10 years after the Indian Ocean Tsunami: What have we learned?

Last updated:24 April 2023

Though no one realised it at the time, the 2004 Sumatra-Andaman earthquake was the first in a series of massive earthquakes to shake the globe. At magnitude 9.1 it is among the three largest earthquakes ever recorded since instrumental recordings began at the turn of the 20th century. The following seven years saw the occurrence of another three of the ten largest earthquakes ever recorded (including the giant earthquake and tsunami in northeast Japan in 2011). This sequence of four megaquakes occurring over 2004-11 rivals another series that occurred over 1952-1965, when another four of the ten largest earthquakes ever recorded occurred around the Pacific Rim - including a magnitude 9.5 offshore Chile in 1960 - the largest earthquake ever recorded.

We still know very little about why megaquakes occur in clusters like this. One explanation for the recent large events is the mechanism of stress transfer, in which the Sumatra subduction zone 'unzipped' in a series of massive earthquakes that ruptured sequentially from northwest to southeast from 2004-07. But this does not explain the presence of a large 'gap' in central Sumatra. In this section the recurrence of earthquakes that occurred in 1797 and 1833 was - and still is - widely anticipated, not only by scientists but by the many residents of the city of Padang, who are crowded into a low-lying coastal strip that will almost certainly be inundated by the resulting tsunami. However, the 'unzipping' of the subduction zone mysteriously skipped this segment during the 2004-07 sequence. Likewise, we don't know what role, if any, the 2004 Sumatra Andaman earthquake may have played in triggering distant events like the 2010 Maule, Chile and the 2011 Tohoku, Japan earthquakes. And we don't know if the present sequence of megaquakes is at an end.

What we do know is that the 2004 Sumatra Andaman earthquake generated a massive tsunami - the Indian Ocean Tsunami - that killed over 227 000 people, more than ten times the number of lives lost in the remaining nine of the ten largest earthquakes combined (the total fatalities for these was about 21 000). Why did the Indian Ocean Tsunami kill so many compared to other earthquakes of similar size? First, the earthquake occurred just offshore of a major population centre. The population of Banda Aceh before the tsunami was over 264 000, and these in addition to the populations of towns along the western coast of Aceh were severely affected by the tsunami. The town of Lhok Nga, where observations of tsunami run-up height reached over 30 metres, had a pre-tsunami population of 7000, reduced to 400 after the tsunami. Banda Aceh itself suffered over 61 000 fatalities, almost 25% of its population. In all, Indonesian fatalities are thought to number at least 167 000 (estimates range as high as 220 000), over 70% of the total Indian Ocean Tsunami fatalities. Even when only the Indonesian fatalities are considered, the Indian Ocean Tsunami is the world's deadliest tsunami disaster.

But the Indian Ocean Tsunami was unique among tsunami disasters in the scale of fatalities caused on a regional scale. Because the rupture extended far north from Sumatra into the Andaman Sea, both India and Sri Lanka in the west, and Thailand in the east, were directly in the path of the main plume of tsunami energy. In addition to the 167 000+ fatalities in Indonesia, over 61 000 died in Sri Lanka, India and Thailand. Around 2000 Europeans, many tourists visiting Thailand, were killed, including over 500 each from Sweden and Germany. 26 Australians also died while overseas in southeast Asia, and dozens were swept to sea by the large waves and strong currents generated when the tsunami reached Australia's western coast.

If the Indian Ocean Tsunami occurred today, the number of lives lost would be significantly less than the devastating number of fatalities seen in 2004

In addition to the large coastal populations exposed to the tsunamis, the major contributing factor to the massive loss of life was a lack of preparedness. A large tsunami in the Indian Ocean was not without historical precedent. The tsunami generated by the eruption of Krakatau in 1883 killed over 35 000 people along the Sunda Strait separating Java and Sumatra. Massive earthquakes in 1797, 1833 and 1861 had occurred off Sumatra, and both the earthquakes and the tsunamis they generated were well documented by Dutch historians. However, these earthquakes had occurred well south of Aceh. While these large local tsunamis destroyed coastal villages, there were at that time no major population centres along this part of the Sumatra coast. The population of Padang, now over 800 000, was only 4000 in 1797. Furthermore, the ruptures of these earthquakes were too far south to have affected Thailand, India and Sri Lanka. So, while there had been large earthquakes and tsunamis in the Indian Ocean, there was no historical precedent for a tsunami affecting large coastal populations. As a consequence there was no warning system, and coastal populations did not know to evacuate low-lying coastal areas in the event of a large earthquake. An exception was the island of Simeulue, to the west of Aceh, where an oral tradition preserved from experience of a smaller tsunami in 1907 caused residents to run to higher ground when they felt the earthquake, saving many lives. In retrospect, it seems clear that better preparedness could have prevented many 10 000s of deaths.

Much has changed in terms of preparedness in the ten years since the 2004 Indian Ocean Tsunami. A warning system for the Indian Ocean has been established, and many at-risk populations are well aware of the danger of tsunamis, and in many cases are drilled in evacuation procedures. The tsunami risk is taken seriously even in subduction zones that have not historically experienced a megaquake. Were an event like the Indian Ocean Tsunami to occur again today, it seems extremely unlikely that the fatalities caused at regional and greater distances would be anywhere near the scale of the death toll in India, Sri Lanka and Thailand in 2004. This is because, with lead times of several hours between detection of an event and its impact on regional or distant shores, conventional tsunami warning systems are generally very effective.

However, it is important to bear in mind that over 70% of the Indian Ocean Tsunami fatalities, 167 000 or more, were killed by the local tsunami that arrived on the shores of Sumatra within minutes after the earthquake rupture. Local tsunami warning remains a hideously difficult problem, in which decisions must be made and warnings disseminated within minutes, and coastal populations must evacuate within 10s of minutes. This in urban areas that are congested at the best of times, and possibly impassable after suffering the effects of a major earthquake. False alarms are inevitable, and the consequent erosion of public confidence in warning systems are difficult to avoid. Japan's experience of the 2011 Tohoku earthquake and tsunami serves notice that even in the presence of the best warning systems, sophisticated communications and tsunami-aware coastal communities, local tsunamis are still able to inflict massive fatalities. Indonesia and its neighbours can minimise losses by strengthening all these elements of mitigation, but it is still a big challenge to completely avoid high fatality events.

Tsunamis are not the only hazard that can cause massive fatalities. Large as the Indian Ocean Tsunami death toll in Aceh was, there are at least 40 cities in Indonesia larger than Banda Aceh, including the megacity of Jakarta. By virtue of location many of these cities may be relatively sheltered from tsunamis. But they, like many of their cousins in neighbouring countries, have highly concentrated urban populations that typically reside in poorly constructed, masonry homes prone to collapse if subjected to strong earthquake ground motion. Such strong ground motion does not have to come from a megaquake: 316 000 deaths were caused in Port-au-Prince by the 2010 Haiti earthquake, with a magnitude of 'only' 7. Virtually every city in the belt of active tectonics stretching from the Himalayas, through Bangladesh and Burma, Indonesia and the Philippines, as well as much of Papua New Guinea, could potentially experience such an earthquake.

Is a massive-fatality earthquake/tsunami event in the Southeast Asian region inevitable in the 21st century, and if so are mitigation efforts pointless? The explosion in population and urbanisation over the latter half of the 20th century in such a seismically active area would indeed seem to make the eventual occurrence of such a mega-disaster all but certain. So the question then becomes "are mitigation efforts worthwhile"? Absolutely. If the Indian Ocean Tsunami were to occur today, it is likely that the regional warning system would reduce the 61 000 fatalities at regional and greater distance to at most a few thousand. Even if local warning and evacuation procedures were only partially successful, they could reduce fatalities to tens rather than hundreds of thousands. If the effectiveness of tsunami warning systems and community awareness can be maintained over the long term, and if this can be combined with improved building practices, hundreds of thousands - if not millions - of lives can be saved. But we need to draw upon these efforts now, while memories of the Indian Ocean Tsunami are still fresh and it is still possible to channel some of the region's recourses into mitigation efforts. It's the best thing we can do to give meaning to such a devastating natural hazard that changed the lives of so many on that day in 2004.