Australia: The Land Where Time Began

A biography of the Australian continent 

Landslides, Tsunamis and Climate Change

During the transition from glacial to interglacial conditions episodes of intense landslide activity have been identified in the latest Pleistocene and Holocene in places such as the UK, the Pyrenees, European Alps, Carpathian Mountains, the peaks of the Canadian Rockies, the Andes and The Southern Alps of New Zealand.  Of all the factors that may have had a role water was involved with all of them.  The author¹ suggests that the increased development of instability and slope failure that occurred in post-glacial times may have been contributed to by such factors as the glaciers and ice caps retreating leaving expanses of bare rock and debris, rising temperatures leading to the thawing of mountain permafrost, and the increase of precipitation rates as weather patterns changed. In some places other factors such as rapid uplift and increased earthquake activity resulting from the unloading of ice may also have contributed by deforming slopes that were already destabilised and shaking them.  There are a number of options for slopes to be destabilised and landslides triggered in a world where so much water is around as the climate transitioned from cold to warm. In high mountain ranges where ice physically binds together a rock face or slope rising temperatures melting the ice can destabilise the slope, and the possible expansion of the warming rock may also have contributed slightly. Persistent heavy precipitation can saturate and mobilise loose debris that has been exposed by retreating ice. Rock masses can be destabilised, especially where it is strongly fractured, by rainwater or the melting of snow or ice. In an unstable rock mass a failure surface needs to form before it can be the site of a landslide. This can happen over periods of days or weeks, or in the case of the greatest volcanic landslides, it can take decades, centuries or even longer. Some idea of the operation of the process can be gained from analysis of a notorious landslide that occurred in Italy in the 1960s.

A small watercourse in the mountainous north of Italy that flowed into the larger Piave River through a spectacular narrow defile about 300 m high was chosen for the site of a dam. In 1956 work began on the dam that would be the tallest in the world at that time. By 1959 the dam was 262 m high. New cracks were found in the northern face of Mt Toc that formed the southern margin of the Vajont valley as a road was being constructed along the planned reservoir. Serious concerns about the stability of Mt Toc were expressed by a number of scientists, but construction continued and the dam was completed in February 1960. Later that year almost 1 million cubic metres of  rock collapsed into the reservoir, as well as a number of smaller landslides,  but they continued with the finishing of the dam. When the water reached 170 m deep the north face of the mountain began slowly moving towards the reservoir, an M-shaped crack in the face of the mountain opened that was  2 km long and half a kilometre above the floor of the valley, marking the detachment of an enormous mass of rock from the mountain face, though the work continued, in spite of the crack widening by a millimetres per day for the next 3 years. It was noticed that the opening of the crack slowed every the water level was lowered, which it was periodically. It was assumed that the widening of the crack could be controlled by controlling the water level of the reservoir. By the time the water depth reached 245 m the opening rate of the crack increased to 3 cm per day. This caused the decision to be made to drain the dam to slow the widening. At 10:30 PM on 9 October 1963 250 million tonnes of rock was carried by a landslide into the reservoir, completely filling it. The 115 million m³ of water in the reservoir was pushed over the dam wall with a splash wave that was 150 m high that struck the town of Longarone that was close to the foot of the dam in seconds almost completely destroying it and all its residents before carrying on along the valley, the wall of water and the cushion of air in front of it rushing down the valley destroying the villages of Pirago, Villanova, Rivalta and Fae. In all it killed 2,500 people.

The engineers had assumed that the pore water pressure in the rock of Mt Toc increased as the water level in the reservoir rose and decreased as it was lowered,. They believed they could control the cracking by manipulating the level of water in the reservoir. 40 years later Chris Kilburn, a volcanologist and landslide expert of University College London and Dave Petley of University of Durham determined what actually happened. They found that as the water level rose water had been percolating down to a clay layer 200 m below the surface. Water is able to corrode the rock at the pressures encountered at these depths. The tips of existing cracks had been weakened by the water, which encouraged growth of these cracks, a process called slow cracking. Over time larger discontinuities will be formed by the coalescence of these cracks. Deformation of the destabilised rock mass will occur slowly as long as these cracks are largely isolated from one another. The fractures begin to join up as the coalescence continues, movement accelerating until there is only a single discontinuity, a failure plane is formed that completely detaches the deforming mass of rock above from the stable rock below. From this point the process cannot be stopped and the sliding that follows is catastrophic.

According to the author¹ the Vajont Dam, water entered the clay, which is hard and brittle at these depths, as the water level in the dam rose. New cracks opened as the clay slowly expanded, these cracks joining up progressively. Cracking stopped or slowed when the water levels were lowered but didn’t close, the growth and coalescence of the cracks continued. Lowering the water level did nothing to halt the process of development of instability, slowing the cracking rate but not stopping it. Slope failure was certain once the water level reached 245 m as by then the fractures within the clay had joined up to a sufficient extent.

This same slow cracking process is likely, according to the author¹, to have a role in triggering large landslides in episodes of rapid climate change in the past, especial in the post glacial period when meltwater from retreating ice sheets, increased precipitation, and surface runoff at higher levels were contributing to the formation of many new water bodies around the margins of the melting ice sheets and glaciers in mountainous areas. The author¹ suggests perfect conditions for the destabilising of adjacent slopes and rock faces resulted from water levels that were rising and around the margins of the ocean basins that had been depleted during the glacial phase may have promoted increased instability levels. He also suggests that the draining of some of the extremely large pro-glacial lakes such as Lake Agassiz may have resulted, at least in part, from the overtopping and erosion of the barriers retaining the meltwater by the triggering of waves by landslides.

In the oceans

According to the author¹ submarine slope collapses are among the largest of the landslides that have been identified up to the present, their possible triggers being one or more a number of possibilities. There is hard evidence that submarine landslides do not occur at a constant rate, being triggered at times of abrupt climate change. The author¹ suggests this should not be surprising as at times when the climate swings from glacial to interglacial and back again there are very large changes in the global sea level, especially at high latitudes, together with a variety of disturbances in the crust along the margins of the oceans as the ice sheets expand and contract.

Off the coast of Norway, in the North Atlantic, something happened in a huge mass of glacial sediment in the Early Holocene a bit over 8,000 years ago. A chunk of sediment with a volume of 3,500 km³ separated from the submarine continental shelf and slid downslope to spread out on the floor of the deep sea, extending more than half way to Iceland. The scar remaining from the collapse was nearly 300 km wide. This Storegga Slide triggered a tsunami that has been estimated to have been 25 m high, the evidence of which as been found in the form of distinctive sand layers in peat on the Shetland Islands, the northernmost part of the UK. The sand was recognised in 1990 as the evidence of the Storegga Slide by David Smith, a Quaternary scientist in the UK, and others. The author¹ says this is similar to the maximum run-up of the Indian Ocean tsunami that occurred in 2004 on Sumatra near the epicentre of the earthquake. Since then evidence of the Storegga Slide has been found at several places along the coast of Norway, where the tsunami run-up appears to have been 10-12 m, and along the east coast of Scotland and northeast England were there were run-ups of 3-6 m. The author¹ suggest that no other evidence has been found because either they are yet to be recognised or that conditions were unsuitable for the preservation of such evidence.  He also says there is no doubt that the tsunami would not have been limited to the vicinity of Norway and the northernmost part of the UK, and it had probably inundated most coastal areas in the region of the North Atlantic.

Surveys carried out since the Shetland findings have identified more sandy deposits that suggest the islands were struck by 2 other, though smaller, tsunamis, one of which occurred about 5,000 years ago and the 2^nd during Roman or early Medieval times, both almost certainly being the result of submarine slides somewhere in the North Atlantic.

The author¹suggsts that as there are 3 tsunamis recorded in the Shetlands it indicates that there are many more tsunamis such as that triggered by the Storegga Slide. There are many deposits on the ocean floor around the world that originated in island and coastal volcanoes, though most originated as submarine slides around the margins of the continents. Regions showing evidence of particular instability can often be recognised where there have been a number of landslides, a case in point being the continental shelf off Norway that has evidence of 3 landslides besides the Storegga Slide. According to the author¹ a number of features are shared by the places in the submarine environment displaying evidence of repeated landslides. There are typically thick sediment deposits, a ready source of landslide material, the instability is increased by the sloping seafloor, and once a slide commences, the slope being steep enough gives added impetus. Some features usually found in submarine environments prone to slope collapse are unusually large sediment deposits as are found near the deltas of rivers, or by earthquake activity, a characteristic of deep ocean trenches at the subduction zones around the world. The large changes in global sea level, rises and falls, and ancient landmasses that are subject to uplift and subsidence as the ice sheets expanded and retreated, would have added dynamism to the marine endowments that were prone to landslides.

Building up to a slide

According to the author¹ there is not much doubt that the occurrence of the Storegga Slide resulted directly from the changing climate; the change from a glacial phase to an interglacial phase that occurred rapidly at the start of the Holocene, producing perfect conditions for landslides. During the last glacial phase massive ice sheets covered the Scandinavian countries, as well as much of Northern Europe, in places reaching depths of 2.5 km. According to the author¹ at this time the 2 things which prepared the conditions for the Storegga Slide when the glacial phase finally ended were the subsidence of the lithosphere that was covered by ice by several hundred metres and the exposure of the continental shelf in the area as a result of lowered sea levels. The fast-moving glaciers that carried large amounts of sediment and debris across the continental shelf to the sea where it was dumped on the continental slope that continued on to the deep ocean floor. The main body of the Scandinavian Ice Sheet had gone by 8,000 years ago and sea levels had risen to almost modern levels, resulting in the continental shelf adjacent to Norway was again submerged. Following the unloading as the ice melted the lithosphere rose, releasing the pressure that had kept the underlying faults inactive during the glacial phase, resulting in a marked increase of seismic activity. In the underlying clays the pressure of the pore-water increased, weakening them and instability was promoted by the addition to the top of the clay-rich continental shelf of large amounts of glacial sediment. Seismic activity opened a failure surface within the clays and started the slide.

The most widely supported mechanism for the Storegga Slide is increased earthquake activity resulting from uplift of Scandinavia that occurred after the close of the glacial phase, and the author¹ suggests this was not the only time this occurred. The site of the most recent landslide has been found to be the source of a number of other slides, the timing of which matches that of the glacial-interglacial cycle in a similar way to the Storegga Slide in the Holocene.

At least 1 link between climate change and the occurrence of submarine landslides appear to be now well established. At times of low sea level during glacial phases glaciers carry large volumes of sediment to the ice sheet margins. When ice sheets melt and contract in interglacial periods seismic activity associated with uplift trigger landslides in the thick deposits of sediment.

Evidence of other large landslides has been found off the coast of Ireland and near Rockall. 24 submarine landslides have been found adjacent to the area formerly covered by the American Ice Sheet. Most of these slides occurred during the last ice age and the Holocene. More than 50 landslides have been recognised further south off the Atlantic margin of the US, and most are believed to have formed in a similar manner at some time in the glacial period or the Holocene. As with other submarine landslides the trigger is believed to have been the seismic activity associated with the post-glacial rebound following the melting of the North American Ice Sheet. The author¹ suggests that the Storegga model can’t be used to explain all submarine landslides. In the middle and low latitudes that were not covered by ice sheets in the last ice age the situation is more complex where destabilisation can occur by a range of different factors, though earthquakes are believed to be the trigger, but not as a result of lithospheric rebound, rather to local tectonic circumstances. Halite (a salty mineral) domes that rise into the sediments, and eventually punch through the sediments on the continental margin resulting from its low density, have been suggested to have a role in destabilising slopes and encouraging sliding. Repeated loading caused by storm waves, or even tides, are other possible triggers that have been suggested.

Many of the slides that are further from the poles also seem to have occurred at the transition from the glacial to the post-glacial conditions, as occurred at the end of the last ice age. The athor¹ suggests that in this context a major role may have been played by rising sea levels by submerging and destabilising continental shelves that had previously been exposed, and in the process maximising the opportunities for slope failure. Rapid, temporary sea level rises that occurred within glacial times, that were associated with interstadials, short warm episodes, appear to have promoted instability and failure At the margins of the ocean. A similar effect is apparent in post-glacial times. A number of major slides occurred off the coast of Mauritania in northwest Africa that appear to have been caused by destabilising sand dunes that had formed on the continental shelf when the sea level was lower. The author¹ suggests it is clear that most submarine landslides occur at times of rising sea level, or at least stable.

Submarine landslides - gas hydrates

Gas hydrates are solid mixtures of gas and water, usually methane, that are ice-like and that are present in marine sediments and beneath the permafrost in the Arctic where there are very large quantities. There are extremely large quantities of carbon sequestered in these gas hydrates on the sea floor, in total 2.5 times that in the atmosphere. They have been implicated in past episodes of sudden onset of exceptional warm spikes in the climate, such as in the Palaeocene-Eocene Thermal Maximum (PETM) that began a bit less than 56 Ma. Gas hydrates remain in solid form indefinitely in the thick sediment deposits on the ocean floor along the margins of ocean basins as long as they are subject to very low temperature and high pressure, the hydrate breaking down if the temperature rises only slightly or the pressure is decreased. If they break down they release huge quantities of methane into the atmosphere.

According to the author¹ some aspects of gas hydrates are not well understood. The constituents of gas hydrates, mostly methane and water, but water with other compounds such as CO₂, hydrogen sulphide, and other compounds. They are not bound tightly enough to form a chemical compound, being loosely attached to one another that allows the transition from a solid to a gas and vice versa easy, the transition being accomplished simply by changing the ambient pressure and temperature. A sufficient supply of water and gas is required to be present together, with pressures that are high enough and temperatures are low enough, or in some cases where the 2 meet to provide the perfect, ‘Goldilocks’, environment. The temperatures of the ocean were significantly lower than at present during the last ice age and that would presumably have allowed gas hydrates to form at shallower depths than they can at the present. Such gas hydrate deposits in relatively shallow water would have been particularly susceptible to changing conditions of temperature and pressure in oceans that were warming rapidly characteristic of post-glacial times.

A number of processes operating in the ocean sediments form methane, the key component in most gas hydrates, mostly by the decomposition of organic debris such as the remains of dead plankton, and by conversion of CO₂ in the sediment of the ocean floor. A sufficiently large change in either the ambient temperature or pressure on the ocean floor can promote the sudden transition from a solid to a gas, as the stability of gas hydrates is controlled by both the temperature and pressure. The risk of the destabilisation of gas hydrates increases as the oceans warm in a warming climate, in which case they would suddenly dump huge quantities of methane into the atmosphere. Countering this, as the temperature of the Earth increase more ice melts at high latitudes and altitudes, this increases the pressure on the gas hydrates tending to stabilise them. Because it is not likely that the 2 effects will exactly balance each other it will depend on which factor will outweigh the other. If the water temperatures in the deep ocean rise faster than the sea level rises this will favour the decomposition of gas hydrates with the release of methane into the atmosphere that would increase the atmospheric temperature even faster. If, on the other hand, the sea level rises faster than the atmospheric temperatures then the gas hydrates would remain stable stabilised, at least for a time.

Exactly how the gas hydrate is released is relevant both in the past and in the future is an area of interest. According to the author¹ the question is how it reaches the atmosphere, is it by bubbling through that water to the atmosphere or by some other method? A suggestion that has attracted some attention is that proposes a role for submarine landslides. According to this proposal a submarine landslide that was sufficiently large would be capable of triggering the breakdown of the gas hydrates beneath them by removing the overlying sediments that would virtually instantly reduce the pressure acting on them, the released gas contributing to a self-feeding runaway process in which the methane acts to increase the area of the slope failure, leading to the release of more methane from more gas hydrate as the pressure acting on the new hydrates is lowered. Mark Maslin et al., a Quaternary scientist, University College London, is one of the scientists that has driven the idea of a link between marine gas hydrate breakdown in the past and marine landslides. They proposed a number of short-lived carbon isotope excursions in the marine sediments that have occurred at various times in the history of the Earth which they propose resulted from large volumes of C¹², which is lighter and considered to be a significant carbon that is sourced from methane in gas hydrates. One such event is the excursion that occurred coincident with the Chicxulub impact that defines the KT boundary 65 Ma, that is also coincident with a mass extinction that finally killed off most of the dinosaurs. According to Maslin and Simon Day the cataclysmic event was so violent that it triggered submarine landslides on a global scale. The removal of large volumes of sediment in these landslides around the world is suggested by Maslin and Day to have caused sudden drops in pressure acting on gas hydrates within the sediments and causing them to transform from solid to gas almost instantaneously with the immediate release of 300-1300 billion tonnes of methane to the atmosphere. According to the author¹ this proposal is attractive but speculative, though it could account for the large rise in atmospheric carbon dioxide that occurred following the impact. The author¹ suggests the evidence is strong for a contribution from marine gas hydrates towards the warming event, that was sudden and severe, that characterised the PETM.

The Maslin supporters point to evidence of slope failures on a large scale in connection with gas hydrate destabilisation at this time along the Atlantic Basin’s western margin during the PETM. It has also been speculated that in Quaternary the switchback climate of that time may have involved the release of marine gas hydrates, the resul being the climate switched back and forth between extremely cold to balmy and back again over the last 2 million years. According to the author¹ submarine landslides have been suggested as the key, the suggestion being that more occur at times of rapid climate change, which leads to the destabilisation of more gas hydrate and a dramatic rise in the atmospheric methane, which would accelerate warming even further. As for the PETM and earlier periods the proposed link will probably remain no more than speculative for some time to come. This results, at least in part, from uncertainties in the atmospheric methane and partly because of the difficulty with the dating of Submarine landslides. It has also been suggested, though again with a lack of hard evidence, that at times of rapid climate change destabilisation of marine gas hydrate, due either to falling sea levels or ocean temperatures that are rising, may actually trigger the marine landslides.

The author¹ suggests it appears reasonable that marine landslides and gas hydrate in marine sediments breakdown are linked in some way, and plenty of circumstantial evidence links them, though it could be some time before a robust cause and effect connection becomes apparent. According to the author¹ it is clear that a connection exists between the incidence of both marine landslides and methane gas release and periods in the past when the climate was being disrupted as it was transitioning to a new relatively stable state. The author¹ suggests we would be wise to take note given the projections for our future climate.

According to the author¹ at the time of the megaflip when the Late Pleistocene cold climate transitioned to the mild climate of the Holocene with melting ice, increased rainfall, filling ocean basins and glacial lakes that were expanding acted together to greatly increase the erosion rate, mobilise slopes of loose debris, detach rock faces and raise the pressure of pore water and promote failure surface growth. There are already signs that avalanche and landslide activity in mountainous areas may be increasing, probably as a result of climate warming.

Valley of the Oxen (Velle del Bove)

This is an amphitheatre in the side of Mt Etna that has been dated to about 7,500 years ago, in the Early Holocene, the period of time when the crust in many places on the Earth was rebounding after the loss of their ice cover and there was a burst of volcanic and seismic activity. It was at this time when the cold climate of the ice age had given way to wet and warm conditions that rainfall that was heavy and persistent eroded large valleys on the sides of Mt Etna, as well as raising water tables in the region, and saturating the sides of the volcano. Torrential rain caused the ashy sides of the volcano to be dislodged as huge mudflows as an explosive combination of water and magma blasted out rock columns, debris and dust. As magma was forcing its way to the surface from beneath the summit the pore water of the sodden rocks was heated to boiling point forcing it to expand. As a result of the boiling groundwater the pressure broke the volcano in two. The immediate result was this was the detachment of a huge chunk of the eastern flank and rushed down the side of the volcano. When this mass of volcanic debris landed in the Ionian Sea it caused a tsunami that raced to the east. Based on a computer model by Maria Pareschi et al. the researchers estimated that all of the eastern Mediterranean would have been affected by the tsunami, Greece, the Levant and North Africa, and according to the author¹, if the model proves to be accurate, the run up at Calabria in southern Italy may have been as high as 40 m, in Greece and Libya about 8-13 m, and on the coasts of Egypt and the Levant, about 2-4 m. It has been suggested by the Italian researchers that it may have been this tsunami that destroyed Atlit-Yarn, Neolithic village in coastal Israel, the village having been suggested by archaeological excavations to have been destroyed by a tsunami at about this time.

Natural rubble piles

Most volcanoes are not solid, but formed a giant mound of volcanic and lava rubble that can be easily destabilised and the hearts of many volcanoes have also been weakened by hydrothermal weathering caused by the hot fluids that are constantly being circulated, the result being interiors that are a clayey mush that destabilises much easier than solid rock. There are also many other reasons for volcanoes to be destabilised rapidly and making them prone to lateral collapse. When the substratum beneath volcanoes is uplifted, subsides or is tilted it may promote instability over a long time period. Environmental factors such as rising or falling sea levels, especially when the climate is changing abruptly. The location of most volcanoes on or close to the margins of tectonic plates, sites that are subject to many earthquakes, there are many episodes of seismic activity that can trigger collapse of most, if not all, of a volcano’s flank. Volcanoes often are either snow-capped or have a wet local climate, and water in either solid or liquid form often have a role in the triggering of earthquakes, and earthquakes are good at destabilising volcano flanks, making them more prone to collapse. The promotion of landslides can also be achieved by heavy rain, and the melting of snow or ice, especially through their effect of pressurising pore water in the structure of the volcano. According to the author¹ this is a key reason for the instability and collapse of a volcano apparently being linked to climate change.

According to the author¹ in general the largest volcanoes which have been active for thousands or even millions of years are subject to the largest collapses. The largest volcanoes, such as the Canary Island and the Hawaiian Archipelago volcanoes, both of which are ocean island volcanoes, periodically have collapses of large chunks of their margins into the sea, leaving large amphitheatres of solid rock. El Hierro, one of the Canary Islands, is a good example of this; huge collapses that occurred in prehistoric times have sculpted the island into the shape of a tricorne hat. The have been at least 12 large collapses on the 7 islands of the archipelago that are known either by the scars they left on the land or by the trains of disrupted debris they left on the ocean floor. A massive amount of rock collapsed into the sea from the island of Tenerife about 500,000 years ago, the catastrophic event leaving 2 spectacular valleys, the Orotava Valley and the Icod Valley, and on Gran Canaria, a neighbouring island, an 80 m high remnant of a giant collapse of much of the southern flank of a volcano about 3 Ma, the iconic monolith called Roque Nublo. The Taburiente Caldera on the nearby island of La Palma is a vast hole that was gouged out by a landslide more than 500,000 years ago dominates the islands northern part, and is bounded by vertical cliffs more than 1 km and 2 days are required to walk I to it. The Cumbre Vieja volcano to the south of the Taburiente Caldera is the most active in the archipelago in recent times is of great interest to geologists and geophysicists. The main cause for this interest is that a large chunk of the western flank sunk towards the sea by a few metres, a worrying indication that is becoming unstable. There has been a lively debate among scientists about if and when it drops into the sea and if it does how big with the resulting tsunami be.

In the Canary Islands the largest of the volcano collapses have volumes of a few hundred km³ while that St Helens volcano that appeared to so large had a volume of about 2.5 km³. The collapses of Hawaiian volcanoes put even the collapses in the Canary Islands in the shade, the largest of the 70 known collapses having volumes of more than 1,000 km^3. The Nuuanu landslide, the largest of the Hawaiian landslides, spread across the submarine flank of the Koolau volcano that is believed to probably be extinct, and forms the eastern half of Oahu, the island on which Honolulu is situated. When it collapsed a few million years ago a huge mass of its flank dropped into the sea spreading across 23,000 km² of the adjacent sea bed, an area about as large of the state of Vermont. This landslide had a volume of 5,000 km³ or more. One of the individual blocks that was part of this landslide is the Tuscaloosa Seamount that is 30 by 17 km and nearly 2 km above the floor of the Pacific. As with other giant landslides the Nuuanu landslide travelled at high speed and came to a stop at about 230 km from the volcano. It has been estimated that the debris flow reached a speed of about 270 km per hour or more, and the huge amount of energy generated by this collapse was so great that after crossing the 5 km deep seafloor around the Hawaiian Islands, that resulted from the downward pressure of the archipelago, the debris flow ran about 150 km up the far side of the deep trough. The tsunami this caused must have been truly enormous.

Mount St Helens and other volcanoes on land

Volcanoes on land tend to be much smaller than their marine counterparts, and they are easier to study as well as study their eruptions as they happen. The sizes of the landslides that occur when volcanoes on land breakup are about 1,000 times smaller than the largest of the Hawaiian landslides, the largest being up to a few 10s of km³. The author1 suggests the largest landslide may have been the one that occurred at Mt Shasta near Mount St Helens about 330 Ma that had a volume of about 45 km³, the debris moving about 50 km from the volcano. About 18,000 years ago there was another large landslide that moved about 25³ km of debris from the flank of Nevado de Colimo volcano in Mexico west to the Pacific Ocean 120 km away. Other large landslides from volcanoes have been identified at Fuego and Pacaya in Guatemala, Taranaki (aka Egmont) in New Zealand and Mt Etna in Italy.

In 1956 one of a group of volcanoes on the Kamchatka Peninsula in Russia, Bezymianny, lost its eastern flank in a massive landslide. A blast from the side of the volcano and an eruption that continued for 4 hours was triggered by the sudden reduction of pressure of the magma in the volcano.

Mount St Helens, 1 of 20 volcanic peaks in the Cascade Range, followed the same sequence of events 24 years later in 1980. It had been inactive for more than 120 years and the vent at the top of the volcano had been blocked. Then in March 1980 swarms of small earthquakes in its vicinity indicated things were changing as new magma was building up beneath the volcano. With the usual vent blocked the magma pushed on the northern flank which bulged out more than 150 m. On 18 May a magnitude 5.1 earthquake directly under the bulging flank triggered a collapse that dropped the mass of debris at up to 250 km/hour. The explosive release of gas from the magma sent a shockwave, that the author¹ suggests may have been travelling as fast as speed of sound, that rapidly overtook the landslide and 600 km² of mature forest was obliterated, the destruction reaching almost 30 km from the volcano. The 57 people killed by the eruption were all caught by the lateral blast, including David Johnston the geologist monitoring the north side of the volcano. Once the north flank and the summit of Mount St Helens was gone it erupted sending dust and gas to 25 km into the atmosphere in the first 15 minutes and spreading ash across 11 states. It was the largest eruption in the US excluding Alaska since 1915 when Lassen Peak in California erupted. The damage bill ultimately reached $1 billion.

The author¹suggests the Mount St Helens collapse is the best studied and was the beginning of acute scientific interest in volcano stability, as well as the dangers resulting from the failure of volcanic edifices and volcanic landslides. As a result of research carried out partly as a result of the collapse of Mount St Helens it has been found that such events occur much more often than was previously realised. There have been 20,000 deaths caused by 8 volcanic landslides over the last 350 years, though the average number of collapses over the last 500 years may be up to 20. 480 collapses at 316 volcanoes have been found by a recent survey,  based on the knowledge of the frequency of such events from historical records, and the author¹ suggests this is much less than the actual number. According to the author¹ 100 million people now live near volcanoes that have collapsed previously, and as such collapses appear to occur preferentially at times of the transitions of climate, this suggests there could be increased danger faced by those people.

Landslides

According to the author¹ it is now clear there is a climate dimension associated with geological hazards such as earthquakes and volcanic eruptions. Another geological phenomenon that displays such an association with times when the climate is in transition, as are temperature, rainfall and other factors, is landslide, whether associated with volcanoes or not. According to the author1 it is difficult to prove irrefutably that the instability and collapse of volcanoes are influenced or modulated by climate change, in part as a result of the incomplete record, and partly as a result of dating errors of some collapses, both of which make correlation much more difficult.

Largely as a consequence of the construction of a database of all known collapses of volcanoes at University College London, there is the beginning of a picture emerging. In the geological record there are 500 collapses that have been recognised, most of which occurred in the Late Pleistocene of Holocene, having ages of less than 40,000 years. As the present is approached there are increasing numbers of collapses recognised, as might be expected, as younger collapses will have had less time to erode or be covered by the products of later volcanic eruptions. Nearly half of all known volcanic collapses occurred during the last 10,000 years during the Holocene, as a consequence of more recent collapses being less affected by erosion and later volcanic activity. Researchers have tried to reduce this age bias in the record by a number of methods, one of which is to ignore all recent collapses, such as those that have occurred over the last 2,000 years and older than 40,000 years. This just about marks the limits of the carbon-14 dating method. Ignoring all collapses having volumes of less than 1 km³ is used to reduce bias associated with the improved chances of survival of smaller collapses of younger age. The author suggests that this method increases the chances of being able to work out collapse rate variations that may give a clue to a link with climate change, though the record provided by the remaining collapses is still not close to providing a complete record of the number of collapses in that period.

A time of climate extremes encompassing the Last Glacial Maximum is included in the period 40,000-2,000 years ago when the ice sheets had reached their maximum extent, and the Holocene in which the climate was transformed into the wetter, warmer state of the present. The author¹ suggests that any correlation between volcanic collapses and changing climate should become apparent over this time period, even with an incomplete record. Research by Rachel Lowe, UCL, has found a distribution of volcano collapse that is far from consistent, a number of distinct troughs and peaks becoming apparent since the Last Glacial Maximum, the troughs correlating with cold snaps and the peaks correlating with warmer phases. A trough in the number of volcano collapses occurs at the same time as the start of the Younger Dryas, about 13,000 years ago, and another at the start of the 8.2 ka cooling event. The succeeding warm phase is marked by a peak in the volcano collapse record.

6,000-2,000 years ago there were a number of switches from cold and dry to warm and wet and back again that apparently correlate with the collapse record. The author¹ suggests a changing sea level could possibly be a possibility, rising sea levels being driven by a warming climate promoting the collapse of the flanks of volcanoes that face the sea and island volcanoes as the crust beneath them is bent down. In the latest Pleistocene and Holocene there is not much evidence that significant global sea level changes were associated with the rapid switches from cold to warm that appear to influence collapse behaviour of volcanoes. The author¹ suggests a far more likely explanation may be availability of water changes and loss of ice mass at transitions from cold to warm climates. Loss of ice from a glaciated volcano can be a very effective way of removing buttressing from the volcano flanks leading the flank instability and the promotion of flank collapse. When combined with a wetter climate increased runoff from melting snow and ice can raise the pore-water pressure within the edifice which can lead to a situation favouring sideways collapse, as in the case of Valle del Bove on the flank of Mt Etna in Italy.

Simon Day, a British geologist, who has studied the possible future collapse of Cumbre Vieja volcano on the island of La Palma, suggests volcanoes are more likely to collapse when the climate is warm and wet than when it is cold and dry. He reports that over the past couple of hundred thousand years the greatest collapses of ocean island volcanoes can be correlated with sea surface temperature record, in the Hawaiian and Canary Islands, but also in other island archipelagos such as the Cape Verde Islands off the African west coast and Réunion Island in the Indian Ocean. Over this period major collapses appear to occur at times of high sea surface temperatures, though not low temperatures, as far as can be determined, being mindful of dating errors. It is suggested that sea surface temperature reflects other environmental conditions that can have the effect of promoting instability of volcanoes and edifice failure, though there is no direct effect on volcanoes by the surface temperatures of the ocean. The temperature of the ocean surface is a an indicator of the general global climate, the oceans being cool during glacial phases and considerably warmer during interglacials, and when they are warm the sea level is high and the global conditions tend to be generally clement and humid compared to the situation during glacials. Day has noted that at low latitudes, where the largest of the oceanic islands are situated, changes and peculiarities of Trade Winds, the prevailing easterly winds in the tropics, resulted from the warmer seas. The wind systems consequently dumped more rainfall on the upper flanks of any oceanic island volcanoes in their path than in cooler times that tended to be more arid prevailing during glacials. Day suggests this would have raised water tables by several hundred metres thereby increasing the opportunity for pore waters in the cores of volcanoes to be heated and pressurised by rising magma that would prime the flank for collapse.

These suggestions are supported by Gary McMurtry et al., School of Ocean and Earth Sciences and Technology (SOEST), University of Hawaii. In the 1990s their research was centred on the Alika-1 and Alika-2 landslides that resulted from lateral collapse of the largest active volcano in the world, Mount Loa, on the island of Hawaii, both of which occurred in interglacial periods at times of warm, wet climates with high sea levels, just as at present. According to the author¹ there seems little doubt that out present climate is conducive to promoting the lateral collapse of volcanoes, and will probably become even more so over time, though according to the author¹ this doesn’t mean that volcanoes are completely stable and unmoving under different climate conditions. This is exemplified by some of the Canary Island collapses that are believed to have occurred during times when the climate was colder and drier and sea levels were rapidly falling, a correlation that the author¹ finds not entirely surprising as the sudden reduction of the buttressing effect as huge volumes of water were removed from the flanks of the volcano. The author¹ suggests that in broad terms it is looking increasingly as if the collapse of volcanoes occurs preferentially at time of warmth and humidity.

It has been cautioned by McMurtry et al. that there are many regions where there are volcanoes that have collapsed in the past, and if increased precipitation is the primary driver for volcano collapse then as the world continues to warm, with the accompanying increased humidity, these volcanoes may become increasingly likely to collapse again, as it is expected that there will increasingly be a higher frequency of extreme rainfall events. As there are many coastal and island volcanoes that could be subject to increased rainfall there is also the possibility that collapse of these island volcanoes and the flanks of coastal volcanoes closest to the sea could generate tsunamis, which could then occur more frequently than they have been in the recent past. Tsunamis are often associated with undersea earthquakes such as when 1,200 km of the Sunda megathrust ripped open in 2004 and the large megathrust that occurred off the northeast coast of Japan in 2011. Huge volumes of water can also be displaced by volcanoes with comparatively catastrophic consequences, such as the cataclysmic eruption of Krakatoa in 1883 that had appeared innocuous until it erupted between Sumatra and Java in Indonesia that generated a large tsunami that killed 30,000 people as it destroyed many coastal villages. The majority of the 12 volcanic tsunamis that have occurred over the last 400 years have been triggered by volcanic landslides. Included among these is a substantial landslide on the Ritter Island volcano 100 km north of Papua New Guinea in the Bismarck Sea. The largest volcanic island collapse in historical times, that was twice as big as the Mount St Helens slide, generated a 20 m-high tsunami that killed 3,000 people on the adjacent islands.

The author¹ suggests the tsunamis generated by volcanic landslides should not be underrated, nor should their potential for generating tsunamis, as they are extremely violent events with the amounts of energy involved being staggering in the formation and transport of even the smallest volcanic landslides. The comparatively tiny Mount St Helens landslide generated almost as much energy as 2000 Hiroshima-sized bombs, would have been dwarfed by the energy generated by the great Hawaiian landslides of the past, the energy of which would have been on a par with the largest and most destructive earthquakes. The tsunami results from the transfer of this enormous amount of energy to the water into which the slide drops, giving it the capacity to cause destruction of the opposite sides of oceans. The debris remaining from a tsunami that was found more than 60 m above sea level on the flanks of the Kohala Volcano in Hawaii was tracked down by Gary McMurtry and Dave Tappin on the British Geological Survey, indicating it was a truly awesome tsunami to reach that high up a volcano’s flanks. The height of this enormous tsunami was actually much higher as for about the last 500,000 years the Kohala Volcano, along with the remainder of Hawaii’s Big Island, has been subsiding under its own weight to the extent that at the time of the tsunami about 120,000 years ago the volcano was nearly 350 m higher than at present, thus changing the height reached by the tsunami to more than 400 m. This incredible wave, or more probably a series of waves, resulted from the sliding into the sea of about 500 km³ of the neighbouring Mauna Loa Volcano during the Alika-2 collapse. There are other such deposits on a number of islands in the Hawaiian Archipelago as well as in the Canary Islands. Layers of cobbles and shell material have been found more than 180 m above sea level on the island of Gran Canaria, the result of a huge volcanic collapse in pre-historic times, the author¹ suggesting it could possibly have originated on the island of Mount Teide on the neighbouring island of Tenerife, though the tsunami could have arrived from further away. There may be other evidence of other giant tsunamis as there have been a large number of collapses in the Canaries, the Hawaiian Archipelago, Cape Verde and elsewhere. The author¹ suggests there may be some that have been mapped but misinterpreted as having resulted from other geological phenomena, though it is likely that they have still to be found.

The Cumbre Vieja volcano on La Palma in the Canary Islands a volcano is presently unstable following an eruption in 1949 in which a large chunk began to separate from the main body of the volcano and continuing instability and movement has been found by monitoring. This may become a problem in the future as evidence from ground deformation monitoring indicates. In a paper published by Steve Ward, University of California, Santa Cruz & Simon Day in which the presented the first computer model forecasting the possible scale and extent of a tsunami that would be generated by the flank falling into the sea. The prediction for a worst-case scenario that involves a rock mass comparable in size to the Alika-2 slide from Mauna Loa a tidal dome of water nearly 1 km high would form within 2 minutes of the failure as the wast flank entered the waters of the North Atlantic. After 10 minutes, according to the prediction, a series of waves hundreds of metres high that are spread out across a distance of about 250 km that inundate the shores of La Palma, El Hierro and La Gomera, the 3 westernmost islands in the Canary Archipelago. The coasts of the other islands of the archipelago, Tenerife, Gran Canaria, Fuerteventura, and Lanzarote, would be swamped over the about 1 hour. After about 6 hours the northwest coast of Africa is struck by waves a bit short of 100 m high. Between 3 and 6 m waves strike the shores of Spain and the UK and the coast of Brazil is struck by waves of about 20 m high. About 12 hours after the landslide waves strike the eastern Caribbean and the coast of Florida is struck by a series of waves, about 12 or more, up to about 25 m high.

This predicted scenario depends on a number of factors, principally the mass of the chunk that separates from the volcano, its speed of entry and the persistence of the tsunami as it crosses the ocean. The collapse is expected to move very rapidly, as no slow-moving collapses are known of, either at the historical period or from the geological record. The most heated part of the debate over this potential tsunami concerns the how rapidly the energy of these moving waves will be dissipated and quickly their height will diminish as they radiate from the source. There is a difference between the models about the height of the waves as they reach the US. Ward & Day favour a wave height of more than 20 m, others say they wave heights are likely to be a few metres. The author¹ says it is not known if a warmer, wetter climate would destabilise the western flank of the Cumbre Vieja volcano sufficiently to drop it into the North Atlantic, what can be predicted is that the chances of many glaciated volcanoes around the world being destabilised under such a climate is significantly increased.

Sources & Further reading

1. McGuire, Prof. Bill, 2012, Waking the Giant: How a changing climate triggers earthquakes, tsunamis, and volcanoes, Oxford University Press.