information of each submarine landslide such as the name, location, typology, age, volume, source, dimensions (width, elongation), minimum and maximum height, minimum and maximum seafloor slopes. As a direct application and advantage of this digital catalogue, a first approach to a susceptibility map has been elaborated using GIS analysis tools taking into account the following available factors sourced from the geological maps of the Spanish continental margins: (i) faults map, (ii) active faults map, (iii) earthquakes density map, (iv) seafloor slope map and (v) seabed composition map. The potential and difficulties of this GIS catalogue of landslide as a first step in the submarine risk analysis are discussed. This first version of the GIS catalogue is conceived as the origin of a national submarine landslides database within the European network, which is being built as part of the EMODNET-Geology project.

Due to their potential volume and speed, large submarine landslides can generate destructive tsunamis or damage expensive seafloor structures. Understanding their timing is therefore important for hazard assessments; however, dating large numbers of landslides close to their origin is logistically difficult. Previous landslide studies are typically limited to fewer than ten observations of ages. To address this we analyse extensive, continuous and long-term turbidite records from four deep-sea basins which are interpreted to be the distal deposits of large, disintegrative landslides. Our records include sufficient numbers of turbidites (

N

 = 151–1571) for robust statistical analysis of long-term controls on event timing and testing for relationships with triggering mechanisms such as earthquakes, sea level, climate change and volcanic activity. We explore statistical methods developed by medical, economic and biological disciplines and show how they can be applied to analysis of submarine landslide frequency and triggering. Frequency analysis of field data reveals two different distribution forms for landslide recurrence – exponential and log-normal. We discuss possible individual and combined effects of controlling factors that result in these distributions. Rescaled range and Gaussian finite mixture models determine whether and how landslides are clustered in time. Parametric Generalised Linear Models and non-parametric Proportional Hazards Models are used to test for the significance and influence of different variables and their rate of change. We demonstrate the value of unusually detailed long-term landslide records, and show how statistical analysis provides quantitative inputs for future hazard assessments, landslide-climate studies and understanding the tempo of deep-sea sediment flux.

An understanding of submarine mass movements is of great importance to the hydrocarbon industry due to the risk they pose to sea floor infrastructure. Technological developments in deepwater surveying methods have produced datasets of the sea floor that rival the best terrestrial ones; however, the study of submarine mass movements remains poorly-developed. Multivariate statistical analysis has a well-established track record for producing quantitative estimates of associated risk for terrestrial landslides and given the often homogenous nature of sea floor sediments, a morphological control on mass movements seems viable. In this study, we perform a statistical analysis on an inventory of shallow slab slides in the West Nile Delta to identify morphometric controls on failure. We find that slopes with planar plan curvature and slope angles <6° account for approximately 95 % of observed landslides and that, beyond this, increasing plan concavity stabilises submarine slopes. This presents a foundation to ultimately reconcile geomorphological observation with geotechnical modelling, and provide additional insight on the controls on submarine instability.

First-order, second-moment (FOSM) formulations are useful tools for assessing uncertainty in GIS based submarine slope stability models. In the simplest applications, variables are assumed to be uncorrelated. In some cases, however, correlation among variables may be significant enough to require inclusion. Correlations among variables can be incorporated by creating an empirical covariance matrix and combining it with analytically derived expressions for partial derivatives of the factor of safety equation. Example calculations show that ignoring correlated variables over-predicts the probability of sliding for gentle slopes and under-predicts the probability of sliding for steep slopes, with small differences for moderate slopes. GIS-based application is illustrated using a hypothetical example motivated by an actual deepwater geohazard assessment, showing areas in which the use of uncorrelated rather than correlated variables over-predicts the likelihood of instability.

This study investigates how progressive oversteepening and fault kinematics impact on slope failure initiation and subsequent landsliding along subsiding basin flanks using 2D DEM simulations. We use large assemblages of granular particles to simulate the deformation behaviour of slope sediments with varying peak strength. Sediments with high peak strength deform preferentially on major faults and produce a stepped topography and a stable slope in the long-term. Mass failures in these sediments occur as large, compact slides of short run out. In contrast, slopes with lower peak strength deform diffusely and present large numbers of faults that fail frequently and maintain the slope at its critical angle of inclination. The resulting slope topography is smoother and laterally more elongated. These differences in mass movements are governed by (i) characteristic fault patterns, and (ii) repeated oversteepening during ongoing basin subsidence, which is an important prerequisite for failure initiation. Our experiments indicate quantitatively that the failure distribution, dimension, and transport mechanism, as well as the recurrence rate of landslides are essentially controlled by the peak strength of the failed material.

Numerical modelling of submarine mass movements is often used to estimate gravity mass flow runout distances, velocities, and the final shape of the sediments in offshore geohazards studies. This paper proposes an approach for the use of numerical models in a practical way, based on calibration against back analyses of known events, to obtain meaningful and reliable results. The proposed approach consists of estimating correction factors that quantify the approximations incurred by the numerical modelling, assuming that the input anchoring facts (i.e. geometrical, geotechnical and rheological information) are based on reliable information, therefore the difference between field evidence and simulations is merely due to the limitations of the numerical model. The approach is exemplified in the paper by simulating submarine debris flows, focused on runout distances, using the numerical model BING with the bilinear rheological model. The anchoring facts are obtained from a database of seafloor slope stability and empirical correlations. The results show the need to assess the level of uncertainty of the assumed anchoring facts in order to narrow the range of the proposed correction factors and use them for predicting runout distances in practical applications. At present the estimated correction factors range from 0.60 to 1.10.

Rapid emplacement of a mass via pyroclastic flows, or edifice failure, generates volcanic tsunamis. Physical modelling demonstrates that the efficiency of tsunami generation is influenced by the angle the mass enters the ocean. Efficiency decreases with increasing slope angle from 20° to 60°, before increasing to a maximum at 90°, which corresponds to a mass falling directly into the ocean without interacting with the slope (impact tsunami). Further, in the case of surging pyroclastic flows or regressive failures, successive closely spaced events may generate larger tsunami waves than a single event of comparable volume.

It is difficult to assess if physical model results are meaningful for real world tsunami events due to limited observational data. Two New Zealand palaeo-events – pyroclastic flows from Mt Tarawera and edifice failure at Whakaari (White Island) – can be linked to tsunami deposits, which constrains numerical simulations of the source mechanisms. The Mt Tarawera event involved multiple pyroclastic flows entering a lake during the AD 1314 ± 12 Kaharoa Eruption. The interaction of multiple closely spaced pyroclastic flows is necessary to generate the 6–7 m maximum wave height inferred from near source tsunami deposits. Tsunami deposits in the Bay of Plenty, dated to 2962 ± 52 BP, are consistent with edifice failure at Whakaari. In this case a single event with a volume of 0.23 km

3

is sufficient to account for the tsunami deposits. Hence, if the failure was regressive, the successive stages were sufficiently close together to be indistinguishable from a large single event.

Clay-rich landslides commonly involve retrogressive mass and momentum release mechanisms. Motivated by the retrogressive behaviour of major landslides offshore Norway, previous studies have demonstrated substantial effects of the release rate on the generation of the tsunami. However, the few existing models are limited to overly idealized conditions. In the present study, we explore further the wave generation due to a continuous retrogressive landslide model, quantifying the effects of the wave model, landslide configuration, and the continental slope. In the present examples, we find that the landslides involve large accelerations that may be crucial for tsunami-genesis. Tsunami footprints due to individual short blocks comprising the landslide are smeared out by dispersion. Keeping landslide material properties constant, we investigate mobilised landslide mass and maximum tsunami crest elevations for three different slopes: 1°, 1.5°, and 2° respectively. In the present examples, the smaller volume landslides are stronger tsunami generators than the larger ones because they are situated in shallower water, thereby clearly demonstrating the importance of the water depth on the tsunami generation.

A tsunami landslide which caused hundreds casualties and lots of damage took place on Lembata Island in 1979. In order to understand the characteristics of the landslide mechanism, a field survey was conducted in 2013 which sampled both the origin soil and landslide material, and the water from hotspring around the landslide site. The physical properties of the soil were obtained and show that the origin soil has dominantly coarser grain than the landslide material (80.5 % coarser grain compared to 11.8 % coarse grain respectively) which indicates that the soil has been altered to be finer and softer. Hot spring analysis determined that the mineral content of the water was 99.48 % SO

4.

This shows that magmatism process are involved which caused the soil to become acidic and may have fragilised the system. Results of X-ray Diffraction Mineralogy Analysis (XRD) show that the origin soil is composed of minerals of cristobalite, quartz, and albite, while the landslide material consists of clay minerals such as quartz, saponite, chabazite, silicon oxide and coesite which are typical minerals in a hydrothermal environment. Based on these results, it can be concluded that the area is influenced by an active geothermal system that could be the main source mechanism behind this disastrous event.

The Cook Strait Canyon of central New Zealand was identified as a priority area to quantify landslide-generated tsunami hazard in a national study in 2005. Therefore the canyon system has seen increasing research interest over the last decade. Landslide scars have been mapped throughout the whole of the Cook Strait Canyon area and analysis of landslide morphology demonstrates that the majority of landslides have some dependence on the topography of the canyon system. Axial downcutting destabilising lower canyon walls is proposed as the principal factor preconditioning slopes for failure. The canyons occur in an active tectonic environment and earthquakes are inferred to be the overriding failure triggering mechanism.

To develop an underpinning magnitude frequency model for probabilistic landslide-tsunami hazard assessment we have created a Monte Carlo based framework to assess the spatial distribution earthquake triggered slope stability within the canyon. The framework is object-oriented and the individual components required to calculate slope stability are implemented in a modular fashion. We drive the model using a long term synthetic earthquake catalogue based on known earthquake parameters for upper-plate and subduction zone fault sources. It is using an empirically derived landslide volume distribution for the Cook Strait Canyon. The model predicts about 1.35 potential slope failures in the Cook Strait Canyon over a period of 500 years with a volume exceeding 0.1 km

3

.

Evidence of previous submarine mass failures in the form of excavation scars has been widely documented in the Cook Strait Canyons of New Zealand. Recent bathymetry surveying has identified a well-defined submarine landslide scar and its associated debris deposit on the northern slope of southern Hikurangi Trough. The newly acquired multi-beam data allowed determination of the location and extent of the deposit, estimation of its volume, as well as reconstruction of both the pre-failure bathymetry and the initial state of the mass failure. A dynamically coupled two-layer model was used to numerically investigate this submarine debris avalanche and its resulting tsunami impact on the coasts of central New Zealand. The modeling results show a fairly good overall agreement with the observed debris deposition and also suggest that tsunami associated with the debris avalanche quite possibly inundated the coasts of central New Zealand, with maximum run-up elevations of between 3 and 5 m in several nearby locations.

On April 15 2014, a landslide occurred on the east shore of a lake in the municipality of Lac-des-Seize-Îles, about 100 km north of Montréal, Québec, Canada. The coastal landslide, with a length of 94 m and a width of 55 m, had a volume of about 30,500 m

3

of sandy to gravelly till. It was likely triggered by a significant amount of water infiltration caused by a heavy rain coincident with an accelerated snow melt. The displaced mass spread on the lake floor and triggered a tsunami that broke and partly lifted the ice cover. Water and ice damaged several seasonal residences and boathouses on the periphery of the lake, in an area extending 450 m north and 500 m south of the landslide debris location. Observations on aerial photographs taken shortly after the event revealed the existence of radial structures on the ice cover in the affected area resulting from the evolution of the tsunami wave. Investigations carried out on east and west shores showed evidences of net marks on tree trunks explained by a maximum inundation height which can be as much as 1.8 m directly in front of the landslide position on the west shore.