Premium
Toward reliable automated estimates of earthquake source properties from body wave spectra
Author(s) -
Ross Zachary E.,
BenZion Yehuda
Publication year - 2016
Publication title -
journal of geophysical research: solid earth
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.983
H-Index - 232
eISSN - 2169-9356
pISSN - 2169-9313
DOI - 10.1002/2016jb013003
Subject(s) - directivity , deconvolution , waveform , seismology , seismic moment , spectral line , geology , window (computing) , energy (signal processing) , event (particle physics) , seismic wave , acoustics , algorithm , computer science , physics , mathematics , statistics , operating system , telecommunications , fault (geology) , radar , quantum mechanics , astronomy , antenna (radio)
We develop a two‐stage methodology for automated estimation of earthquake source properties from body wave spectra. An automated picking algorithm is used to window and calculate spectra for both P and S phases. Empirical Green's functions are stacked to minimize nongeneric source effects such as directivity and are used to deconvolve the spectra of target earthquakes for analysis. In the first stage, window lengths and frequency ranges are defined automatically from the event magnitude and used to get preliminary estimates of the P and S corner frequencies of the target event. In the second stage, the preliminary corner frequencies are used to update various parameters to increase the amount of data and overall quality of the deconvolved spectral ratios (target event over stacked Empirical Green's function). The obtained spectral ratios are used to estimate the corner frequencies, strain/stress drops, radiated seismic energy, apparent stress, and the extent of directivity for both P and S waves. The technique is applied to data generated by five small to moderate earthquakes in southern California at hundreds of stations. Four of the five earthquakes are found to have significant directivity. The developed automated procedure is suitable for systematic processing of large seismic waveform data sets with no user involvement.