Global earthquake predictor models shown to work

CSEP network includes Italy, California, New Zealand and Japan

(ANSA) - Rome, June 12 - The first models to forecast earthquakes have been shown to work.
    Though the road is still long, the tests provide accurate probability margins, according to the Collaboratory for the Study of Earthquake Predictability (CSEP) network.
    Italy takes part in CSEP with researchers from the Istituto Nazionale di Geofisica e Vulcanologia (INGV), along with California, New Zealand and Japan. "We are not able to say whether there will be an earthquake tomorrow, but we are able to calculate the probability" of it happening, ANSA was told by INGV seismologist Warner Marzocchi.
    The first results of tests conducted by CSEP covered the 2009-2014 period and are published in the Seismological Research Letters journal. They also include tests based on data from the 2012 earthquake in Emilia Romagna and the article was signed by Marzocchi and Matteo Taroni.
    Marzocchi added that if the probability of an earthquake "is 5%, this means that on average one out of every 20 times there will be an earthquake in those specific conditions". This means, he added, that "we are still light years away for what is generally meant when we say 'forecast', but we are certainly able to say how the probability changes over time and space". For the past few years, data have been transmitted at an experimental level both to the Major Risks Commission (CGR) and the civil protection service. All the data from August 2009 until 2014 recorded by the national seismic network - including the earthquakes in Emilia Romagna in 2012 - were compared with the probability forecasts provided by many models tested by CSEP. Analysis will also be conducted on a series of earthquakes that began on August 24, 2016 in central Italy. "The comparison between the forecasts made and the data observed also make it possible to identify the most effective model,"Marzocchi added. "Thanks to the CSEP network, we have been able to move all discussions on forecasts to a scientific environment." After this initial round, another five years of testing on the models is expected.