Being able to distinguish foreshocks and aftershocks of earthquakes in real time could be useful for earthquake prediction (see L. Gulia and S. Wiemer Nature 574, 193–199; 2019). For example, the authors claim that — in retrospect — their method could have predicted the biggest such event in the 2016–17 cluster of earthquakes that occurred in the Apennines in central Italy: the magnitude-6.6 earthquake that hit the town of Norcia in October 2016. There were no casualties, yet the death toll from a similar event in the region — the Avezzano earthquake of magnitude 6.7 in January 1915 — was 30,000. How could this difference be explained? It could be because Italy’s Major Risk Committee, of which we were members at the time, found that a large event had a higher probability of occurring than usual, based on the persistence of the earthquake sequence in the region, and recommended putting the entire area under official alert. The committee issued a warning 40 hours ahead of the earthquake to the public, the press and the Civil Protection organization (see go.nature.com/2ecmvwk). As a result, prefects and mayors enforced mass evacuation.
Language: en
LA - en SN - 0028-0836 UR - http://dx.doi.org/10.1038/d41586-019-03941-2 ID - ref1 ER -