Abstract. Here we propose a methodology for Earthquake Early Warning able issuing the alert based on the real-time estimation of the epicentral area where a ground Intensity measure is expected to exceed a user-set ground shaking level. The method provides in output a P-wave-based, time-evolutive “early” shake map. The P-wave displacement, velocity and acceleration amplitudes are jointly measured on a progressively expanded time window while the earthquake location and magnitude are evaluated using data at near source stations. A retrospective analysis of the 2016, Mw 6.5 Central Italy earthquake records shows that the method naturally accounts for effects related to the earthquake rupture directivity and spatial variability of strong ground motion related to source and path and site effects. Five seconds after the origin time the simulated performance of the system in predicting the event impact is very high: in the 40 km-radius area that suffered an Intensity MCS VIII-IX, 41 over 42 strong-motion instrumented sites would have been successfully alerted, with only one false alert. Even considering the 15-km-radius blind-zone, a 15-55 km wide annular area would have received the alert 2-14.5 sec before the occurrence of the strong ground shaking.The proposed EEW method evolves with time in a way that it minimizes the missed alarms while increasing successful alarms and to a lesser extent false alarms, so it is necessary for the end-user to accept these eventualities and account for them in a probabilistic decision scheme depending on the specific safety actuation measure to be undertaken in real-time

In September 2016 and again in November 2017, Korea witnessed some of the largest magnitude earthquakes occurred in-land in its recent history. They were located in the South-Eastern part of the peninsula, near the cities of Gyeongju and Pohang, with magnitude ML 5.8 and 5.4 respectively. The regional PRESTo Early Warning System has been in the testing phase at KIGAM (Korea Institute of Geoscience and Mineral Resources) since late 2013, using the nation-wide KIGAM and KMA (Korea Meteorological Agency) real-time accelerometric streams, after a calibration based on moderate-to-low seismicity. In this work the PRESTo algorithm parameters and regression laws were optimized and calibrated for the region of South Korea around the Gyeongju earthquake, specifically for a sub-network with a side of about 200 km. A waveform database spanning ten years was analyzed (99 earthquakes during 2007-2017, with M 2 ÷ 5.8, 15681 waveforms), to derive an optimized configuration for this network geometry and seismicity, including GMPEs and empirical regression laws for magnitude estimation from peak displacement, measured in short P-waves and early S-waves time-windows. This configuration enables a fast characterization of the local earthquake sources and effects, i.e. during the playback of onshore earthquakes with magnitude 3.0 and above the EW system produces, in 90% of the cases: a first alert within 6 seconds from the first pick; stable magnitude estimations (variations within 0.5 magnitude points); a final estimate performance against the reference bulletin of ∆M ≤ 0.5, ∆Repi ≤ 10 km.

A primary task of a network-based, earthquake early warning system is the prompt event detection and location, needed to assess the magnitude of the event and its potential damage through the predicted peak ground shaking amplitude using empirical attenuation relationships. Most of real-time, automatic earthquake location methods ground on the progressive measurement of the first P-wave arrival time at stations located at increasing distances from the source but recent approaches showed the feasibility to improve the accuracy and rapidity of the earthquake location by using the additional information carried by the P-wave polarization or amplitude, especially unfavorable seismic network lay-outs. Here we propose an evolutionary, Bayesian method for the real-time earthquake location which combines the information derived from the differential P-wave arrival times, amplitude ratios and back-azimuths measured at a minimum of two stations. As more distant stations record the P-wave the posterior pdf is updated and new earthquake location parameters are determined along with their uncertainty. To validate the location method we performed a retrospective analysis of mainshocks (M>4.5) occurred during the 2016-2017 Central Italy earthquake sequence by simulating the typical acquisition layouts of in-land, coastal and linear array of stations. Results show that with the combined use of the three parameters, 2-4 sec after the first P-wave detection, the method converges to stable and accurate determinations of epicentral coordinates and depth even with a non-optimal coverage of stations. The proposed methodology can be generalized and adapted to the off-line analysis of seismic records collected by standard local networks.