An Examination of the Threshold‐Based Earthquake Early Warning Approach Using a Low‐Cost Seismic Network

Earthquake early warning (EEW) systems have been researched and developed in Japan, Mexico, the United States, Taiwan, and other countries (e.g., Allen et al. , 2009; Satriano et al. , 2011). One important aspect of EEW is to rapidly assess potential earthquake damage using the early portion of ongoing ground vibration. There are two types of EEW systems, namely, regional and onsite. Regional EEW systems collect and analyze seismic data from multiple stations near the epicenter and provide earthquake information (e.g., magnitude and location) to distant sites. In contrast, onsite systems use the early part of the P wave to predict impending ground motion for the later S and surface waves at the same site without necessarily estimating the source location and magnitude. Thus, onsite systems are generally faster than regional systems for near-source sites. Over the past decade, the initial peak ground displacement ( P d) and predominant period ( τ c) for the initial P wave have been the two most important early warning parameters used to rapidly estimate the magnitude (e.g., Wu and Kanamori, 2005a; Shieh et al. , 2008). Furthermore, several authors (Wu and Kanamori, 2005b; Hsiao et al. , 2009) have found empirical relations between the P d and peak ground velocity (PGV) and peak ground acceleration (PGA). A threshold-based EEW approach was recently tested in several studies. For the threshold-based method, once the initial P -wave amplitude exceeds a certain threshold, an alarm is issued. Therefore, the EEW processing time is reduced for a faster warning. Wu and Kanamori (2005b) used 26 earthquake events ( M w>5.0 and focal depth 0.5  cm is a good …