Interpretation of Earthquake Epicenter and CMT Centroid Locations, in Terms of Rupture Length and Direction

Gideon P. Smith1, and Göran Ekström

Department of Earth and Planetary Sciences, Harvard University, Cambridge, MA,  USA.
1Now at Department of Earth and Planetary Sciences, Washington Univ., St. Louis, MO, USA
Phys. Earth and Plan. Int., 102, 123-132, 1997.

Abstract

A comparison is made between hypocenter and centroid estimates of shallow earthquakes to investigate whether the difference between these two quantities contains useful information on fault rupture length and direction. An initial comparison of ISC and Harvard CMT catalogue locations suggests that for earthquakes below MW > 6.5 the difference in location is dominated by location errors, whilst larger earthquakes appear to show the effect of increasing spatial rupture. However, for this data set combined location errors greater than 25 km are observed for a majority of earthquakes imposing a limit on interpretation. We use these constraints to investigate this approach using a consistent Earth model in the re-analysis of 8, previously well-studied, events. New hypocenter estimates based on teleseismic P-wave arrivals from the ISC Bulletin, and a 3-D velocity model of the mantle (S&P12/WM13) are compared to new centroid locations obtained using the same 3-D model, and waveform data from IDA, GDSN, and re-calibrated HGLP stations. The distance between our new hypocenter and centroid locations is shown to provide reliable information on both the rupture length and direction. The rupture lengths and azimuths generally agree with the most recent, previous results from other methods and provide an absolute estimate of the location of the rupture.

This Figure shows the match between 6 initial rupture vectors inferred in this study (white arrows) and those from previous studies (black arrows). Each arrows base is at the travel-time location and are proportional in length to each other.