Simple model could predict rare 'tsunami earthquakes'

[SANTIAGO] Tsunami alert systems may soon benefit from a simple computer model that predicts more accurately those earthquakes that can potentially cause large, destructive, tsunamis.

These rare 'tsunami earthquakes' rupture more slowly and cause much larger tsunami waves, such as the 17-metre wave that followed last year's (25 October) magnitude 7.8 event in Sumatra, Indonesia, than more common 'subduction earthquakes' of the same magnitude, such as the magnitude 9 event off Japan earlier this month  (11 March) which caused waves of up to ten metres.

Existing warning systems may miss tsunamis caused by such earthquakes or predict much smaller waves, according to Andrew Newman, a scientist at the Georgia Institute of Technology, United States, whose team has devised a new system, RTerg (Realtime erg), described in Geophysical Research Letters earlier this month (5 March).

"Slow-rupturing tsunami earthquakes are difficult to observe and can be misrepresented by current systems … they create a tsunami much larger than expected for the magnitude of the earthquake."

RTerg uses data received from around 150 seismic stations on the energy radiated by an earthquake and the rate of rupture to calculate how long it takes for the earthquake's energy to build up.

The ratio of the two is ideal for identifying tsunami earthquakes because a deficiency in radiating energy and an extended rupture period are hallmarks of these events, said Newman.

Using live data, RTerg identified the Sumatra earthquake as a potential tsunami earthquake in eight and a half minutes. The team is currently working on further reducing the warning time.

Newman said that any earthquake processing or tsunami warning centre that receives global seismic information in real time should be able easily to adopt the methodology.

"Since the code was written by earth scientists, and not computer programmers, we need to rewrite the codes this year to provide a more standardised version for tsunami warning centres," Newman said.

Jörn Lautaerjung, head of the Indian Ocean Tsunami Early Warning System, told SciDev.Net: "This tool could improve the prediction of tsunamis that take more than ten minutes to reach the shore. If it is faster than the systems used by tsunami early-warning centres, these may benefit from integrating it into their seismic software."

Bernardo Aliaga, acting head of the Tsunami Unit at UNESCO's Oceanographic Commission, said: "The RTerg system advances our knowledge of how slow earthquakes capable of generating a tsunami behave. It would be ideal to test it under real conditions and in a warning centre."

Link to abstract in Geophysical Research Letters