Douglas G. Masson

Submarine landslides: processes, triggers and hazard prediction

Huge landslides, mobilizing hundreds to thousands of km3 of sediment and rock are ubiquitous in submarine settings ranging from the steepest volcanic island slopes to the gentlest muddy slopes of submarine deltas. Here, we summarize current knowledge of such landslides and the problems of assessing their hazard potential. The major hazards related to submarine landslides include destruction of seabed infrastructure, collapse of coastal areas into the sea and landslide-generated tsunamis. Most submarine slopes are inherently stable. Elevated pore pressures (leading to decreased frictional resistance to sliding) and specific weak layers within stratified sequences appear to be the key factors influencing landslide occurrence. Elevated pore pressures can result from normal depositional processes or from transient processes such as earthquake shaking; historical evidence suggests that the majority of large submarine landslides are triggered by earthquakes. Because of their tsunamigenic potential, ocean-island flank collapses and rockslides in fjords have been identified as the most dangerous of all landslide related hazards. Published models of ocean-island landslides mainly examine ‘worst-case scenarios’ that have a low probability of occurrence. Areas prone to submarine landsliding are relatively easy to identify, but we are still some way from being able to forecast individual events with precision. Monitoring of critical areas where landslides might be imminent and modelling landslide consequences so that appropriate mitigation strategies can be developed would appear to be areas where advances on current practice are possible.

A giant landslide on the north flank of Tenerife, Canary Islands

The extent of mass wasting along the north flank of Tenerife has been mapped using swath bathymetry, GLORIA side-scan sonar, and 3.5-kHz echo sounder data. The marine surveys show that, north of Tenerife, a giant landslide is exposed over an area of 5500 km2 of the seafloor, more than twice the surface area of the island. The landslide truncates an older ridge and valley topography that is associated with the shield building basalts on Tenerife. We interpret the ridge and valley topography as the result of subaerial erosion. The landslide is estimated to have a length of 100 km, a width of up to 80 km, and a volume of about 1000 km3. It extends onshore into the Orotava and Icod valleys which have been interpreted as of landslide origin. K-Ar dating of basaltic flows in the steep headwall of Orotava suggests an age of formation for the valley is younger than 0.78 Ma and may even be younger than 0.27 Ma. The Icod valley is located immediately to the north of the most recent volcano on Tenerife, Las Canadas, and has been associated with the collapse of its caldera, between 1.2 and 0.2 Ma. A young age for the landslide is supported by the 3.5-kHz echo sounder data which show that the landslide is draped by a thin (< 10 m) layer of younger sediment. The landslide did not form, however, during a single catastrophic event but represents the amalgamation of a number of separate landslides. The occurrence of the ridge and valley topography in water depths of up to 2.5 km suggests that the shield-building basalts have subsided by at least this amount since they formed, 3.3–8.0 Ma. We speculate that this subsidence is caused by some form of stress relaxation that occurs in the underlying lithosphere. The giant landslide imaged in our sonar data is associated with the late stages in the development of the most recent volcano on Tenerife, Las Canadas, which only began at about 1.8 Ma. Thus landsliding may be a particular feature of the time soon after emplacement when because of incomplete isostatic adjustment, oceanic volcanoes have their greatest elevations above sea-level and therefore are most susceptible to slope failure.

Onset of submarine debris flow deposition far from original giant landslide

Submarine landslides can generate sediment-laden flows whose scale is impressive. Individual flow deposits have been mapped that extend for 1,500 km offshore from northwest Africa. These are the longest run-out sediment density flow deposits yet documented on Earth. This contribution analyses one of these deposits, which contains ten times the mass of sediment transported annually by all of the world’s rivers. Understanding how this type of submarine flow evolves is a significant problem, because they are extremely difficult to monitor directly. Previous work has shown how progressive disintegration of landslide blocks can generate debris flow, the deposit of which extends downslope from the original landslide. We provide evidence that submarine flows can produce giant debris flow deposits that start several hundred kilometres from the original landslide, encased within deposits of a more dilute flow type called turbidity current. Very little sediment was deposited across the intervening large expanse of sea floor, where the flow was locally very erosive. Sediment deposition was finally triggered by a remarkably small but abrupt decrease in sea-floor gradient from 0.05° to 0.01°. This debris flow was probably generated by flow transformation from the decelerating turbidity current. The alternative is that non-channelized debris flow left almost no trace of its passage across one hundred kilometres of flat (0.2° to 0.05°) sea floor. Our work shows that initially well-mixed and highly erosive submarine flows can produce extensive debris flow deposits beyond subtle slope breaks located far out in the deep ocean.

Characterization and recognition of deep-water channel-lobe transition zones

The channel-lobe transition zone (CLTZ) is an important, but commonly overlooked, element of many deep-water turbidite systems. Recognizing this zone is difficult in both modern and ancient environments and depends largely on the quality and resolution of the data obtained. In this article, three case studies of modern CLTZs are presented, largely based on high-resolution side-scan sonar imagery. These data are then compared to other well-defined CLTZs, both modern and ancient, and the common characteristics identified. CLTZs occur at canyon/channel mouths and are commonly associated with a break of slope. Most sediment bypasses this zone, and consequently only coarse sands and gravels are deposited, although these are commonly patchily distributed and extensively reworked. The CLTZ is characterized by abundant erosional features, including isolated spoon- and chevron-shaped scours up to 20 m deep, 2 km wide, and 2.5 km long. In areas of more widespread erosion, these merge to form amalgamated scours several kilometers across. Depositional bed forms include sediment waves with wavelengths of 1-2 km and wave heights up to 4 m. The presence or absence of a CLTZ has important implications for hydrocarbon exploration and development, especially in terms of the connectivity between sandy channel-fill and lobe facies.

The origin of deep-water, coral-topped mounds in the northern Rockall Trough, northeast Atlantic

Mounds associated with the cold water coral Lophelia pertusa are widespread in the North Atlantic, although the factors controlling their distribution are not well understood. Here we examine a group of small, coral-topped mounds (the Darwin mounds) which occur at 1000 m water depth in the northern Rockall Trough, northwest of the UK. Individual mounds are up to 75 m in diameter and 5 m high, although some ‘mound-like’ targets seen on sidescan sonar have little or even negative relief. Some mounds are associated with ‘tail-like’ features, imaged as elongate patches of moderate backscatter up to 500 m long, elongated parallel to prevailing bottom currents. High-resolution sidescan images and seabed photographs show hundreds of coral colonies, each a metre or so across, on each individual mound. Many other organisms, mainly suspension feeders, occur in association with the coral. Piston cores from the mounds contain predominantly quartz sand with only scattered coral fragments, showing that bioclastic material is not a major contributor to mound building. A field of seabed pockmarks occurs immediately south of the Darwin mounds. On sidescan sonar data, pockmarks are low relief, circular depressions, typically around 50 m in diameter. The seafloor around the pockmarks consists of uniform, heavily-burrowed, muddy sediments and no specific biological communities, nor any sedimentological or photographic evidence for active seepage, were observed. The distribution of mounds and pockmarks suggests a gradual transition from mounds in the north to pockmarks in the south. This, combined with the lack of bioclastic material in the mound sediments, suggests that both mounds and pockmarks are created by fluid escape from below the seafloor. Mounds occur where fluids carry subsurface sand to the surface, where it forms mounds because bottom currents are not strong enough to disperse it. Pockmarks form where muddy material is eroded by fluid escape but dispersed by bottom currents. Despite the origin of mounds through fluid escape, we suggest that it is the elevated mound topography, rather than any fluid escape, that is advantageous to the corals. This is supported: (1) by the wide variety of suspension-feeding organisms that occur on the mounds, since all of these are unlikely to have a specialised seepage-related lifestyle, and (2) because corals and their associated community do not occur around pockmarks, where seepage has also occurred but elevated topography is absent.

The northwest African slope apron: a modern analogue for deep-water systems with complex seafloor topography

The Northwest African slope apron is an interesting modern analogue for deep-water systems with complex seafloor topography. A sediment process map of the Northwest African continental margin illustrates the relative roles of different sedimentary processes acting across the entire margin. Fine-grained pelagic and hemipelagic sedimentation is dominant across a large area of the margin, and is considered to result from background sedimentary processes. Alongslope bottom currents smooth and mould the seafloor sediments, and produce bedforms such as erosional furrows, sediment waves and contourite drifts. Downslope gravity flows (debris avalanches, debris flows and turbidity currents) are infrequent but important events on the margin, and are the dominant processes shaping the morphology of the slope and rise. The overall distribution of sedimentary facies and morphological elements on the Northwest African margin is characteristic of a fine-grained clastic slope apron. However, the presence of numerous volcanic islands and seamounts along the margin leads to a more complex distribution of sedimentary facies than is accounted for by slope apron models. In particular, the distribution and thickness of turbidite sands are controlled by the location of the break-of-slope, which is itself controlled by the pre-existing submarine topography.

The morphology of the submarine flanks of volcanic ocean islands: A comparative study of the Canary and Hawaiian hotspot islands

The submarine flanks of volcanic islands are shaped by volcanic constructional processes, landslides, erosion, sediment deposition and tectonic movements. We use a newly acquired multibeam sonar dataset from the westerly Canary Islands (El Hierro, La Palma and Tenerife) to develop a comparison with the Hawaiian Islands, which suggests differences in the processes constructing and modifying their flanks. Landslides affect the flanks of both island groups. Debris avalanches (fast-moving shallow landslides) have left smooth chutes and blocky deposits in both cases, but blocks within some Hawaiian avalanche deposits are markedly larger. We attribute the larger block sizes in the Hawaiian Islands to the fact that their avalanches were relatively unconfined, whereas many Canary and Hawaiian avalanches with small block sizes appear to have been constrained down narrow chutes, forcing interactions between blocks within the flows and encouraging disintegration. Furthermore, the Hawaiian avalanches with the largest blocks initiated near sea-level, whereas many of the Canary avalanches initiated above sea-level, so hydraulic resistance of water entering cracks may be an additional factor in resisting block disintegration during flow. Slow-moving deep-seated slumps or volcanic spreading have produced submarine benches and tabular escarpments due to thrust faulting adjacent to several Hawaiian rift zones, but are not well-developed in the Canaries. Although volcanic morphology is partly obscured by sedimentation in the Canaries, we are able to interpret lava terraces around the deep flanks of El Hierro which are similar to those found in the Hawaiian Islands. However, cones rather than terraces are the most common volcanic forms in the Canary Islands, implying that flank eruptions have involved magma with significant volatile contents, assuming that volatile contents dictate whether cones or terraces are formed. These differences may ultimately originate from the different building rates of the two island groups. For example, the lack of evidence for high-level magma chambers in the Canaries, associated with their lower outputs, implies that there is less possibility for degassing of magma below the summit before lateral intrusion down rift zones, hence cones rather than lava terraces are more commonly observed. The apparent lack of slumping or volcano spreading could also reflect a lack of driving pressure from extensive high-level magma chambers in the Canaries.

Turbidity current sediment waves on the submarine slopes of the western Canary Islands

Two sediment wave fields have been identified on the flanks of the western Canary Islands of La Palma and El Hierro, using a high-quality 2-D and 3-D dataset that includes GEOSEA and TOBI imagery, 3.5-kHz profiles, and short sediment cores. The La Palma sediment wave field covers some 20,000 km2 of the continental slope and rise, and consists of sediment waves with wave heights of up to 70 m and wavelengths of up to 2.4 km. The wave crestlines have a complex morphology, with common bifurcation and a clear sinuosity. Waves have migrated upslope through time. Cores recovered from the wave field contain volcaniclastic turbidites interbedded with pelagic/hemipelagic layers. The wave field is interpreted as having formed beneath unconfined turbidity currents. A simple, previously published, two-layer model is applied to the waves, revealing that they formed beneath turbidity currents flowing at 10–100 cm/s−1, with a flow thickness of 60–400 m and a sediment concentration of 26–427 mg/l. The El Julan sediment wave field lies within a turbidity current channel on the southwest flank of El Hierro. The sediment waves display wave heights of about 6 m and wavelengths of up to 1.2 km. The waves are migrating upslope, and migration is most rapid in the centre of the channel where the flow velocity is highest. This wave field has been formed by channelised turbidity currents originating on the flanks of El Hierro.

Fluid seepage at the continental margin offshore Costa Rica and southern Nicaragua

A systematic search for methane-rich fluid seeps at the seafloor was conducted at the Pacific continental margin offshore southern Nicaragua and northern central Costa Rica, a convergent margin characterized by subduction erosion. More than 100 fluid seeps were discovered using a combination of multibeam bathymetry, side-scan sonar imagery, TV-sled observations, and sampling. This corresponds, on average, to a seep site every 4 km along the continental slope. In the northwestern part of the study area, subduction of oceanic crust formed at the East Pacific Rise is characterized by pervasive bending-induced faulting of the oceanic plate and a relatively uniform morphology of the overriding continental margin. Seepage at this part of the margin typically occurs at approximately cone-shaped mounds 50 - 100 m high and up to 1 km wide at the base. Over 60 such mounds were identified on the 240 km long margin segment. Some normal faults also host localized seepage. In contrast, in the southeast, the 220 km long margin segment overriding the oceanic crust formed at the Cocos-Nazca Spreading Centre has a comparatively more irregular morphology caused mainly by the subduction of ridges and seamounts sitting on the oceanic plate. Over 40 seeps were located on this part of the margin. This margin segment with irregular morphology exhibits diverse seep structures. Seeps are related to landslide scars, seamount-subduction related fractures, mounds, and faults. Several backscatter anomalies in side-scan images are without apparent relief and are probably related to carbonate precipitation. Detected fluid seeps are not evenly distributed across the margin but occur in a roughly margin parallel band centered 28 ± 7 km landward of the trench. This distribution suggests that seeps are possibly fed to fluids rising from the plate boundary along deep-penetrating faults through the upper plate.

Landslides and the evolution of El Hierro in the Canary Islands

Seismic and sonar data have been used to evaluate the extent and characteristics of giant landslides on the flanks of El Hierro in the Canary Islands. As the youngest and most southwesterly of the Canary Islands, El Hierro has experienced rapid growth and destructive events in its 1.12 million year history. At least four giant landslides (El Golfo, El Julan, San Andres, and Las Playas) have modified ~450 km3 of El Hierro during the last 200–300 thousand years, with each landslide event removing around 3% of the total edifice volume. The extent of landsliding indicates that it is the main process of decay. We characterise flank morphology around El Hierro and distinguish between rugged, unfailed flank, failed flank and steep gullied ridge. Flanks affected by landsliding have downslope long profiles with distinctive b coefficients and exponential forms. The El Golfo landslide is the most recent (15 ka), best described and clearly defined landslide in the Canary Islands. The El Julan landslide (SW flank) has an estimated volume of 130 km3, an age of >200 ka and is characterised by gravitational slumping. On the SE flank, two new landslide events are reported. The younger landslide (Las Playas) occurred 145–176 ka, has a narrow, steep-sided embayment and a corresponding blocky debris avalanche deposit. The older landslide (San Andres) is recognised on the basis of a highly chaotic seismic facies offshore and reduced upper flank gradients. Its lack of an upper flank embayment and offshore blocky debris avalanche lead us to interpret that the landslide involved gravitational slumping, possibly a series of events, which reduced upper flank gradients, but did not catastrophically collapse to produce a debris avalanche.