Abstract
Online Material: Figures of array correlation detection with an alternative embedded event, correlation score as a function of time‐bandwidth product, and correlation score using alternative time periods. Seismic waveform correlation offers the prospect of greatly reducing event detection thresholds when compared with more conventional processing methods. Correlation is applicable for seismic events that in some sense repeat, that is they have very similar waveforms (e.g., Gibbons and Ringdal, 2006). A number of recent studies have shown that correlated seismic signals may form a significant fraction of seismicity at regional distances (e.g., Schaff and Richards, 2011; Dodge and Walter, 2015; Slinkard et al. , 2015). For the particular case of multiple nuclear explosions at the same test site, regional distance correlation also allows very precise relative location measurements (e.g., Waldhauser et al. , 2004; Wen and Long, 2010) and could offer the potential to lower thresholds when multiple events exist (e.g., National Research Council, 2012). Using the Comprehensive Nuclear‐Test‐Ban Treaty (CTBT) International Monitoring System (IMS) seismic array at Matsushiro, Japan (MJAR), Gibbons and Ringdal (2012) were able to create a multichannel correlation detector with a very low false alarm rate and a threshold below magnitude 3.0. They did this using the 2006 or 2009 Democratic People’s Republic of Korea (DPRK) nuclear explosion as a template to search through a data stream from the same station to find a match via waveform correlation. In this article, we extend the work of Gibbons and Ringdal (2012) and measure the correlation detection threshold at several other IMS arrays. We use this to address three main points. First, we show the IMS array station at Mina, Nevada (NVAR), which is closest to the Nevada National Security Site (NNSS), is able to detect a chemical explosion that is well under 1 ton with the right template. …
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have