Timeliness of Data Delivery from Q330 Systems

Abstract

Research in rapid analysis of earthquake data, broadly classified as early warning (EW) or early alarm studies, has spurred interest in the capability of digital seismic data systems to provide low‐latency data. Because of the wide usage of Quanterra Q330 data systems in operational networks, published results analyzing the latency of deployed systems have considered these devices. Several years ago, studies (e.g., Allen, 2009; Brown et al. , 2011) showed that the effectiveness of EW systems was then limited by the total delay required to deliver waveform data for analysis of 4 s or more within instrumentation systems. Recent results (Kuyuk et al. , 2014), however, following the American Recovery and Reinvestment Act (ARRA)‐funded upgrade of instrumentation in California show that instrumentation‐related delays have been reduced in some operational networks to a comparatively small component, typically <10%, of the overall time required to issue EW alerts. We present a simple, robust method to measure the latency of data delivery arising in the Q330 family of data loggers, including the typical contribution to total latency of both hardwired and cellular wireless communication paths over the public internet. These latter two categories of communication represent the majority of those used in operational seismological networks. Current EW systems are capable of issuance of an alert within 8.68±3.73 s after the first P ‐wave detection for events in California (Kuyuk et al. , 2014). We show that the actual latency introduced by Q330 data loggers—including public internet latencies—is typically <1.5 s maximum, with a mean value <1 s, when communications network bandwidth is capable of sustaining continuous data delivery. The method is applicable to any data system. Measured values of data latency are consistent with in situ measurements in EW systems now operating in California. The sub‐1 s mean latency capability of Q330 systems now deployed widely in seismically active areas …