Predicting earthquakes

Earthquakes can be charted by analysing infrasound high in the atmosphere, or so Delft researchers found. Their discovery could save lives.

On 12 January 2010, a magnitude seven earthquake devastated Haiti. It took days before the extent of the damage was understood, partly because there were no seismometers in the region at the time of the earthquake.

Delft researchers have found a way to circumvent this problem if future disasters like this occur. They are able to pick up acoustic signals – also called infrasound – in the atmosphere caused by earthquakes.

“Some of the energy caused by earthquakes leaks into the atmosphere,” says seismic expert Dr Shahar Shani-Kadmiel of the Geoscience and Engineering Department (CEG faculty). “Even if there are no seismic stations anywhere near the epicentre, you can still pick up the signals in the atmosphere thousands of kilometres away and use these to reconstruct a shakemap, which indicates the distribution of shaking intensity.”

The mathematical model that explains how this energy leaks into the atmosphere has been known for over a century. But until now it was impossible to filter out the faint waves from the background noise in the atmosphere with pressure sensors (barometers).

“Now we can do this because we have had plenty of data to improve our algorithms and calibrate the models,” says Shani-Kadmiel. “Many of these data come from submarine volcanic eruptions. Another important source was the underground nuclear test that North Korea carried out on 3 September 2017 at Punggye-ri, in the north east of the country. This was a much bigger explosion than the ones in the preceding years. It was the equivalent of a magnitude six earthquake.”

The research was carried out in the framework of the Comprehensive Nuclear-Test-Ban Treaty. The leader of the project, Prof. Läslo Evers (CEG Faculty), is also employed by the Royal Netherlands Meteorological Institute KNMI which surveys the international ban on nuclear tests.

The Delft researchers used the new insights on the wave propagation to ameliorate their shakemap algorithms for earthquakes. An article about how they reconstructed the shakemap of the 2010 Haiti disaster has been submitted to Nature as a proof-of-concept.