Six audio files have been created of the first 4 minutes and 22 seconds of the September 28, 2004 M 6.0 Parkfield main shock as it was observed in at the EarthScope's San Andreas Fault Observatory at Depth (SAFOD). The earthquake "sounds" were detected by a geophone installed deep underground in the SAFOD Pilot Hole. The "sound" is the vertical component of velocity as observed at a depth of 3465' on a 15 Hz geophone. The high-cut filter in the original recordings is at 200 Hz (500 samples/s data). The seismograms were provided by Peter Malin, Eylon Shalev and Andres Chavarria of Duke University.
During the 2004 Parkfield earthquake, the San Andreas Fault ruptured from southeast to northwest from an epicenter 30 km southeast of SAFOD. Initial models of the rupture suggest that it stopped less than 5 km from the Pilot Hole. The recording begins with the P-waves from the M 6.0 main shock and rapidly grows to a crescendo in the first 4 seconds as the S-waves arrive. The main shock coda begins 8-10 seconds into the record. As it decays away, aftershocks can be heard as distinct burst and bangs.
The sound files were created by using SAC2000 to upsample the original data and Matlab to write the wav files. The wav files are automatically clipped to fall within a range of -1 to 1 (e.g. amplitudes larger than 1 are replaced by 1).
The sped up ones sound a lot better on most speakers. The main shock sounds great at real-time on a sound system with subwoofers, but much of the remainder is very low frequency and hard to hear. The real-time files sound better if your player has a 'fast forward' control (4x in Windows Media Player).
Andy Michael and Bill Ellsworth
U.S. Geological Survey
Menlo Park, CA