Tsunami risk in East Africa

The 26 December 2004 Indian Ocean tsunami was one of the deadliest natural disasters since the start of the 20th century, with an estimated death toll of around 230 thousand. Millions more were deeply traumatised, bereft of homes and possessions, rendered short of food and clean water, and threatened by disease. Together with that launched onto the seaboard of eastern Japan by the Sendai earthquake of 11 March 2011, it has spurred research into detecting the signs of older tsunamis left in coastal sedimentary deposits (see for instance: Doggerland and the Storegga tsunami, December 2020). In normally quiet coastal areas these tsunamites commonly take the form of sand sheets interbedded with terrestrial sediments, such as peaty soils. On shores fully exposed to the ocean the evidence may take the form of jumbles of large boulders that could not have been moved by even the worst storm waves.

Sand sheets attributed to a succession of tsunamis, interbedded with peaty soils deposited in a swamp on Phra Thong Island, Thailand. Note that a sand sheet deposited by the 2004 Indian Ocean tsunami is directly beneath the current swamp surface (Credit: US Geological Survey)

Most of the deaths and damage wrought by the 2004 tsunami were along coasts bordering the Bay of Bengal in Indonesia, Thailand, Myanmar, India and Sri Lanka, and the Nicobar Islands. Tsunami waves were recorded on the coastlines of Somalia, Kenya and Tanzania, but had far lower amplitudes and energy so that fatalities – several hundred – were restricted to coastal Somalia. East Africa was protected to a large extent by the Indian subcontinent taking much of the wave energy released by the magnitude 9.1 to 9.3 earthquake (the third largest recorded) beneath Aceh at the northernmost tip of the Indonesian island of Sumatra. Yet the subduction zone that failed there extends far to the southeast along the Sunda Arc. Earthquakes further along that active island arc might potentially expose parts of East Africa to far higher wave energy, because of less protection by intervening land masses.

This possibility, together with the lack of any estimate of tsunami risk for East Africa, drew a multinational team of geoscientists to the estuary of the Pangani River  in Tanzania (Maselli, V. and 12 others 2020. A 1000-yr-old tsunami in the Indian Ocean points to greater risk for East Africa. Geology, v. 48, p. 808-813; DOI: 10.1130/G47257.1). Archaeologists had previously examined excavations for fish farming ponds and discovered the relics of an ancient coastal village. Digging further pits revealed a tell-tale sheet of sand in a sequence of alluvial sediments and peaty silts and fine sands derived from mangrove swamps. The peats contained archaeological remains – sherds of pottery and even beads. The tsunamite sand sheet occurs within the mangrove facies. It contains pebbles of bedrock that also litter the open shoreline of this part of Tanzania. There are also fossils; mainly a mix of marine molluscs and foraminifera with terrestrial rodents fish, birds and amphibians. But throughout the sheet, scattered at random, are human skeletons and disarticulated bones of male and female adults, and children. Many have broken limb bones, but show no signs of blunt-force trauma or disease pathology. Moreover, there is no sign of ritual burial or weaponry; the corpses had not resulted from massacre or epidemic. The most likely conclusion is that they are victims of an earlier Indian Ocean tsunami. Radiocarbon dating shows that it occurred at some time between the 11th and 13th centuries CE. This tallies with evidence from Thailand, Sumatra, the Andaman and Maldive Islands, India and Sri Lanka for a major tsunami in 950 CE.

Computer modelling of tsunami propagation reveals that the Pangani River lies on a stretch of the Tanzanian coast that is likely to have been sheltered from most Indian Ocean tsunamis by Madagascar and the shallows around the Seychelles Archipelago. Seismic events on the Sunda Arc or the lesser, Makran subduction zone of eastern Iran may not have been capable of generating sufficient energy to raise tsunami waves at the latitudes of the Tanzanian coast much higher than those witnessed there in 2004, unless their arrival coincided with high tide – damage was prevented in 2004 because of low tide levels. However, the topography of the Pangani estuary may well amplify water level by constricting a surge. Such a mechanism can account for variations of destruction during the 2011 Tohoku-Sendai tsunami in NE Japan.

If coastal Tanzania is at high risk of tsunamis, that can only be confirmed by deeper excavation into coastal sediments to check for multiple sand sheets that characterise areas closer to the Sunda Arc. So far, that in the Pangani estuary is the only one recorded in East Africa

Risks of sudden changes linked to climate

The Earth system comprises a host of dynamic, interwoven components or subsystems. They involve processes deep within Earth’s interior, at its surface and in the atmosphere. Such processes combine inorganic chemistry, biology and physics. To describe them properly would require a multi-volume book; indeed an entire library, but even that would be even more incomplete than our understanding of human history and all the other social sciences. Cut to its fundamentals, Earth system science deals with – or tries to – a planetary engine. In it, the available energy from inside and from the Sun is continually shifted around to drive the bewildering variety, multiplicity of scales and variable paces of every process that makes our planet the most interesting thing in the entire universe. It has done so, with a variety of hiccups and monumental transformations, for some four and half billion years and looks likely to continue on its roiling way for about five billion more – with or without humanity. Though we occupy a tiny fraction of its history we have introduced a totally new subsystem that in several ways outpaces the speed and the magnitude of some chemical, physical and organic processes. For example: shifting mass (see the previous item, Sedimentary deposits of the ‘Anthropocene’); removing and modifying vegetation cover; emitting vast amounts of various compounds as a result of economic activity – the full list is huge. In such a complex natural system it is hardly surprising that rapidly increasing human activities in the last few centuries of our history have hitherto unforeseen effects on all the other components. The most rapidly fluctuating of the natural subsystems is that of climate, and it has been extraordinarily sensitive for the whole of Earth history.

Cartoon metaphor for a ‘tipping point’ as water is added to a bucket pivoted on a horizontal axis. As water level rises to below the axis the bucket becomes increasingly stable. Once the level rises above this pivot instability sets in until the syetem suddenly collapses

Within any dynamic, multifaceted system-component each contributing process may change, and in doing so throw the others out of kilter: there are ‘tipping points’. Such phenomena can be crudely visualised as a pivoted bucket into which water drips and escapes. While the water level remains below the pivot, the system is stable. Once it rises above that axis instability sets in; an external push can, if strong enough, tip the bucket and drain it rapidly. The higher the level rises the less of a push is needed. If no powerful push upsets the system the bucket continues filling. Eventually a state is reached when even a tiny force is able to result in catastrophe. One much cited hypothesis invokes a tipping point in the global climate system that began to allow the minuscule effect on insolation from changes in the eccentricity of Earth’s orbit to impose its roughly 100 ka frequency on the ups and downs of continental ice volume during the last 800 ka. In a recent issue of Nature a group of climate scientists based in the UK, Sweden, Germany, Denmark, Australia and China published a Comment on several potential tipping points in the climate system (Lenton, T.M. et al. 2019. Climate tipping points — too risky to bet against. Nature, v. 575, p. 592-595; DO!: 10.1038/d41586-019-03595-0). They list what they consider to be the most vulnerable to catastrophic change: loss of ice from the Greenland and Antarctic ice sheets; melting of sea ice in the Arctic Ocean; loss of tropical and boreal forest; melting of permanently frozen ground at high northern latitudes; collapse of tropical coral reefs; ocean circulation in the North and South Atlantic.

The situation they describe makes dismal reading. The only certain aspect is the steadily mounting level of carbon dioxide in the atmosphere, which boosts the retention of solar heat by delaying the escape of long-wave, thermal radiation from the Earth’s surface to outer space through the greenhouse effect. An ‘emergency’ – and there can be little doubt that one of more are just around the corner – is the product of ‘risk’ and ‘urgency’. Risk is the probability of an event times the damage it may cause. Urgency is the product of reaction time following an alert divided by the time left to intervene before catastrophe strikes. Not a formula designed to make us confident of the ‘powers’ of science! As the commentary points out, whereas scientists are aware of and have some data on a whole series of tipping points, their understanding is insufficient to ‘put numbers on’ These vital parameters. And there may be other tipping points that they are yet to recognise.  Another complicating factor is that in a complex system catastrophe in one component can cascade through all the others: a tipping may set off a ‘domino effect’ on all the others. An example is the steady and rapid melting of boreal permafrost. Frozen ground contains methane in the solid form of gas hydrate, which will release this ‘super-greenhouse’ gas as melting progresses.   Science ‘knows of’ such potential feedback loops in a largely untried, theoretical sense, which is simply not enough.

A tipping point that has a direct bearing on those of us who live around the North Atlantic resides in the way that water circulates in that vast basin. ‘Everyone knows about’ the Gulf Stream that ships warm surface water from equatorial latitudes to beyond the North Cape of Norway. It keeps NW Europe, otherwise subject to extremely cold winter temperatures, in a more equable state. In fact this northward flow of surface water and heat exerts controls on aspects of climate of the whole basin, such as the tracking of tropical storms and hurricanes, and the distribution of available moisture and thus rain- and snowfall. But the Gulf Steam also transports extra salt into the Arctic Ocean in the form of warm, more briny surface water. Its relatively high temperature prevents it from sinking, by reducing its density. Once at high latitudes, cooling allows Gulf-Steam water to sink to the bottom of the ocean, there to flow slowly southwards. This thermohaline circulation effectively ‘drags’ the Gulf Stream into its well-known course. Should it stop then so would the warming influence and the control it exerts on storm tracks. It has stopped in the past; many times. The general global cooling during the 100 ka that preceded the last ice age witnessed a series of lesser climate events. Each began with a sudden global warming followed by slow but intense cooling, then another warming to terminate these stadials or Dansgaard-Oeschger cycles (see: Review of thermohaline circulation, Earth-logs February 2002). The warming into the Holocene interglacial since about 20 ka was interrupted by a millennium of glacial cold between 12.9 and 11.7 ka, known as the Younger Dryas (see: On the edge of chaos in the Younger Dryas, Earth-logs May 2009). A widely supported hypothesis is that both kinds of major hiccup reflected shuts-down of the Gulf Stream due to sudden influxes of fresh water into North Atlantic surface water that reduced its density and ability to sink. Masses of fresh water are now flowing into the Arctic Ocean from melting of the Greenland ice sheet and thinning of Arctic sea ice (also a source of fresh water). Should the Greenland ice sheet collapse then similar conditions for shut-down may arise – rapid regional cooling amidst global warming – and similar consequences in the Southern Hemisphere from the collapse of parts of the Antarctic ice sheets and ice shelves.  Lenton et al. note that North Atlantic thermohaline circulation has undergone a 15% slowdown since the mid-twentieth century…

See also: Carrington, D. 2019. Climate emergency: world ‘may have crossed tipping points’ (Guardian, 27 November 2019)

Landslides and multiple dangers

English: A rock landslide in Guerrero, Mexico....
A landslide in Guerrero, Mexico in August, 1989. (credit: Wikipedia)

Just as modern humans were establishing a permanent foothold in Britain and engaging in the transition to settled farming and livestock husbandry disaster struck some of the most attractive Mesolithic real estate. Around 8 000 years ago the east coast of Scotland, from the Shetland Isles to the Firth of Forth, was struck by a tsunami as big as that affecting the north eastern island of Honshu in the Japan archipelago in 2011. It washed over low lying islands of Shetland and Orkney and roiled up the great inlets or firths of eastern mainland Scotland to leave thick sand deposits containing carcases of whales and other large sea mammals. At that time, Britain was joined to the rest of Europe by marshy lowlands linking East Anglia and the Netherlands dubbed ‘Doggerland’ at the southern end of a huge gulf that became the North Sea. Final sea level rise removed that initial gateway to Britain, so we cannot judge what damage the tsunami wrought, but tools and animal bones dredged from the area show that it was full of game and people. A disaster, but not one linked to seismicity. The driving force has been recognised in a series of submarine scars off the west coast of Norway that witness massive slides of sediment on the sea bed area known as Storegga. Similar scars around the Hawaiian Islands and those making up the Azores and Canaries in the mid Atlantic bear witness to many large slippage events, on the sea bed and from the islands themselves. Recognising signs of past tsunami damage in coastal areas worldwide reveals plenty of cases triggered by landslides rather than earthquakes.

The March 2011 Sendai tsunami and those which ravaged lands around the Indian Ocean in late 2004 formed because of vertical movements on major faults that dropped or shoved up the oceanic crust itself. Yet any sudden change in the shape of the sea floor will displace all the ocean water above, the difference from seismic tsunamis lies in the energy source: instead of tectonic plate forces, gravitational potential energy is released by slumps and slides. That may happen because of erosion producing unstable steep slopes, build up of sedimentary piles, large outpourings of lavas or slopes being destabilised by minor earthquakes or release of gases from the sediments themselves. The Mesolithic submarine slide at Storegga may have been set in motion by massive release of methane from gas-hydrate deposits, and such is the extent of scarring of the sea floor there that it must have happened before and may do so again.

1755 copper engraving showing Lisbon in flames...
Copper engraving showing the 1755 Lisbon tsunami overwhelming ships in the harbor. (credit: Wikipedia)

Realisation of the potential for tsunamis to be triggered by submarine and coastal and slides has spurred bathymetric studies in a number of likely areas, including the Gorringe Bank that lies on the Atlantic floor just west of the Iberian Peninsula. It is tectonic in origin but has a thick veneer of sediment brought by Iberian river systems. On its northern flank is a 35 km long scar of a slip that moved 80 km3 of sediment (Lo Iacono, C. And 11 others 2012. Large, deepwater slope failures: implications for landslide generated tsunamis.  Geology, v. 40, p. 931-934). The Spanish-British-Italian group estimate that the slip would have generated a 15 m tsunami most likely to have affected the Iberian coast south of Lisbon. Conditions for slides of si,ilar magnitude still exist on the Gorringe Bank. One unstable system ripe for collapse is present far out in the Atlantic on the south-east coast of the island of Picos in the Azores (Hildenbrand, A. et al. 2012. Large-sale active slump on the southeast flan of Picos Island, Azores. Geology, v. 40, p. 939-942). This is in a coastal area where repeated volcanism has piled up lavas on the flanks of the island’s main volcanic edifice. Failure has already started, with a number of prominent arcuate scars having developed. The Picos slide moves very slowly sideways but vertical displacements ar estimated at up to a centimetre a year. The volume of the slowly moving mass is an order of magnitude less that the fossil slide on the Gorringe Bank. Yet should it fail entirely, the slopes involved, the absence of water’s slowing effect and the height of the mass might ensure comparable energy is delivered to the Atlantic Ocean, though the likely trajectory of tsunamis would be parallel to the coast of Africa rather than directly towards it.

Landslides of all kinds, though hazardous, have long been thought to be less of a risk to life globally than the more spectacular seismic and volcanic hazards, but there are few data to support that view. In an attempt to assess the annual risk properly, David Petley of Durham University, UK ‘mined’ world-wide landslide records for the seven years since 2004 (Petley, D. 2012. Global patterns of loss of life from landslides. Geology, v. 40, p. 927-930). There were more than 2600 recorded slope-failures that killed people and caused a total of more than 32 thousand fatalities: ten time more than previous vague estimates. This is a minimum because many landslides occur in very remote areas, especially in the mountainous regions of China and the Himalaya. The number of fatalities accompanying each event shows distinct signs, on a country-by-country basis, of a relationship with population density. Several international agencies are emerging that aim at means of measuring disaster risk, one being the Integrated Global Observing Strategy for Geohazards (IGOS).