Regional Small-Event Identification Using Seismic Networks and Arrays.
Abstract
The CTBT negotiations have endowed small seismo/acoustic events (mb ^ 2.5) with a much greater significance than they have had in the past and have increased the need for automated regional discrimination. We have been contracted by AFTAC to develop a small event discriminant that uses a time frequency expansion (sonogram) of seismic coda to discriminate ripple fired mining explosions from single explosions and earthquakes. The Automated Time Frequency Discriminant (ATFD) which we have developed uses a binary sonogram which is derived from the original, spectral, sonogram by the application of filters which replace the spectral information with a binary code which simply reflects local spectral highs and lows. We have found that the binary patterns which we extract from the coda of ripple-fired events are banded and thus distinct from those obtained from single explosions and earthquakes. The bands, which result from source finiteness, intershot delays or a combination of the two, are largely independent of time and of the recording component. The ATFD uses three statistical tests to measure the time and recording component independence and automatically recognize these bands. All of the raw discrimination parameters produced by these tests are merged into a single discriminant score with multivariate statistics.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 15, 1996
- Accession Number
- ADA319766
Entities
People
- Frank Vernon
- J. B. Minster
- John A. Orcutt
- Michael Hedlin
Organizations
- University of California, San Diego