Classical and Bayesian Seismic Yield Estimation: The 1998 Indian and Pakistani Tests

-- The nuclear tests in May, 1998, in India and Pakistan have stimulated a renewed interest in yield estimation, based on limited data from uncalibrated test sites. We study here the problem of estimating yields using classical and Bayesian methods developed by Shumway (1992), utilizing calibration...

Full description

Saved in:
Bibliographic Details
Published in:Pure and applied geophysics 2001-10, Vol.158 (11), p.2275-2290
Main Author: Shumway, R H
Format: Article
Language:English
Subjects:
Availability
Bayesian analysis
Calibration
Confidence intervals
Density
Geophysics
Indian
Intervals
Nuclear tests
Regression
Seismology
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:-- The nuclear tests in May, 1998, in India and Pakistan have stimulated a renewed interest in yield estimation, based on limited data from uncalibrated test sites. We study here the problem of estimating yields using classical and Bayesian methods developed by Shumway (1992), utilizing calibration data from the Semipalatinsk test site and measured magnitudes for the 1998 Indian and Pakistani tests given by Murphy (1998). Calibration is done using multivariate classical or Bayesian linear regression, depending on the availability of measured magnitude-yield data and prior information. Confidence intervals for the classical approach are derived applying an extension of Fieller's method suggested by Brown (1982). In the case where prior information is available, the posterior predictive magnitude densities are inverted to give posterior intervals for yield. Intervals obtained using the joint distribution of magnitudes are comparable to the single-magnitude estimates produced by Murphy (1998) and reinforce the conclusion that the announced yields of the Indian and Pakistani tests were too high.[PUBLICATION ABSTRACT]
ISSN:0033-4553
1420-9136
DOI:10.1007/PL00001149