Highlights

PRESTo Earthquake Early Warning System: Feasibility Study For A Nation-Wide Deployment Using An Integrated Regional and On-site Approach

AGU Fall Meeting, San Francisco, USA, 3 - 7 December, 2012
Aldo Zollo, Simona Colombelli, Luca Elia, Antonio Emolo, Gaetano Festa, Claudio Martino, Sandro Marcucci

Abstract

PRESTo (PRobabilistic and Evolutionary early warning SysTem) is a highly configurable and easily portable platform for Earthquake Early Warning (EEW) that integrates algorithms for real-time probabilistic earthquake location, magnitude estimation and damage assessment. The system processes the live accelerometric streams from the stations of a seismic network and promptly performs event detection providing location and magnitude estimations, as well as shaking prediction at the regional scale. The earthquake location and magnitude are obtained by an evolutionary, probabilistic approach that uses information from both triggered and not-yet-triggered stations at each time step. Peak ground motion is estimated at target sites by GMPEs using location and magnitude.
The above regional approach is integrated with a threshold-based early warning method for the definition of alert levels at stations and the estimation of the Potential Damaged Zone (PDZ), in which the highest intensity levels are expected. At each station the characteristic P-waves period (τc) and the peak displacement (Pd) are simultaneously measured on the initial P-waves signal. They are compared with threshold values, previously established through an empirical regression analysis, to produce an alert level at each station that can be correlated to the expected local damage in a robust way. Integrating the measured on-site parameters at stations (Pd, τc) and the estimated regional parameters (hypocenter), PRESTo can identify the damage area in few seconds with a good matching with the instrumental intensity maps produced much later. Alarm messages containing the evolutionary estimates and uncertainties of the above parameters can reach target sites before the destructive waves, enabling automatic safety procedures. Since late 2009, PRESTo has been under continuous real-time experimentation in Southern Italy, on the data streams of the Irpinia Seismic Network (ISNet), producing a bulletin of more than 200 events (M < 3.7). Since June, 2012, s similar deployment is implemented at KIGAM accelerometer network, in South Korea. Likewise, PRESTo has been off-line tested on the Mw >= 4 deep earthquakes recorded since 2005 by the Romanian Seismic Network and earthquakes off-shore the Iberian Peninsula, using recordings from the Spanish and Portuguese networks. An application to the KOERI seismic network in Turkey is also being studied.
This work shows the feasibility of a nation-wide Early Warning System, by analyzing the performance of PRESTo EWS on scenarios involving the Italian Accelerometric Network (RAN). Moderate to large magnitude events in areas with differing network geometries and densities are simulated by playing back in the system either synthetic traces, for historical earthquakes, or actual recordings for more recent events. Using these simulations we measured the performance of the integrated regional and on-site Early Warning system, in terms of both available lead-times and capability to predict the peak ground motion. When a dense seismic network featuring real-time telecommunication is deployed in the fault area, PRESTo produces reliable estimates of the source parameters and the earthquake damaging effects within 5-6 s from the event origin time.
http://isnet.na.infn.it/presto_index.php