Abstract
Abstract : A weakness of traditional methods of seismic event location is their susceptibility to errors in picking arrival times of regional phases such as Pn, Pg, and teleseismic phases. We evaluate the practical utility of waveform cross-correlation as a way to reduce or remove pick error, applied to large datasets in seismically active regions of East Asia, for purposes of obtaining significant improvements in event location. Key ideas in this project emerged during 2000-2003 from work of a Lamont-led consortium to calibrate stations in East Asia. Thus, we found for a number of regions that one of the best methods for obtaining ground truth events was to use waveform cross-correlations that enabled excellent (sub-kilometer) precision in relative locations. Additional (often, non-seismic) information then allowed absolute locations to be estimated. A preliminary study of China has shown that 1300 out of 14,000 earthquakes (approximately 9%) exhibit high cross-correlations with at least one other earthquake, and on this basis we have found 494 sets of cross correlated multiplets, ranging from doublets to one multiplet with 26 events. We apply waveform cross correlations to the problem of event location in four project areas, namely China, parts of Eastern Canada, the Central United States (New Madrid), and (for purposes of validation of our overall method) California. For each project area, we carry out the three steps of data acquisition (digital waveforms, traditional phase picks, catalog information), waveform cross-correlations (exploring the effects of different filters, spatial event separation, and differences in magnitude), and event location (double-difference). We produce data sets of waveforms, cross-correlation measurements of arrival differences, and sets of precise locations.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call