Real-time characterization of large earthquakes using the predominant period derived from 1Hz GPS data

Earthquake early warning (EEW) systems’ performance is driven by the trade-off between the need for a rapid alert and the accuracy of each solution. A challenge for many EEW systems has been the magnitude saturation for large events (Mw>7) and the resulting underestimation of seismic moment magnitude. In this study, we test the performance of high-rate (1Hz) GPS, based on seven seismic events, to evaluate whether long-period ground motions can be measured well enough to infer reliably earthquake predominant periods. We show that high-rate GPS data allow the computation of a GPS-based predominant period (τg) to estimate lower bounds for the magnitude of earthquakes and distinguish between large (MW>7) and great (MW>8) events and thus extend the capability of EEW systems for larger events. It is also identified the impact of the different value of the smoothing factor α on the τg results and how the sampling rate and the computation process differentiates τg from the commonly used τp