Azimuthal Effects on Magnitude Calculations for a Sparse Network: Eastern Canada

Session: Seismicity and Tectonics of Stable Continental Interiors
Type: Oral
Date: 4/30/2020
Time: 01:30 PM
Room: 240
Description: 

Earthquake magnitudes may be influenced by many factors. Distance, attenuation, frequency and site effects are generally considered in the magnitude equation. Azimuthal variations are not. For earthquakes recorded by sparse networks such as that in eastern Canada, station coverage is rarely uniform. This project seeks to determine whether and if so, how much, magnitudes are affected by an azimuthally uneven station distribution. The data set consists of earthquakes in eastern Canada over a ten year period with magnitudes calculated at a minimum of twenty stations. Station magnitudes were averaged for predefined azimuthal windows and an event magnitude was calculated by taking the mean of these values. These were compared to magnitudes calculated treating all stations equally. The mean difference was 0.02, suggesting that the magnitudes are not greatly influenced by the station distribution. There was no correlation between the magnitude difference and either the total number of azimuthal windows containing stations nor the percentage of windows with only a single station. One interesting observation was that for the small number of events for which the magnitude difference was greater than 0.1, the azimuthally averaged magnitude was almost always less than the initial magnitude, which, with some added complexity, is consistent with the premise behind maximum likelihood magnitude calculations. A second test treated stations in the Charlevoix Seismic Zone (CSZ), which cover a narrow azimuth and distance range with respect to epicenters not in the CSZ, as a single point and all other stations as individual points. The mean difference was 0.0 with respect to the original magnitudes. These analyses will be repeated including site corrections, not applied in eastern Canada but shown to have an effect comparable to azimuthal averaging, to determine whether the effects of one cancels or enhances the other, and the data set will be expanded.

Presenting Author: Allison L. Bent

Authors