The damaged Sunlive Kengun in the pedestrian arcade in downtown Kumamoto following earthquakes in 2016.
An international research team led by Western University has developed a new methodology for forecasting the magnitude of the largest possible earthquake within a series of seismic events like the deadly ones that shook Kumamoto, Japan in 2016.
The Kumamoto earthquakes, which featured a magnitude 7.3 mainshock less than two days after a magnitude 6.5 foreshock, killed at least 50 people and injured an estimated 3,000 others in total. The 2016 event was investigated by researchers from Western, the Institute of Statistical Mathematics (Japan) and the University of Potsdam (Germany) to create a new statistical approach that estimates the probabilities for such extreme earthquakes during a prolonged seismic sequence of events to be above certain magnitudes.
The findings were published today in the high impact journal Nature Communications.
Historical data shows most earthquakes, like Kumamoto, occur unexpectedly and often trigger subsequent events far more powerful than the initial shock. Forecasting the largest expected earthquake within a self-exciting seismic series is critically important in mitigating hazard, damage and loss of life.
Robert Shcherbakov, an associate professor in Western's Department of Earth Sciences and first author on the study, stresses that it is important to differentiate between forecasting and predicting earthquakes.
The renowned statistical seismologist explains: "Predicting earthquakes means providing a narrow range of times and locations where large earthquakes are going to occur which is rather unrealistic at the moment."
The model did however allow Shcherbakov and his collaborators to estimate, retrospectively, the probabilities of having large subsequent earthquakes during several stages of the evolution of 2016 Kumamoto seismic event.
"The probabilities for large earthquakes proved rather low but it is still very important to have such estimates," says Shcherbakov.
For the complete study, please visit https://doi.org/10.1038/s41467-019-11958-4
