Rapid Estimation of Macroseismic Intensity for On-site Earthquake Early Warning in Italy from Early Radiated Energ

Abstract:

Earthquake Early Warning System (EEWS) are effective tools for the risk mitigation in active seismic regions. Recently, a feasibility study of a nation-wide earthquake early warning systems has been conducted for Italy considering the RAN Network and the EEW software platform PRESTo. This work showed that a reliable estimations in terms of magnitude and epicentral localization would be available within 3-4 seconds after the first P-wave arrival. On the other hand, given the RAN’s density, a regional EEWS approach would result in a Blind Zone (BZ) of 25-30 km in average. Such BZ dimension would provide lead-times greater than zero only for events having magnitude larger than 6.5. Considering that in Italy also smaller events are capable of generating great losses both in human and economic terms, as dramatically experienced during the recent 2009 L’Aquila (M_L 5.9) and 2012 Emilia (M_L 5.9) earthquakes, it has become urgent to develop and test on-site approaches.

The present study is focused on the development of a new on-site EEW metodology for the estimation of the macroseismic intensity at a target site or area. In this analysis we have used a few thousands of accelerometric traces recorded by RAN related to the largest earthquakes (M_L>4) occurred in Italy in the period 1997-2013.

The work is focused on the integral EW parameter Squared Velocity Integral (IV2) and on its capability to predict the peak ground velocity PGV and the Housner Intensity I_H, as well as from these latters we parameterized a new relation between IV2 and the Macroseismic Intensity. To assess the performance of the developed on-site EEW relation, we used data of the largest events occurred in Italy in the last 6 years recorded by the Osservatorio Sismico delle Strutture, as well as on the recordings of the moderate earthquake reported by INGV Strong Motion Data. The results shows that the macroseismic intensity values predicted by IV2 and the one estimated by PGV and I_H are in good agreement.