The Papua New Guinea Tsunami, 20 years on ... by Prof Dave Tappin

20 years ago today, on the evening of the 17th July 1998, 2200 people died when a 15-metre high tsunami devastated an idyllic lagoon on the north coast of Papua New Guinea (PNG).  The event was to prove a benchmark in tsunami science as the tsunami was generated, not by an earthquake, but by a submarine landslide. Most tsunamis are generated by earthquakes and, previously, submarine landslides were an under-appreciated mechanism in tsunami generation. This was because there had been no recent historical event to prove just how dangerous they could be.

The Sissano Lagoon devastated after the 1998 tsunami  (Image courtesy of Jose Borrero, University of Southern California)
The Sissano Lagoon devastated after the 1998 tsunami
(Image courtesy of Jose Borrero, University of Southern California)
Sissano Mission school carried 65 metres inland by the tsunami wave
(credit; NOAA/NGDC, Hugh Davies, University of PNG)

At a water depth of 1600 metres, on the landslide headscarp
we found slumped limestone blocks, together with cold water
chemosynthetic mussels and tubeworms feeding on the methane
rich fluids expelled when the seabed failed. 
Credit: Japan Agency for Marine Earth Science
Back in 1998, there had been few recent destructive earthquakes, they were to strike later.  Although earthquake mechanisms were generally well understood in tsunami generation, the mechanisms by which submarine landslides cause tsunamis, were not. In fact it was generally believed that submarine landslides could not generate destructive tsunamis.

PNG, changed all this.

The importance of Papua New Guinea

PNG was a ‘wake-up call’ for tsunami hazard. The tsunami was the most devastating event since Sanriku 1933, when a tsunami struck the east coast of Japan, leaving 1500 dead and the same number missing.

The massive death toll, generated a surge of scientific interest in non-earthquake tsunami mechanisms, which subsequently extended outside of convergent margins, where earthquakes are most common, to passive margins, and to volcanic collapse.

The tsunami struck at a time when new technology was being used increasingly to map the sea bed as well as topography was being mapped on land.  New numerical models of submarine landslide tsunamis were also being developed, but were still theoretical, and PNG allowed these to be tested in real life conditions.

At the time of PNG, tsunami science was dominated by seismologists because earthquakes were seen as the only major hazard. Research into submarine landslide tsunamis requires the contribution from geologists, so geologists became much more involved. The research on the PNG tsunami was therefore to prove seminal.

Papua New Guinea – the forerunner

As was later to prove, PNG was the first of a series of catastrophic tsunamis which over the next 13 years were to devastate the coastlines of the Indian Ocean (2004) and Japan (2011). These tsunamis killed over 250 000 people and caused billions of pounds worth of damage. These events would ‘rock’ the globe, bringing home to world populations the previously unrecognised hazard from these events.

Aerial view of Banda Aceh, northern Sumatra, where over 100,000 people died in the 2004 Indian Ocean tsunami
(Image U.S. Navy photo by Photographer's Mate 2nd Class Philip A. McDaniel)

In the case of the Indian Ocean, there was a realisation that geological hazards, such as volcanic eruptions, earthquakes and tsunamis do not just impact on ‘other people’ in far distance places. With ever increasing international travel made so much easier by a general drop in prices, an idyllic holiday in an exotic location could quite rapidly turn into a nightmare.

Earthquake tsunamis are not the only hazard

Sanriku, 1933 was an earthquake-generated tsunami resulting from a Mw 8.4 event. The scale of the tsunami from Sanriku earthquake, although devastating, was commensurate with the earthquake magnitude. Both Japan and PNG are sited along plate boundaries, termed convergent margins, where earthquakes are quite common. More recent events along these types of margins were in 2011 off the east coast of Japan and in 2004 in the Indian Ocean. The tsunamis from these events were also devastating, but in scale with their associated earthquakes.

The elevation of the tsunami which struck PNG, however, was completely out of proportion to the associated earthquake Mw of 7.1. Most earthquakes are caused by movement, or ‘slip’, along the interface between the plates which are colliding along convergent margins. Although there are ‘special’ types of earthquakes, termed ‘tsunami earthquakes’, which may generate tsunamis larger than their magnitude would suggest. Tsunami earthquakes are usually associated with heavily sedimented convergent margins, and the Papua New Guinea margin is not of this type.

There were several other aspects of the PNG tsunami which suggested that the earthquake was not the cause. There was a 20 minute delay between the earthquake and the tsunami striking the coast. The earthquake was located quite close to shore, so this was immediately anomalous. Field surveys conducted immediately after the event also found that the distribution of the tsunami elevations along the coast had the highest wave heights focused on the low-lying Sissano Lagoon.

Scientists confused and the generosity of others 

After the PNG event and as the results of the first field surveys were circulated, there was much discussion in science circles on why the tsunami was so elevated in relation to the earthquake magnitude. For example, at the AGU international scientific meeting in San Francisco in December of 1998 there was a special session during which the PNG tsunami was discussed.

Without further research the event would have remained an enigma. Except that, in response to a plea for help from Alf Simpson, the Director of the regional geoscientific organisation, SOPAC, which assisted PNG in mitigating their geological hazards, the Government of Japan funded four marine scientific research expeditions on state of the art vessels, to survey the area offshore of the devastated area. The USA diverted one of its vessels working in the region to acquire further marine data. This was the first time that marine surveys had been carried out in response to a major tsunami disaster, and the first time this region had been surveyed using these sophisticated technologies.

The JAMSTEC Kairei which in January 1999 was the first research vessel to research the PNG 1998 tsunami
(Credit Dave Tappin)

Marine surveys provide the answers

The first surveys took place in January 1999 and, from mapping the sea bed, discovered a submarine landslide just offshore of the area devastated by the tsunami.  Based on the mapping, the landslide discovered was used as the basis for numerical models of the tsunami. This was a major challenge as this had only been attempted once previously. The numerical models demonstrated that the landslide was the most likely cause of the tsunami.

The 1998 Papua New Guinea tsunami was triggered 12 minutes after the earthquake by a rotational slump,
~6km3 in volume, located 20 km offshore of the devastated area. Note the circular expansion
of the tsunami waves, characteristic of a point-source, submarine landslide tsunami. Credit: Phil Watts

Because landslides were considered not to cause hazardous tsunamis, this result on the tsunami mechanism was controversial, but gradually as other events were identified and more new numerical models were developed, they became more generally accepted.

The area offshore of northern Papua New Guinea mapped in 1999 by the Kairei (Credit. Tappin et al 2001).

The simulation is based on a dual, earthquake/submarine landslide mechanism,
with the landslide triggered three minutes after the earthquake.
Note the linear tsunami wave front from the earthquake in the south, and
the circular waves from the submarine landslide in the north. Credit: Stephan Grilli

Unforeseen downstream impacts

The generous investment made by Japan in funding the marine research on PNG was to be repaid in full in 2011, when the east coast of Honshu Island was devastated by a tsunami up to 40 metres in elevation which killed 18 000 people and cost 200 billion dollars in damage. Although the earthquake magnitude 9.0-9.1 could explain most of the tsunami, the elevated 40-metre-high run-ups along the northern Honshu could not. So, a submarine landslide was proposed and numerically modelled as the cause of these. Without the research carried out on the PNG tsunami, this would have been impossible.

Overview shot of Minamisanriku, northern Honshu, showing the destruction from the 2011 Japan tsunami (Credit Dave Tappin).
Overview shot of Minamisanriku, northern Honshu, showing the destruction from the 2011 Japan tsunami
(Credit Dave Tappin).

The destruction of Minamisanriku from the Japan 2011 tsunami (Credit Dave Tappin).

The future

Dave Tappin emerging after the first Shinkai
2000 submersible dive onto the tsunami 
landslide – November 1999 (Credit Horst Letz).
Since PNG, we have come a long way in understanding how submarine landslides generate tsunamis, but they are a major hazard which is still not fully understood or appreciated. Although mapping of the sea bed now demonstrates the almost universal presence of submarine landslides offshore of most coastal areas, there are still too few well studied events to form a sound basis for similar mitigation to that from earthquakes, which are addressed by warning systems in all the world’s ocean basins. In addition, the numerous different submarine landslide mechanisms means that ‘one size doesn’t fit all’ so the development of generalised models is still in its infancy.

As with all high impact – low frequency hazards, our experience from the recent tsunami events identified here is that memories fade fast after the immediate response. As memories fade, so does the investment needed to understand and mitigate the impacts of tsunamis in the future. Research into the submarine landslide hazard is ongoing, but is harder to fund as other research priorities take over. The next major challenge is to tackle dual earthquake/submarine landslide mechanisms, such as Japan, 2011, and to extend the ocean basin early warning systems, now operational for earthquakes, to include tsunamis from submarine landslides – because undoubtedly, at some time in the future there will be another event.

Note. Professor Dave Tappin of BGS participated in the research on the PNG tsunami, taking part in all of the four marine surveys funded by Japan. At first, only a one-off opportunity, it led to a career in tsunami science as later events in the Indian Ocean and Japan proved the massive hazard from tsunami events globally.  Dave acknowledges all of his numerous colleagues and friends with whom he has collaborated on this research.

Further reading

The Sissano Papua New Guinea tsunami of July 1998 - offshore evidence on the source mechanism 

The Papua New Guinea tsunami of 17 July 1998: anatomy of a catastrophic event

Submarine Mass Failures as tsunami sources - their climate control

Did a submarine landslide contribute to the 2011 Tohoku tsunami?

Tsunamis from submarine landslides

The Generation of Tsunamis

The importance of geologists and geology in tsunami science and tsunami hazard