600 channel recording of pseudo nuclear bomb

Thirty or forty years ago, underground nuclear explosions were common but computers and high density seismology were unknown. International treaty now prohibits such explosions, a benefit to humanity, but a loss to seismology. In 1993 at the Nevada Test Site (NTS) there was a chemical explosion of 1.3 kilotons. Fortuitously, the Subsurface Exploration Company of Pasadena, CA was operating a 609 channel petroleum-exploration seismic array 200 km distant. This recording equipment was activated at midnight and recorded the blast. This data set is remarkable in two respects: The 609 channel linear array is transverse to the direction of propagation. On a layered earth, all 609 channels would exhibit the same signal, yet we see they are remarkably diverse. The recorders are open to much higher frequencies than conventional earthquake seismometry. Here we see 10 Hz energy. This remarkable data set was nearly lost now I have a tape image here and a movie of it. We have recovered only the first two minutes of what is thought to be a 20-30 minute recording. The time axis is so long that I present it as 16 frames of 8 seconds each. You'll notice the frequencies getting lower with time as the angular bandwidth increases. The lateral variations on this data contradict the assumptions of most people's seismic data analysis which expects no variation transverse to the wavefront.