Incorporating Love- and Rayleigh-wave magnitudes, unequal earthquake and explosion variance assumptions and interstation complexity for improved event screening

Conference · Thu Jan 01 04:00:00 EST 2009

OSTI ID:992208

Anderson, Dale N ^[1]; Bonner, Jessie L ^[2]; Stroujkova, Anastasia ^[2]; Shumway, Robert ^[3]; Russell, David R ^[4]

---

1. Los Alamos National Laboratory
2. WESTON GEOPHYSICAL
3. UC/DAVIS
4. AFTAC

Our objective is to improve seismic event screening using the properties of surface waves, We are accomplishing this through (1) the development of a Love-wave magnitude formula that is complementary to the Russell (2006) formula for Rayleigh waves and (2) quantifying differences in complexities and magnitude variances for earthquake and explosion-generated surface waves. We have applied the M{sub s} (VMAX) analysis (Bonner et al., 2006) using both Love and Rayleigh waves to events in the Middle East and Korean Peninsula, For the Middle East dataset consisting of approximately 100 events, the Love M{sub s} (VMAX) is greater than the Rayleigh M{sub s} (VMAX) estimated for individual stations for the majority of the events and azimuths, with the exception of the measurements for the smaller events from European stations to the northeast. It is unclear whether these smaller events suffer from magnitude bias for the Love waves or whether the paths, which include the Caspian and Mediterranean, have variable attenuation for Love and Rayleigh waves. For the Korean Peninsula, we have estimated Rayleigh- and Love-wave magnitudes for 31 earthquakes and two nuclear explosions, including the 25 May 2009 event. For 25 of the earthquakes, the network-averaged Love-wave magnitude is larger than the Rayleigh-wave estimate. For the 2009 nuclear explosion, the Love-wave M{sub s} (VMAX) was 3.1 while the Rayleigh-wave magnitude was 3.6. We are also utilizing the potential of observed variances in M{sub s} estimates that differ significantly in earthquake and explosion populations. We have considered two possible methods for incorporating unequal variances into the discrimination problem and compared the performance of various approaches on a population of 73 western United States earthquakes and 131 Nevada Test Site explosions. The approach proposes replacing the M{sub s} component by M{sub s} + a* {sigma}, where {sigma} denotes the interstation standard deviation obtained from the stations in the sample that produced the M{sub s} value. We replace the usual linear discriminant a* M{sub s}+b*{sub m{sub b}} with a* M{sub s}+b*{sub m{sub b}} + C*{sigma}. In the second approach, we estimate the optimum hybrid linear-quadratic discriminant function resulting from the unequal variance assumption. We observed slight improvement for the discriminant functions resulting from the theoretical interpretations of the unequal variance function. We have also studied the complexity of the ''magnitude spectra'' at each station. Our hypothesis is that explosion spectra should have fewer focal mechanism-produced complexities in the magnitude spectra than earthquakes. We have developed an intrastation ''complexity'' metric {Delta}M{sub s}, where {Delta}M{sub s} = M{sub s}(i)-M{sub s}(i+1) at periods, i, which are between 9 and 25 seconds. The complexity by itself has discriminating power but does not add substantially to the conditional hybrid discriminant that incorporates the differing spreads of the earthquake and explosion standard deviations.

• Research Organization: Los Alamos National Laboratory (LANL)
• Sponsoring Organization: DOE
• DOE Contract Number: AC52-06NA25396
• OSTI ID: 992208
• Report Number(s): LA-UR-09-05324; LA-UR-09-5324
• Country of Publication: United States
• Language: English

Related Subjects

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEFENSE
ATTENUATION
EARTHQUAKES
EXPLOSIONS
HYPOTHESIS
METRICS
MIDDLE EAST
MONITORING
NEVADA TEST SITE
NUCLEAR EXPLOSIONS
PERFORMANCE
RAYLEIGH WAVES
SEISMIC EVENTS
SEISMIC SURFACE WAVES
SPECTRA