What’s changed, are we more prepared?
On the 26 December 2004, the third largest earthquake ever recorded by seismographs with a magnitude of Mw9.2 struck 85km off the coast of Northern Sumatra, Indonesia at a 30km depth. The fault rupture was over 1,500km long[1], shook for nearly 10 minutes and produced a maximum MMI of IX (Violent).
In total, over 227,000 fatalities[2], hundreds of thousands were injured, and 1.7 million people were displaced across 14 countries with economic losses estimated at $9.9 billion ($16.5 billion in 2025).[3] However, the majority of these human and monetary losses were not caused by the earthquake shaking itself, but by the resulting tsunami. Waves were reported to exceed 20 metres and travel up to 3km inland[4], inundating populated areas with debris and seawater.
Mw9.2 third largest earthquake ever recorded, off coast of Northern Sumatra, Indonesia
1500km length of fault rupture
227k fatalities exceeded
1.7m displaced population across 14 countries
$9.9b estimated economic loss ($16.5 b in 2025)
One of the reasons for such high human and monetary losses was the lack of centralised Indian Ocean tsunami warning system at the time. Furthermore, although the Pacific Tsunami Warning Center (PTWC) detected the earthquake and was aware a large tsunami was likely generated, they had no capacity to warn the locations that would be affected.[5] Countries were unprepared, and when waves began rising, there was no procedure for where people should go or what they should do. So, what advancements have there been in the last 20 years to ensure similar tragedies won’t be repeated in the Indian Ocean or elsewhere around the world? What scientific research is being done to improve the modelling of potential tsunami waves and the resilience of populations and infrastructure? How have models and markets changed?
Building on its experience establishing the Pacific Tsunami Warning System in 1965, the United Nations Educational, Scientific and Cultural Organization Intergovernmental Oceanographic Commission (UNESCO-IOC) began creating a global warning and mitigation system to coordinate locally established warning systems and procedures for evacuation, minimizing risks from future tsunami events.[6],[7]
The IOC-UNESCO Tsunami Programme supports its 150 member states in assessing tsunami risk across the Caribbean, Indian, Pacific, Mediterranean and Northeastern Atlantic Oceans. This is done through a number of service providers that monitor seismic and sea level activity in order to issue timely tsunami threat information within an Intergovernmental Coordination Group (ICG) framework, to National Tsunami Warning Centres and one another operating within an ocean basin. There are also multiple information centres that provide education, outreach, technical and capacity building assistance to member states and the public, for how best to prevent, prepare for, and mitigate against tsunamis.
The result of these schemes is that 100 communities, emergency management agencies and governments across 34 countries have been recognised as ‘Tsunami Ready’ by UNESCO[8],[9] according to 12 indicators covering the assessment, preparedness and response.[10] This readiness is also assessed in worldwide tsunami exercises and drills, the largest of which occurred in the Caribbean in 2019 and had 800,000[11] participants. UNESCO aims to make all at-risk communities Tsunami Ready by 2030, and although no small task, the progress made since the programme’s creation in 2015 suggests this may be achievable.
Examples of how UNESCO’s programme has improved tsunami preparedness measures can be seen in Japan. Depending on the magnitude, hypocentre location and seismic intensity of an earthquake, the Japanese Meteorological Agency (JMA) can issue different levels of tsunami warning or advisory for sections of coastline within two to three minutes of occurrence. If tsunamis are generated by seismic events far from Japan, the Agency engages in coordinated action with the Pacific Tsunami Warning Centre (PTWC) in Hawaii, to issue warnings for tsunamis propagating over long distances.[12] This was seen following the 2022 eruption of the Hunga Tonga–Hunga Ha‘apai volcano, which triggered warnings across much of the Pacific as NOAA tide gauges showed elevated wave heights.[13] As detailed by UNESCO, these warnings provide information for communities, residents and policymakers on if, where and how far they may need to evacuate based on estimated wave heights and runup distances.
Risk managers have benefited from significant improvements in tsunami hazard research and model development since 2004. In particular, there is now an improved understanding of tsunami trigger mechanisms, wave propagation and their effects on coastlines, buildings and infrastructure.
Furthermore, researchers have developed models that account for the complex interactions between landslides and the surrounding water, improving predictions of landslide-induced tsunami hazards. However, one area that has seen less research is volcanically generated tsunamis, particularly in the Pacific Ocean, even if the Tonga eruption did demonstrate effective dissemination of tsunami warnings and advisories.
Considerable improvements in bathymetry and topographic information have improved the realism and accuracy of inundation modelling. The data can be used in coupled models which integrate various physical processes, such as wave propagation, fluid-structure interaction, and structural response. Remote sensing has improved the understanding of where assets and infrastructure are, and their respective vulnerabilities. Coupled with improved modelling this enhances the understanding of potential losses for insurers and helps risk managers to prioritise and understand and mitigate the potential impacts of an event.
As well as enhancing understanding of potential future events, these developments can also be brought together with real time data from observation systems like wave buoys for warning purposes. Many of these developments would be unthinkable without huge increases in computing power for relatively low cost. Although projects such as the General Bathymetric Chart of the Oceans (GEBCO) aim to collate all available bathymetric data to produce a definitive map of the world’s ocean floor by 2030, this is still ongoing.
While these advancements in research have improved the ability to model tsunami, the translation into mainstream catastrophe models has been piecemeal. There is an understandable tendency to roll out detailed modelling first in countries like New Zealand, Japan and the United States, where insurance markets are more established. However, in general model resolution has improved the representation of events and understanding of exposure and vulnerability is greatly improved.
A 2023 report produced by Dr. Natalia Zamora of the Barcelona Supercomputing Centre, in collaboration with the WTW Research Network, found that the main sources of uncertainty for tsunami research are the limitations or availability of datasets such as topography or bathymetry to simulate tsunamis, and a lack of information on the built environment that can add caveats for hazard and risk assessment. Tsunami hazard maps based on one or few seismic events are commonly used for evacuation purposes, assuming a worst-case scenario, and inundation maps using historical data or geological interpretation as sources for the maximum credible hazard may not account for uncertainties in these given scenarios.
In addition to research programs aiming to better our understanding of tsunami triggering events and modelling of scenarios based on earthquakes and topography and bathymetry, physical modelling of tsunami wave effects is also an important area of study. ‘MAKEWAVES’, an international multi-partner collaborative project bringing together academic institutions such as WRN partners University College London as well as commercial consultancies, models wave generation and runup using a pneumatic tsunami simulator to measure wave impact on buildings such as hydrostatic/hydrodynamic loads, air entrainment, debris modelling and scour of sediments. This research has fed into new building regulations such as in Indonesia, which adopted US code ASCE 7–16 for tsunami loads,[14] and improved construction practices around the Pacific, allowing waves to flow through and around structures while avoiding collapse and improving resilience against future tsunami events.
Owing to the sheer scale of the challenge of defending against tsunami, it is often unaffordable for all but the wealthiest countries to build effective defences. In the aftermath of the 2004 event, many reports emerged of the effectiveness of mangroves and other ecosystems’ ability to dissipate waves and prevent items and people from being washed back out to sea. Villages that retained their mangroves had far fewer casualties than those where they had been removed for industries like shrimp farming. Anecdotal evidence has been supported by subsequent modelling, and is being used to refine approaches and roll out nature-based solutions[15] in many locations around the world. Where it is difficult to defend against tsunami hazard, it is all the more important that consideration is given to new development in vulnerable areas and whether it is appropriate. Unfortunately, it has been demonstrated that more properties are being built on areas vulnerable to coastal flooding and tsunami around the world at an increasing rate.[16],[17]
Another approach used in strengthening the resilience of tsunami-prone regions, like Indonesia and the Pacific, to the impacts of disasters is through the use of disaster risk financing. By using risk assessment tools and catastrophe models to estimate damage and calculate losses for an event, loss thresholds can be agreed that trigger insurance payouts when exceeded. This solution has been implemented by initiatives such as the Pacific Catastrophe Risk Insurance Company (PCRIC)[18] and the Southeast Asia Disaster Risk Insurance Facility (SEADRIF)[19], to provide resilience for countries with low insurance penetration.
Despite ongoing strands of research in modelling, and increasing global preparedness measures, the consequences of tsunami can be dire and the potential for economic and human losses is still significant. Alongside the advances made in science, engineering and insurance solutions, populations and exposure to tsunami hazard has also grown significantly since 2004. Yet despite these risks, tsunami still does not feature within the scenarios run by many insurers outside of the US, Japan and New Zealand. This is likely because insurance penetration in some tsunami-prone regions like Indonesia are much lower, meaning less demand for catastrophe models. As a result, there is a significant insurance protection gap in areas like South East Asia, which alongside continuing technological improvement is an area that still needs to be addressed.
When tsunami-triggering events occur close to coastlines, there can be as little as 20 minutes warning before the waves reach shore, leaving limited time for evacuation and response. Therefore, greater understanding of tsunami triggering mechanisms is crucial to focus preparedness, supported by effective warning systems, engineering resilience measures and community preparedness. As new technologies and monitoring systems emerge, such as using existing submarine fibre optic cable networks to detect seismicity[20], they should be used to increase the catalogue of hazard scenarios that inform tsunami models.
Progress in warning systems, modelling, data and disaster risk financing will not prevent future tsunamis, but have increased preparedness and offer a better chance for affected populations to recover. Remembering rare but traumatic events such as the 2004 Boxing Day earthquake is important to support those advances, especially given estimates that by 2030, 50% of the world’s population will live in coastal areas exposed to tsunamis, sea level rise, storm surges and tropical storms.
