Exploring the next generation of EEW testing: Large magnitude events, geodetic data and station latency information, with a focus on the Ridgecrest sequence

Earthquake early warning algorithms must undergo rigorous real-time and offline testing before being accepted into the ShakeAlert production system. The ShakeAlert Testing and Certification platform attempts to simulate how the production system will perform by deploying the proposed algorithms, configurations, earthworm rings/modules, and ActiveMQ messaging on separate, nearly identical hardware. For real-time testing, the test servers are fed the same West Coast input waveform data as the production servers, and the algorithms' performance is compared between the test and existing production systems after a minimum of two weeks. For offline testing, four simultaneous instances of the algorithms are run on one machine with a historic test suite of earthworm tankplayer files (format used for replaying data into an earthworm ring with realistic timing). These tankplayer files include data from West Coast earthquakes (M4.1 to M7.2), teleseisms, regional events, and problematic data sets (re-centering channels, calibrations, noisy data, etc.).

We have made significant progress in developing a testsuite v.2.0, for the ShakeAlert EEW testing platform. In particular, this testsuite v.2.0 will include tests of algorithm performance for many more large magnitude (M6+) events that were recorded by dense strong-motion networks. We are augmenting the large magnitude suite of earthquakes by showing results for numerous crustal and subduction zone Japanese events, the M7.1 Anchorage event, the M6.7 Northridge event, the M6.4 and M7.1 Ridgecrest events, and possibly other events. In this testsuite v.2.0, we will also will incorporate seismo-geodetic data (including data from the Geodetic Algorithms Testing and Implementation subcommittee) for testing geodetic algorithms in EEW. It is important to assess geodetic algorithm behavior for large magnitude events, such as the Ridgecrest sequence, in order to eventually incorporate geodetic data into real-time alerts.

Two other areas of development in testing include better ground motion assessments of algorithms, and exploring how to incorporate station latency information into the testsuite. These improvements will lead to a more accurate testing environment and yield better comparisons to the real-world performance of ShakeAlert.