An elephant pulls debris near the coast of Banda Aceh in Indonesia, after the 2004 Boxing Day tsunami. Credit: Phillippe Desmazes/AFP/Getty

When a magnitude-9.1 earthquake shuddered to life off the Sumatran coast on 26 December 2004, there was no systematic way to alert communities across the Indian Ocean that a devastating wave might be coming. Afterwards, with some 230,000 people dead and US$14 billion in damages, international disaster experts resolved to reduce the toll next time a tsunami struck.

Science pours in from Rosetta comet mission GM microbes created that can’t escape the lab Crunch time for pet theory on dark matter

Ten years on from the deadliest tsunami in history, almost all the countries bordering the Indian Ocean are hooked into a network of seismometers, sea-level gauges and satellite-linked buoys. In close to real time, this Indian Ocean Tsunami Warning and Mitigation System (IOTWS) notifies nations from Indonesia to Sri Lanka to Oman when a big offshore earthquake has occurred and determines whether it might generate a tsunami. Were the 2004 earthquake to happen today, these nations would be much better prepared.

But despite its technical sophistication, the tsunami warning system remains vulnerable. The initial rush of funding from international donors is drying up, and Indian Ocean nations now face the responsibility of maintaining the system — to the tune of between $50 million and $100 million per year. “We’re definitely safer than we were in 2004,” says Rick Bailey, head of tsunami warning services at the Australian Bureau of Meteorology in Melbourne. “But sustainability will be the next big issue for us.”

The geophysical components of the Indian Ocean tsunami-alert system generally work well (see ‘Early warning’). More than 140 seismo-meters constantly monitor earthquakes around the basin, including in the quake-prone ‘subduction’ zones off Indonesia and the coast of Pakistan, where one plate of Earth’s crust grinds under another. When a big tremor hits, three regional alert centres — in Australia, Indo-nesia and India — spring into action. Scientists there use seismic data to estimate how much the earthquake has displaced the ocean floor. Then they compare the real quake with model scenarios in which they have calculated what size of tsunami might be produced. The centres alert national governments about what to expect, and data from coastal sea-level gauges and a handful of tsunameters — buoys floating in the open ocean that can detect the passage of large waves — can help to confirm whether a major swell is making its way across the ocean.

What happens next is up to each country — but warnings often fail to travel the ‘last mile’ to people living in areas, often remote, that are at risk of being swamped. “We really do need to focus on that last mile,” says Tony Elliott, head of the warning system’s intergovernmental coordination group in Perth, Australia.

In tsunami-prone Indonesia, a German–Indonesian team has worked to develop warning communication chains in 26 provinces and districts. Seven years into the project, only about half of those 26 had implemented a functional warning service that reached all the way down to the local level, says Harald Spahn, a disaster-management consultant formerly with the German development agency GIZ.

And even when alerts do make it to people at risk, those people do not always behave as disaster experts would wish. In April 2012, a magnitude-8.6 earthquake hit off the coast of Sumatra. Instead of going to shelters, as emergency managers had hoped, many people tried to drive away. The roads in Aceh province became clogged. Fortunately, the geology of that quake meant that it produced only a very minor tsunami.

In Indonesia, Spahn and his colleagues focused on four pilot regions to develop ways to complete the communication chain. They developed tsunami-hazard maps to work out which communities were most at risk. Then they devised a brochure that lays out the warning signs of an approaching tsunami and what to do when one might be on the way. Finally, they helped to develop a three-tier alert system that was adopted at the national level. The tiers depend on the height of the expected tsunami, and specify the action that government officials should take — such as to move people off and away from the beach, evacuate in a limited fashion, or evacuate completely.

Spahn says that the tsunami-alert system can be useful even when no tsunami is coming. In September 2009, a magnitude-7.6 earthquake killed more than 1,100 people in and around the city of Padang on the western coast of Sumatra. The tsunami-alert system indicated that there would be no big wave, which let emergency officials respond more quickly to the earthquake damage.

Maintenance will be key to keeping the information flowing. The IOTWS cost more than $450 million to set up, with most funding coming from Australia, Indonesia and India. If a piece of equipment breaks, it is up to the country that installed it to fix it. The deep-sea buoys in particular are expensive and prone to vandalism or accidental damage from passing ships.

The Indian Ocean countries have varying levels of motivation to keep the system going, Elliott notes. Nations that are farther from likely sources of great earthquakes are less engaged.

Experts say that the best chance of keeping the system operating for the next decade and beyond is to make sure that tsunami alerts are woven into the national fabric for dealing with other kinds of emergency, from cyclones to landslides, many of which use the same sensing networks and communication channels.

 “We’ve done a lot,” says Bailey. “We’ve just got to hang onto it now.”