A Review of Broadband Regional Discrimination Studies of NTS Explosions and Western U.S. Earthquakes

Abstract

Verification of a Comprehensive Test Ban Treaty (CTBT) will require the use of seismic stations within regional distances (<2000 km). One important aspect of regional seismic monitoring involves the discrimination of nuclear explosions from other sources such as earthquakes and mining events (e.g. industrial explosions and rockbursts). In this paper, we review earthquake/nuclear explosion discrimination studies in the western U.S. using broad-band seismic data. These studies are important because they are the only ones involving nuclear explosions and other sources in a single geophysical region having excellent ground truth information and a substantial historic database. Additionally, because of access to information from the NTS, much can be learned about the physical basis of regional discriminants. Using multivariate discrimination techniques, it was found that approximately 96% of the events analyzed could be correctly identified down to about magnitude 3.5. Most of the events misidentified were recorded with poor signal-to-noise ratio at only a minimal number of stations. However, a few well-recorded events were misclassified for one or more discriminants. Examples of detailed analysis of a missed violation (nuclear explosion that looks like an earthquake) and a false alarm (naturally occurring event that looks like an explosion) are illustrated. Resolution of anomalous events such as these will be critical to CTBT monitoring.