Statistical Frameworks for Seismic Discrimination,
Abstract
To verify compliance with a Comprehensive Test Ban Treaty (CTBT), low energy seismic activity must be detected and discriminated. Monitoring small-scale seismic activity will require regional monitoring capabilities (within approximately 2000 km, U.S. Congress (1988)). The reliable discrimination of small-scale seismic events requires a multi-dimensional representation of the seismic signal. A multi-dimensional characterization might include wave arrival times, magnitudes, and incidence and azimuth angles. These measurements can be used singly or combined to form discriminants, which are then subjected to a set of discrimination rules to categorize the source event. Statistical discrimination methods of this type require a training sample, i.e., a set of real or simulated seismic data used to optimize or tune the discrimination algorithm by assigning weights to the various discriminants, or by modifying the structure of the algorithm. Identifying the signatures of various seismic sources also requires a geologic characterization of the shallow structure of the earth in each particular region of interest. The results can be used to construct seismic signals representing nuclear test sources. These synthetic or simulated signals can be combined with empirical signals of earthquakes and mining activities to form a training sample for each region. This paper identifies several statistical issues that must be resolved in order to address the CTBT verification mission. These are all associated with uncertainties in the multidimensional characterization measurements or in the correlations among them.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 14, 1995
- Accession Number
- ADP204405
Entities
People
- C. J. Young
- D. N. Anderson
- D. N. Hagedorn
- K. K. Anderson
- S. R. Sain