How seabed mapping is shedding light on a silent tsunami risk
July 17, 1998, is a day that the people of Papua New Guinea will not forget anytime soon. At 6.49 pm local time, a magnitude 7.0 earthquake struck, shaking foundations, cracking open the ground, and damaging buildings along the Aitape coast. Shortly after, a large boom. Then, a roaring sound. The sea boiled and receded. Then it struck - a tsunami comprising three waves, the largest of which reached heights of 10-15 metres. By 7.25 pm, it was all over. Over 2,200 people are thought to have lost their lives that day, with another 1,000 seriously injured by the power of the wave, sand abrasion, and debris. The ten thousand survivors had to resettle elsewhere.
“The Papua New Guinea tsunami was unusual,” says Professor David Tappin, tsunami researcher with the British Geological Survey. “An earthquake of magnitude seven shouldn’t generate waves up to 15 metres high.” In the wake of the tsunami, Tappin led one of five expeditions that mapped the seabed off Papua New Guinea to try and understand what had happened. What those maps revealed changed our perspective on tsunamis forever.
Multibeam reveals a hidden generator
“Previous to our expedition, most scientists thought it was only earthquakes and sometimes volcanoes,” Tappin explains, noting that meteors can also cause tsunamis but haven’t done so for millions of years. Indeed, most tsunamis are caused by earthquakes. A small number – less than a hundred over the past few centuries – have been caused by volcanoes. What Tappin’s seabed mapping found was a submarine mass failure - a submarine landslide approximately 4.6 kilometres long, 4.2 kilometres wide, and with a volume of around 6.4 cubic kilometres.
Submarine landslides act much like landslides or avalanches on land - a huge amount of sediment, rock, and other debris rapidly moves downslope off the continental shelf, reshaping the seabed. “The background is geological. We have erosion on land and rivers flowing into the ocean, bringing sediments and other deposits that accumulate on the continental shelf. If that sediment becomes unstable and fails, you get a submarine landslide,” Tappin explains. The reasons sediments may become unstable and fail are many, including the flow and changes in pressure of water or gas in the sediment, hurricanes, and earthquakes.
Scientists thought that submarine landslides in general, let alone one of the size found off Papua New Guinea, would be incapable of generating a tsunami because they couldn’t generate the velocity needed. “With colleagues, we developed a numerical model that demonstrated the initial acceleration of the landslide generates sufficient velocity,” Tappin explains.
Since that work, seabed mapping has identified numerous landslides that may have played a role in tsunamis. Perhaps the most famous are those located off the Norwegian coast. Formed around 6,200 BCE, the total volume of the three Storegga Slides (Storeggaraset) is approximately 3,500 cubic kilometres. Their formation has been linked to a significant tsunami event in the region. “The scales of some these landslides are just enormous, but as we’ve seen with Papua New Guinea, even a small one can generate a dangerous event,” says Tappin. When an earthquake generates a tsunami, any subsequent submarine landslides can add to the devastation.
When the 2011 Tōhoku earthquake struck off the east coast of Honshu Island, the largest and most populated island in the Japanese archipelago, it generated a tsunami with a runup (the maximum vertical height onshore that a tsunami reaches) of over 20 meters along the Sendai Plain. Using a multibeam echosounder, “we identified a submarine landslide [60 kilometres off the Sanriku coast farther north] which had been triggered by the earthquake,” says Tappin. Forty kilometres wide, 20 kilometres long, and two kilometres thick, the landslide “appears to have added an extra 20 metres of runup height to the tsunami.”
Sounding the alarm
Nothing can be done to prevent a tsunami from striking, but some things can be done to minimise the loss of life. In 1946, the USA set up the first tsunami warning system in the Pacific Islands in response to an earthquake-generated tsunami that struck the Aleutian Islands. When the 2004 Indian Ocean tsunami struck, no warning system was in place. Like the Aleutian Islands, the event has since prompted significant investment in a tsunami warning system. “These warning systems are based on earthquakes. It’s something you can monitor relatively easily 24/7,” says Tappin. Once an earthquake is detected, dedicated organisations determine if an earthquake is likely to cause a tsunami. If there is a risk, mitigation plans are enacted. “The 2011 Japanese tsunami was devastating. They underestimated the potential for the largest magnitude earthquake, so more people died than maybe should have. But if they didn’t have any mitigation strategies, it would have been much worse,” says Tappin.
The amount of warning depends in part on the earthquake’s location. “If there is an earthquake just offshore of Chile, it could strike land in about 20 or 30 minutes, but it might strike Hawaii in about 10-12 hours and Japan about 10 hours after that,” Tappin explains. Where and when a tsunami may strike is also partly driven by the seabed.
First, there is the direction of waves. “One of the most important aspects is the mid-ocean ridges. The waves bend and refract on these ridges, so what starts as a linear wave can change quite dramatically,” says Tappin. Then there is speed. “The deeper the water, the faster they travel. When they’re on the continental shelf – 200 metres depth, they travel at about 30 metres a second. In the deep ocean, say at 3,000 metres depth, they travel at 170 metres a second.” Finally, as the wave approaches the coast, “that’s where the danger starts. As the wave moves across the continental slope onto the shelf, you get this frictional event from contact with the seabed, which causes the water velocity to slow. The wave then builds upwards,” Tappin explains. If the earthquake is large enough, when the wave strikes land, it can be tens of metres high.
Unfortunately, designing warning and mitigation systems for tsunamis driven by submarine landslides is much more problematic. First, we don’t have a method to detect submarine landslides like earthquakes. Second, because submarine landslides occur on the continental shelf, the amount of time people will have to react may be short. Finally, the science of submarine landslide-driven tsunamis is still relatively young, and the events are infrequent, so we don’t fully understand the mechanisms.
Modelling the risks
An estimated 40% of the world’s population lives within 100 kilometres of the coast. In some countries, that percentage is much higher. In Australia, for example, 87% of the population lived within just 50 kilometres of the coast. “Fortunately, we don’t have a huge tsunami risk in Australia, and the majority that do impact Australia have been quite small,” says Dr Kendall Mollison, associate lecturer and postdoctoral researcher at the University of Newcastle, Australia. The last tsunami to hit Australia, for example, was generated by the 2022 Hunga Tonga–Hunga submarine volcanic eruption. Those wave heights reached just 80 centimetres.
Mollison has spent the past six years studying submarine landslides and their potential for generating tsunamis. By analysing submarine landslides from high-resolution bathymetric data and sediment cores collected from the East Australian continental shelf, Mollison has been able to build tsunami models. “The models help us understand what sorts of impacts we might see from a variety of different events and locations and answer questions like, if these landslides were to occur in a modern-day setting, what sort of tsunami waves could we be expecting, what sort of amplitudes and water velocities, what sort of impact” Mollison explains, stressing that the risk of one of these events happening is so small, “it’s not something to lose sleep over.”
Models perform best when the underlying data is high-quality. “We need to have a really good idea of the morphology of that landslide, the extent, and the volume of material. The higher the resolution the seabed mapping data is, the more confident we can be in our model,” says Mollison. “Some of the submarine landslides I’ve modelled are relatively near the shoreline. As [any resulting] tsunami propagates into shallower water, the morphology of those shallow water environments and the morphology of the continental shelf becomes really important. Again, the higher the resolution of that data, the more confident we can be in our modelling.”
Recently, Mollison has started working on submarine landslide assessments across the Tasman Sea. A partnership between GNS Science, NIWA, the University of Auckland, iwi partners Ngāti Tama and Ngāti Mutunga, and Australian researchers like Mollison, the Silent Tsunami project aims to improve our understanding of the causes of submarine landslides, how often they may occur in the Tasman Sea, and the potential impacts to the west coast of New Zealand if one were to occur. For Mollison’s part of the project, “we’re modelling some really large submarine landslides identified off the West Coast of the North Island through seismic data,” says Mollison. “We’re now modelling these landslides to understand what sort of tsunami would be associated with such large landslide events.” So far, preliminary modelling suggests that a tsunami created by these landslides would have impacted the east coast of Australia, too.
All this research would not be possible without seabed mapping. In 2022, Mollison was part of a team on board the Australian research vessel Investigator, where they mapped some 40,000 square kilometres of seafloor. “To map such a massive area is really important in terms of hazard research,” says Mollison. “This is why the ['The Nippon Foundation-GEBCO] Seabed 2030 project is so important. The more high-quality data we have, the more we can understand and hopefully reduce the risks [of tsunami events] for coastal populations.”
