Dip-dependent variations in LFE duration during ETS events

Abstract:

Using data from the Array of Arrays experiment, we create a new, more spatially complete catalog of LFEs beneath the Olympic Peninsula, WA. Using stacked waveforms produced by stacking 1-minute windows of data from each array over the slowness with the greatest power [Ghosh et al., 2012], we pick out peaks in tremor activity that are consistent over multiple arrays. These peaks are potential LFE detections. Fifteen-second windows of raw data centered on each peak are scanned through time. If the waveform repeats, the detection is used as a new LFE family. Template waveforms for each family are created by stacking all windows that correlate with the initial detection. During an ETS event, activity at a given point on the plate interface (i.e. the activity of an LFE family) typically lasts for 3.5 (downdip) to 5 days (updip). Activity generally begins with a flurry of LFEs lasting 8 hours (downdip) to 20 hours (updip) followed by many short bursts of activity separated by 5 hours or more. Updip families have more bursts (5-10) than downdip families (2-5 bursts). The later bursts often occur during times of encouraging tidal shear stress, while the initial flurries have no significant correlation with tides. While updip LFE families are more active during ETS events than downdip families, they seldom light up between ETS events, which only occur every 12-14 months. On the other hand, downdip LFE families are active much more frequently during the year; the most down-dip families exhibit activity every week or so. Because updip families are rarely active between ETS events, it is possible that little stress is released updip during inter-ETS time periods. Hence during ETS events more stress needs to be released updip than downdip, consistent with the longer-duration activity of updip LFE families.