We introduce in the on-site earthquake early warning (EEW) a partially non-ergodic perspective from the site effects point of view. We consider the on-site EEW approach where the peak ground velocity (PGV) for S waves is predicted from an early estimate, over the P waves, of either the peak-displacement (PD) or cumulative squared velocity (IV2). The empirical PD-PGV and IV2-PGV relationships are developed by applying a mixed-effect regression where the site-specific modifications of ground shaking are treated as random effects. We considered a large data set composed of almost 31 000 selected recordings in central Italy, a region struck by four earthquakes with magnitude between 6 and 6.5 since the 2009 L’Aquila earthquake. We split the data set into three subsets used for calibrating and validating the on-site EEW models, and for exemplifying their application to stations installed after the calibration phase. We show that the partially non-ergodic models improve the accuracy of the PGV predictions with respect to ergodic models derived for other regions of the world. Moreover, considering PD and accounting for site effects, we reduce the (apparent) aleatory variability of the logarithm of PGV from 0.31 to 0.36, typical values for ergodic on-site EEW models, to about 0.25. Interestingly, a lower variability of 0.15 is obtained by considering IV2 as proxy, which suggests further consideration of this parameter for the design of on-site EEW systems. Since being site-specific is an inherent characteristic of on-site EEW applications, the improved accuracy and precision of the PGV predicted for a target protection translate in a better customization of the alert protocols for automatic actions.

On-site earthquake early warning: a partially non-ergodic perspective from the site effects point of view

Spallarossa, D;Barani, S;
2019-01-01

Abstract

We introduce in the on-site earthquake early warning (EEW) a partially non-ergodic perspective from the site effects point of view. We consider the on-site EEW approach where the peak ground velocity (PGV) for S waves is predicted from an early estimate, over the P waves, of either the peak-displacement (PD) or cumulative squared velocity (IV2). The empirical PD-PGV and IV2-PGV relationships are developed by applying a mixed-effect regression where the site-specific modifications of ground shaking are treated as random effects. We considered a large data set composed of almost 31 000 selected recordings in central Italy, a region struck by four earthquakes with magnitude between 6 and 6.5 since the 2009 L’Aquila earthquake. We split the data set into three subsets used for calibrating and validating the on-site EEW models, and for exemplifying their application to stations installed after the calibration phase. We show that the partially non-ergodic models improve the accuracy of the PGV predictions with respect to ergodic models derived for other regions of the world. Moreover, considering PD and accounting for site effects, we reduce the (apparent) aleatory variability of the logarithm of PGV from 0.31 to 0.36, typical values for ergodic on-site EEW models, to about 0.25. Interestingly, a lower variability of 0.15 is obtained by considering IV2 as proxy, which suggests further consideration of this parameter for the design of on-site EEW systems. Since being site-specific is an inherent characteristic of on-site EEW applications, the improved accuracy and precision of the PGV predicted for a target protection translate in a better customization of the alert protocols for automatic actions.
File in questo prodotto:
File Dimensione Formato  
Spallarossa_et_al_2019.pdf

accesso chiuso

Tipologia: Documento in versione editoriale
Dimensione 7 MB
Formato Adobe PDF
7 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
11567 941533Spallarossa.pdf

accesso aperto

Tipologia: Documento in Post-print
Dimensione 7.69 MB
Formato Adobe PDF
7.69 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11567/941533
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 16
  • ???jsp.display-item.citation.isi??? 15
social impact