Copula-based earthquake early warning decision-making strategy

This paper presents a new copula-based earthquake early warning (EEW) decision-making strategy, aiming to characterize missed-alarm and false-alarm probabilities for an on-site EEW and determine an optimum threshold at which EEW should be set off with the lowest missed-alarm and false-alarm probabilities combined. On the basis of an existing PD3-PGV (PD3: peak ground displacement within the first three seconds after P-wave arrives; PGV: peak ground velocity) on-site EEW, the analysis shows that a copula model consisting of the Lognormal distribution, Weibull distribution, and Frank copula can satisfactorily model the PD3-PGV joint probability distribution. Accordingly, the optimum PD3 triggering thresholds for different PGV warning thresholds from 5 to 35 cm/s are presented for future references in the use of the PD3-PGV on-site EEW with maximum reliability.