Earthquake Early Warning: Tools for System Assessment

Successful earthquake early warning requires a fast, reliable network of seismic stations and a fast, accurate analysis algorithm to produce useful warnings in a timely manner. Developers and operators of earthquake early warning systems must constantly assess the health of their network and the performance of their algorithms. Hardware and telemetry of the seismic network are evaluated based on station latencies and the numbers of station P-wave triggers. Software and event processing performance can be analyzed by considering the number of successful event detections, of false alarms (alerts generated for nonexistent events), and of missed events (real events for which no alert was generated). Additional parameters for evaluation are the speed of delivery of the alerts and accuracy of magnitude, location and ground shaking estimates. For rapid assessment of system performance, it is useful to provide automatically prepared summaries for review. We present new tools for viewing and assessing performance of the earthquake early warning system from data collection through telemetry to alert delivery in a graphical, intuitive format. The data is collected into a KMZ file for easy viewing in Google Earth. We will show some example of these tools as part of our evaluation of the CISN ShakeAlert system. We will continue to improve automatic procedures for CISN ShakeAlert system.