Tsunami warning: why prediction is so hard

By Natalie Muller May 11, 2012
Reading Time: 5 Minutes Print this page
Predicting tsunamis is a complex task, but scientists’ efforts in improving warning times are saving lives.

IN 2004, A MAGNITUDE 9.1 earthquake off the coast of northern Sumatra, Indonesia, generated the worst tsunami in recorded history. With more than 230,000 people dead and many more hundreds of thousands injured and homeless, this natural phenomenon wreaked widespread destruction.

At the time, people were caught off guard – the world had not seen a massive tsunami for generations and did not recognise the danger. Since then, a heightened awareness has brought tsunamis into consciousness, with almost every earthquake reported having an associated tsunami risk attached to it.

The enormous tsunami in Japan in March 2011 further highlighted the power of big tsunamis. But the most recent magnitude 8.6 earthquake off the coast of Sumatra, Indonesia on 11 April, 2012 – which didn’t generate a destructive tsunami – showcases just how complex tsunami generation can be. Not to mention their prediction.

A cluster of earthquakes?

It takes a massive disruption of the Earth’s crust beneath the sea floor to displace enough water to generate a tsunami. Once in motion, these earthquake-triggered waves can race up to 1000km/h in open water, devastating coastlines in their path.

But they’re rare. On average, destructive tsunamis occur twice a year, and tsunamis that cause ocean-wide devastation, like the one in 2004, happen only once every 15 years, on average.

Paul Somerville, deputy director of Risk Frontiers Natural Hazards Research Centre at Macquarie University, Sydney, says it’s significant that there have been a series of large earthquakes triggering tsunamis in recent years – Sumatra in 2004, Chile in 2010 and Japan in 2011.

“We’re interested in whether it shows that these earthquakes cluster in time or not, and that’s a raging debate at the moment,” he says. “The previous big earthquakes all happened in the late 1950s, 1960s, and since then there haven’t been any magnitude 9 earthquakes.”

Tsunami warning system for the Indian Ocean

Historically, over 80 per cent of tsunamis have happened in the Pacific, especially around the Ring of Fire, and for decades there’s been a Pacific Ocean Tsunami Warning System keeping an eye on the region.

One of the reasons for the huge loss of life in 2004 was the absence of such a system in the Indian Ocean.

“A lot of people died in places like Kenya, India, Sri Lanka, Thailand, Maldives, Madagascar, Tanzania,” says Daniel Jaksa, co-director of the Joint Australian Tsunami Warning Centre (JATWC) at the Geoscience Australia hub in Canberra. “It was 15 hours’ travel time [for the tsunami to reach] Africa, so those people should have been warned. If it happened now, I’d be amazed to see any deaths away from the epicentre.” 

In 2005, the Indian Ocean community set in motion a plan to develop their own warning system, and there are now centres monitoring potential tsunami threats in Australia, India and Indonesia.

Predicting tsunami waves

Like any earthquakes, there’s no way of predicting when tsunami-causing quakes will strike, but thanks to these early warning systems, it’s now possible to get word out about an approaching tsunami within minutes.

Not all big underwater earthquakes will cause big waves, but Daniel says a tsunami will generally happen if there’s a significant earthquake (magnitude 7 or above) in a subduction zone where the tectonic plates meet.

The April 2012 magnitude 8.6 earthquake off the coast of Sumatra didn’t generate a tsunami because it occurred away from the plate boundary. And, unlike the 2004 earthquake, which forced a 1300km-long chunk of the sea floor upward, the more recent earthquake only caused the two plates to grind together horizontally.

Still, this underwater seismic activity put the Indian Ocean tsunami warning system to the test. The earthquake struck on 12 April at 08:38 GMT, and within five minutes, the Indonesian Meteorological Service had issued a tsunami warning across the region. This was followed minutes later by warnings from other centres in India and Australia, as well as from Japan’s Meteorological Agency and the Pacific Tsunami Warning Centre in Hawaii.

People closest to the epicentre scrambled to higher ground, expecting a repeat of the 2004 disaster, while warnings were in place along Africa’s west coast, Thailand, India, Sri Lanka, Indonesia, the Middle East and many Indian Ocean island communities. But the anticipated giant tsunami never came.

Scientists working at the region’s warning centres downscaled the threat after local sea level gauges and deep ocean sensors confirmed no huge tsunami had been generated.

Profiling an incoming tsunami

The JATWC, an organisation formed between Geoscience Australian and the Bureau of Meteorology (BOM), has an arsenal of tsunami-tracking tools at its disposal. Seismic information from Geoscience Australia is combined with ocean data from the BOM to assess a tsunami risk; alerts are then issued via media channels, including the ABC, within 15 minutes of a quake occurring.

Ocean data from DART (Deep-ocean Assessment and Reporting of Tsunami) buoys moored to sensors on the sea floor monitor pressure shifts caused by a passing tsunami and beam changes in sea-surface height back to scientists.

“This gives an idea of the speed, which [scientists] can estimate very closely based upon a whole series of pre-computed scenarios and tsunami travel times around the ocean,” says James Goff, co-director of the Australian Tsunami Research Centre and Natural Hazards Laboratory at UNSW.

The tsunami scenario database is one of the more recent additions to the warning system. After an earthquake, the closest scenario is chosen from thousands of situations, and scaled to match the tsunami’s magnitude. This has made it much easier for scientists to build a detailed picture of a tsunami’s behaviour, such as how tall the waves might be (amplitude) and where they’re headed. But even with current technology, it’s difficult to predict what the impact and magnitude of the waves will be on land, says James.

“For tsunamis generated a good hour or so away from the coast, there is now a system in place,” he says. “Of course, tsunamis can be generated by other mechanisms such as landslides and volcanoes, and also if a tsunami is generated close to the shore then there will be little warning anyway and the system cannot help in those cases.”

World-first warning system

Geoscience Australia is currently building a world-first warning system out of an array of seismometers in the remote Pilbara region of WA that aims to make more accurate tsunami predictions possible. The instruments will monitor earthquakes in the Indian Ocean and beam seismic signals from the initial rupture to the warning centre.

“Essentially what we’re doing is defining the epicentre,” says Daniel. “Mapping that rupture…defines very much what type of tsunami forecast you can do in terms of arrival times, [and] locations of coastline that are impacted by the tsunami.”

But gathering data about an approaching tsunami is only part of the battle. The other part is educating people about how to respond to warnings. Tsunamis may be rare and unlikely events, but adequate preparation could save thousands of lives.

James says Japan is “the best tsunami-prepared country in the world,” he says. “Those countries that have had a recent event are, not surprisingly, quite responsive and people have a better sense of what to do, but others – such as Australia for example, are very poor. The technology is great but it’s the last mile that’s missing. In other words, you can give the message to people, but unless they understand it and know what to do, then nothing has changed.”