North American seismic networks can contribute to nuclear security

In a paper published as part of an upcoming focus section on regional seismic networks in Seismological Research Letters, University of Utah seismologist Keith Koper explains how the work of regional seismic networks in North America is contributing to nuclear test monitoring, particularly in the case of low-yield explosions. Koper is the director of the University of Utah Seismograph Stations.

The need to detect low-yield explosions -- those with a yield less than 0.5 kilotons and causing seismic events between magnitude 1.0 and 3.0 -- has become more pressing, given conflicting reports on Russian low-yield tests at its Arctic island base of Novaya Zemlya and in North Korea, where researchers have debated whether a 2010 event was an explosion or a natural earthquake.

"Whether with an experienced tester like Russia or an inexperienced tester like North Korea, the U.S. government has a very strong interest in these small events," said Koper.

These small seismic events would likely only be well-recorded at local distances of 150 to 200 kilometers, and are difficult to distinguish from earthquakes and other kinds of industrial explosions. Regional seismic networks, which operate over local distances and routinely create catalogs of small earthquakes that must be uncontaminated by non-earthquake seismic noise, are well-equipped to handle these challenges.

"That's sort of our bread and butter -- the detection, the location and estimation of the size of these small events," Koper noted. "It's a way that regional networks in North America can contribute to this important global security issue."

In the SRL paper, Koper offers several examples of how regional seismic networks have provided data that can be useful to international nuclear monitoring. Most regional seismic networks in North America center around a particular "seismo-tectonic feature that warrants extra monitoring versus what you might get from a global or national-scale network," said Koper, such as the New Madrid Fault Zone in the central United States or the Cascadia subduction zone in the Pacific Northwest.

Data from several North American regional networks are helping seismologists refine some of the usual techniques that they use to distinguish explosions from natural earthquakes, such as comparing surface and body seismic waves and calculating the ratio between P-waves and S-waves. (P-waves compress rock in the same direction as the wave's movement, while S-waves move rock perpendicular to the direction of the wave.) While these techniques can identify earthquakes at a regional and global distance, data collected by regional networks suggests that these techniques may not perform in a similar way over local distances.

Specific regions may provide other important information for monitoring. For instance, the geology in the northeastern United States is more similar to Asian test sites than the geology of western North America. "If you're really interested in detecting things in, let's say, North Korea, the Northeastern U.S is a better geological analog," Koper said.

Beyond the IMS, low-yield nuclear test discrimination is also a topic of interest for agencies like the U.S. Department of Defense and national laboratories. In Utah, for instance, the Air Force Research Laboratory has provided funds for the regional network operators to create a catalog of times, locations and sizes of Utah mining blasts. "For the people who are testing new methods of discrimination, maybe based on machine learning or something like that, we can provide this nice, ground-truth catalog," Koper said.