By Kevin Wilcox
Researchers develop filters that can isolate seismic body waves created by ambient noise, opening the possibility of faster, cheaper seismic research.
Researchers are using the ambient noise from such sources as beach waves to economically study how seismic waves move through the earth. Wikipedia Commons/cuchulain
June 9, 2015—Researchers at Stanford University have developed a method to isolate seismic body waves hidden in ambient noise—ocean waves and city traffic, for example—and develop a three-dimensional representation of the pressure-wave velocities beneath the surface of the earth. By isolating these body waves, which travel within the surface of the earth, the team has unlocked the potential to make some forms of seismic research faster and much less expensive.
The research team was led by Nori Nakata, Ph.D., the George Thompson Postdoctoral Fellow in the Department of Geophysics at Stanford. Nakata's team recently published its findings, " Body wave extraction and tomography at Long Beach, California, with ambient-noise interferometry," in the
Journal of Geophysical Research
. The article's coauthors are Stanford colleagues Jesse Lawrence, an assistant professor of geophysics, Pierre Boué, Ph.D., a postdoctoral researcher, and graduate student Jason P. Chang. "Extracting body waves is much harder than surface waves because body waves in the ambient fields are much weaker than surface waves," said Nakata, in written answers to questions posed by
Civil Engineering
online. "However, we also know that ...body waves are useful for mapping elastic parameters of the earth materials (for example, how stiff the media are). This [was my] motivation to tackle this problem. As usual, a challenging problem is attractive for researchers."
Because ambient noise is random and undocumented, using it to isolate body waves has been a challenge for researchers. In contrast, when researchers conduct seismic research with active-source vibrations, they use specialized machines to create the vibrations at specific points, which is expensive and time-consuming. Because such vibrations are strong and their sources are fixed, however, body waves are easily discerned from them. "Ambient noise is not very easy to use because we do not control the origin of noise. But ambient noise always exists and propagates through the earth. If we can use ambient noise fields properly, we can obtain the information about the [earth materials]," Nakata said. And because ambient noise is continuous, researchers can use it to monitor the material for changes over time, which is very expensive to do when using vibrations generated by equipment.
The team focused on a site in Long Beach, California, using an existing data set from a dense array of 2,500 sensor nodes, spaced 100 m apart. The data were recorded by NodalSeismic, of Signal Hill, California, for Signal Hill Petroleum, a local private energy company. Each of the 2,500 nodes included a geophone receiver, a Global Positioning System (GPS) receiver, a battery, and a computer device with memory storage. The self-contained nodes eliminated the need for a network of wired connections over the 8.5 by 4.5 km site. The researchers used a data set from 10 days over a three-month period.
The biggest challenge for the team was isolating the body waves in the data. "Because the signal-to-noise ratio of body waves is very, very low—and body waves are buried in other waves—isolating body waves from recorded ambient noise is difficult," Nakata said. "This is the reason why people [could not] extract them for a long time."
For this reason, filtering out all the other waves was critical. "I knew the key point is how to construct a good filter to reconstruct clean body-wave signals, so I created as many filters as I can conceive," Nakata said. The filters identified data correlations that pointed to relatively strong body-wave energy. Using a time window enabled the team to separate out the surface wave energy. Waveform matching and noise suppression filters were used to further isolate the body-wave energy.
Then through travel-time tomography, the researchers used the isolated body-wave data to develop a 3-D image of the composition and structure of the materials beneath the surface. The oil field was ideal for the research because Signal Hill Petroleum had already surveyed it, employing active sources that provide validation for the new ambient-noise method.
"As [for] creating a filter and testing it, we needed this situation," Nakata said. "I found many similar features in my map of the subsurface structure compared with [those shown by] the active sources, which is great. So, now we can apply this technique to different places to find new seismic elements. And we can use this technique [at Long Beach] for time-lapse surveys, which are not done with active sources."
Nakata said the filter technique can use ambient noise to isolate body waves over the entire United States. Next he hopes to use the technique on a much larger scale and to apply it to a single site over a longer period of time in order to measure changes to the elastic properties of earth materials.