HR: 17:00h
AN: S54B-05    [Abstracts]
TI: Robust Canonical Coherence for Quasi-Cyclostationary Processes: Geomagnetism and Seismicity in Peru
AU: * Lepage, K
EM: lepage@mast.queensu.ca
AF: Queen's University, Department of Math & Stats Jeffery Hall, University Ave., Kingston, ON K7L3N6, Canada
AU: Thomson, D J
EM: djt@mast.queensu.ca
AF: Queen's University, Department of Math & Stats Jeffery Hall, University Ave., Kingston, ON K7L3N6, Canada
AB: Preliminary results suggesting a connection between long-period, geomagnetic fluctuations and long-period, seismic fluctuations are presented. Data from the seismic detector, NNA, situated in ~Naña, Peru, is compared to geomagnetic data from HUA, located in Huancayo, Peru. The high-pass filtered data from the two stations exhibits quasi-cyclostationary pulsation with daily periodicity, and suggests correspondence. The pulsation contains power predominantly between 2000 μ Hz and 8000 μ Hz, with the geomagnetic pulses leading by approximately 4 to 5 hours. A many data section, multitaper, robust canonical coherence analysis of the two, three component data sets is performed. The method, involving an adaptation, suitable for quasi-cyclostationary processes, of the technique presented in "Robust estimation of power spectra", (by Kleiner, Martin and Thomson, Journal of the Royal Statistical Society, Series B Methodological, 1979) is described. Simulations are presented exploring the applicability of the method. Canonical coherence is detected, predominantly between the geomagnetic field and the vertical component of seismic velocity, in the band of frequencies between 1500 μ Hz and 2500 μ Hz. Subsequent group delay estimates between the geomagnetic components and seismic velocity vertical at frequencies corresponding to large canonical coherence are computed. The estimated group delays are 8 min between geomagnetic east and seismic velocity vertical, 16 min between geomagnetic north and seismic velocity vertical and 11 min between geomagnetic vertical and seismic velocity vertical. Possible coupling mechanisms are discussed.
DE: 0520 Data analysis: algorithms and implementation
DE: 3255 Spectral analysis (3205, 3280)
DE: 3265 Stochastic processes (3235, 4468, 4475, 7857)
DE: 3270 Time series analysis (1872, 4277, 4475)
SC: Seismology [S]
MN: 2007 Fall Meeting