Investigation of Backprojection Uncertainties With M6 Earthquakes

TitleInvestigation of Backprojection Uncertainties With M6 Earthquakes
Publication TypeJournal Article
Year of Publication2017
AuthorsFan W.Y, Shearer PM
JournalJournal of Geophysical Research-Solid Earth
Volume122
Pagination7966-7986
Date Published2017/10
Type of ArticleArticle
ISBN Number2169-9313
Accession NumberWOS:000418577900028
Keywords2015 gorkha earthquake; Back-Projection; backprojection; california; california hypocenter relocation; cluster-analysis; denali fault earthquake; earthquake source imaging; form cross-correlation; hybrid; japan; location techniques; source-receiver reciprocity; southern; subduction zone; tohoku-oki earthquake; Uncertainty analysis
Abstract

We investigate possible biasing effects of inaccurate timing corrections on teleseismic P wave backprojection imaging of large earthquake ruptures. These errors occur because empirically estimated time shifts based on aligning P wave first arrivals are exact only at the hypocenter and provide approximate corrections for other parts of the rupture. Using the Japan subduction zone as a test region, we analyze 46 M6-M7 earthquakes over a 10year period, including many aftershocks of the 2011 M9 Tohoku earthquake, performing waveform cross correlation of their initial P wave arrivals to obtain hypocenter timing corrections to global seismic stations. We then compare backprojection images for each earthquake using its own timing corrections with those obtained using the time corrections from other earthquakes. This provides a measure of how well subevents can be resolved with backprojection of a large rupture as a function of distance from the hypocenter. Our results show that backprojection is generally very robust and that the median subevent location error is about 25km across the entire study region (approximate to 700km). The backprojection coherence loss and location errors do not noticeably converge to zero even when the event pairs are very close (<20km). This indicates that most of the timing differences are due to 3-D structure close to each of the hypocenter regions, which limits the effectiveness of attempts to refine backprojection images using aftershock calibration, at least in this region.

DOI10.1002/2017jb014495
Student Publication: 
No
sharknado