Could seismic signals from earthquakes mask the signals of an underground explosion?

It's possible, according to a new review article published in the Bulletin of the Seismological Society of America that contradicts the conventional wisdom about explosion "masking."

The new analysis by Joshua Carmichael and colleagues at Los Alamos National Laboratory found that advanced signal detector technology that can identify a 1.7-ton buried explosion with a 97% success rate only has a 37% success rate when seismic signals from that explosion are hidden within the seismic waveforms of an earthquake that happens within 100 seconds and about 250 kilometers away from the explosion.

The overlapping waveforms of explosion and earthquake "obfuscate the ability of even the most sensitive digital signal detectors we have to identify that explosion," said Carmichael.

The findings could lead experts to reconsider a 2012 report that concluded earthquake signals could not cover up explosion signals. Potential explosion masking by natural seismic signals is a concern for the community of scientists charged with nuclear test monitoring around the globe.

In North Korea, which has held six nuclear tests over the past 20 years, an increase in regional seismic instruments shows that "there's been a lot more low-magnitude seismicity in the vicinity of test sites than we initially realized," Carmichael noted.

The new findings suggest that "background seismicity in regions where there's any sort of seismicity at all is going to measurably and substantially reduce the probability that we can detect signals from an underground explosion at a test site," he added.

The researchers also found that natural signals from earthquake swarms or other repeating seismic events could be similarly hidden by overlapping waveforms. In this case, the masking effect dropped detection from a 92% to a 16% detection rate.

"This may mean that we probably underestimate a lot of the low magnitude seismicity that is sourced during a swarm or an aftershock sequence," Carmichael said. "In other words, we could be largely undercounting the number of earthquakes that occur in these swarms or in certain aftershock sequences."

Explosion masking has been difficult to test because there are so few explosions to examine, and very few data sets that contain both explosion and natural seismic signals.

One way to analyze the potential effect would be to simulate explosion data, but Carmichael said there are "too many unknowns" with the high-frequency signals of explosions to accurately create the seismograms that represent relatively small-scale explosions measured at a distance.

Instead, he and his colleagues used a technique that draws from data collected on explosions and natural seismicity at the Nevada National Security Site.

The researchers developed a way to scale down the amplitude of the explosion data, "so that it mimics waveforms recorded from smaller explosions," Carmichael said. These explosion data are then "injected" back a set of earthquake signals, to see if sophisticated multi-channel correlation detectors can identify the explosion signal.

The technique creates a dense set of data that allowed the researchers to finally test the assertion that natural seismicity could not mask explosion signals, Carmichael said.

Researchers rely on multiple factors in addition to seismic signals as part of nuclear test monitoring, looking for other confirmatory evidence such as the presence of certain radionuclides in the atmosphere. It's unlikely that a coinciding earthquake would be enough to completely hide an explosion event.

But the new study offers "at least a recipe" on how to calculate the probability of explosion detection from seismic signals, so that this information can used together with other monitoring tools, said Carmichael.