The Daily Manila Shimbun

 

Big, small quakes start similarly, making accurate early warning difficult: study

September 6, 2019



Tokyo--A Japanese earth and planetary scientist has found that large and small earthquakes cannot be distinguished at their onsets in many cases, revealing a fundamental difficulty the current early earthquake warning system in Japan has in promptly forecasting a huge tremor from its initial shake.

In the warning system, the Japan Meteorological Agency issues a big earthquake alert by quickly locating it and estimating its magnitude and seismic intensity based on initial waveforms recorded by seismographs.

The agency admits that the system has limitations in making accurate forecasts due to a lack of sufficient time for analyzing the initial signals and other factors. Sometimes, an alert is made after a severe jolt is felt, or no big shake occurs despite an alert.

Satoshi Ide, professor at the University of Tokyo, suggested that efforts to shorten the interval between the detection of the first signals and alert issuance may not produce results as long as analysis of the initial waveforms is given priority.

In his latest study, published online by the British scientific journal Nature, Ide compared relatively large earthquakes with a magnitude of at least 4.5 observed in the Tohoku-Hokkaido subduction zone plate boundary in eastern Japan from June 2002 to April 2018 with smaller ones that occurred in nearby locations. Half of the sampled big earthquakes were subduction-type events, similar to the one that devastated the Tohoku region in March 2011.

As a result, initial waveforms were found to be extremely similar for nearly 20 pct of pairs of big and small subduction-type earthquakes, far higher than around 4 pct for different-type quakes, he said.

While pointing out that the high similarity would make it difficult to estimate the final size of an earthquake from the initial data, Ide said the study suggests that every gigantic event starts with a tiny rock rupture at almost a point and takes a certain development process determined by cascade structures in a plate boundary.

"The probability of cascading should be estimated theoretically or empirically and can then be employed to assess the predictability of future earthquakes with a more quantitative size estimation," he said. Jiji Press