$\varvec{\tau}_{ps}$, a new magnitude scaling parameter for earthquake early warning

Abstract

Rapid and reliable estimation of earthquake magnitude has a significant effect on proper performance of earthquake early warning system. In recent years, three methods including predominant period $(\tau_{p}^{max} )$, characteristic period $\left( {\tau_{c} } \right)$, and log-average period $(\tau_{\log } )$ have been introduced for estimation of earthquake size based on frequency contents within the first few seconds of P-waves. In the current study, new power-spectrum parameter, $\tau_{ps}$ in the modified form of $\tau_{\log }$ approach, is introduced by using waveforms of 70 earthquakes occurred in Japan with magnitudes ranging between M _JMA 3.0 and 8.0. Results demonstrate that $\tau_{ps}$ values have no significant weighted dependency on amplitude and frequency contents, and are better correlated with the reported magnitudes compared to other frequency-based proxies. It is also shown that $\tau_{ps}$ and $\tau_{p}^{max}$ magnitude-scaling relationships provide more precise and robust event size estimation than $\tau_{c}$ and $\tau_{\log }$ approaches, especially for small events. Thus, by using $\tau_{ps}$ parameter it has been expected that performance of earthquake early warning systems (EEWS) will improve by reducing the rate of false alarms.