Contact USGS

  • About
  • Observatories
  • Activity
  • Education
  • Publications

Tiny Bubbles in the Seismic Noise

USGS scientist Phil Dawson and colleagues have applied a novel research approach to voice recognition software. In their January 2012 paper, published in Geophysical Research Letters, they utilize this software to discover that background seismic activity in geyser basins can be intimately linked to daily cycles of heating and cooling.

The authors looked at data obtained in 2003, when colleagues from the Yellowstone Volcano Observatory placed a variety of seismic and other monitoring stations in the Norris Geyser Basin to try and understand the cause for a "disturbance" of increased surface temperatures. There were few, if any, measurable earthquakes originating in the geyser basin during that time period, but a closer look at the constant background seismic noise revealed the presence of millions of tiny events. These events represent the constant bubbling, boiling and crackling of the geyser basin.

Earthquakes in volcanic and geothermal areas often happen in the same place in the same way. This means that what we record at the seismic sensors often produces the same data-output patterns. In effect, the continuous seismic record from a station on a volcano can be considered a song and the individual events as the repetition of specific words in the song. For 5 or 6 years, anyone who looked at the data saw primarily undecipherable "seismic noise". Dawson tried using voice recognition software supplied by his colleague Carmen Benítez, a signal-processing expert at the University of Granada in Spain, to identify the small seismic events. The software was able to detect specific types of small seismic events embedded in the background noise. These events were interpreted to be due to the violent collapse of small steam bubbles during their rise toward the surface.

Dawson trained the software to recognize these events and count their abundance. He also tracked a persistent tremor clearly linked to movement of water in shallow subsurface cracks. He found that the region of the disturbance, in Norris' Back Basin (near seismic station N02 in the upper figure) had more active and consistent noise than surrounding areas at Norris.

Most surprising was that the occurrence of these small seismic events increased in the late morning and persisted at high levels through the early evening. The activity correlated with air temperature, but even more remarkably, tracked with a mid-day decrease in the water temperature in hot pools in Norris' Back Basin. It appears that the drop in temperature of the ground water increased the collapse rate of the bubbles (and thus the noise). Dawson speculates that surface evaporation may contribute to the cooling. His hypothesis appears to be backed up by an increase in the seismic noise during cooling rains.

In a similar study using voice recognition software, Dawson and colleagues tracked the bursting of large slugs of gas in the lava lake in the Halemaʻumaʻu pit crater of Kīlauea Volcano, Hawaiʻi. In that study, published in Geophysical Research Letters in 2010, nearly 50,000 bursts were identified over a 2.5-year-long period. Together, these two papers demonstrate the power of voice recognition in identifying seismic events that are produced by combined effects of fluid movements within the Earth. In the future, this technique will allow scientists to listen to the 'seismic songs' generated by volcanoes and geothermal systems, which will help them identify and characterize the individual and repetitive words imbedded in the music. The article can be found on the journal website.

The research couldn't have been done without the contributions of YVO colleagues Bob Smith from the University of Utah and Henry Heasler at Yellowstone National Park. Smith coordinated the seismic experiment in 2003 and Heasler collected much of the key temperature and meteorological data that proved so useful in the final study.

Reference Dawson, P. B., M. C. Benítez, J. B. Lowenstern, and B. A. Chouet (2012), Identifying bubble collapse in a hydrothermal system using hidden Markov models, Geophys. Res. Lett., 39, L01304, doi:10.1029/2011GL049901
To learn more about the types of monitoring at Yellowstone, read our Frequently Asked Questions.

Other items of interest