Improved Methods for Regionalized Surface Wave Analysis.
Abstract
We develop optimized methods for measuring surface waves for event screening under a CTBT. The optimization techniques are: a regionalized earth model, phase-matched filtering, path corrected surface wave spectral magnitude; maximum likelihood magnitudes and upper bounds. A regionalized earth model is developed by tomographic inversion of surface wave data starting with the Crust 5.1 earth models, and using dispersion measurements from PIDC data, historic explosion data, and data sets from the University of Colorado, Harvard, and St. Louis University. Dispersion curves derived from the final earth models are sufficient to predict dispersion and to generate phase-matched filters for most regions of the world. The detection threshold of the automatic surface wave processor at the PIDC is found to be one magnitude unit lower than NEIC bulletins. Maximum likelihood magnitudes are calculated for a data set of 700 events using 10,000 PIDC and historical explosion seismograms. Maximum likelihood magnitudes combined with maximum likelihood upper bounds allow discrimination to about mb 4.0. The Ms:mb discrimination line between earthquakes and explosions has a slope of 1.4. We recommend redefining Ms as Ms = log A/T+k logdelta+1/2 log(sindelta)+gammadeltalog e+D using the parameters found by Rezapour and Pearce (1997), and recommend using path corrected spectral magnitudes as an alternative to Ms.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1997
- Accession Number
- ADA337234
Entities
People
- J. L. Stevens
- K. L. Mclaughlin