Space-Time Variations in Tidal Stress and Cascadia Tremor Amplitude

We present a new analysis of the influence of tidal stresses on the amplitude of non-volcanic tremor in Washington State. Tremor counts (Thomas et al., 2009), tremor amplitude (Rubinstein et al., 2008), and strain (Hawthorne and Rubin, 2010) are modulated by tidal stresses in Cascadia as well as in California. However, tremor amplitudes have not yet been extensively studied in Cascadia. Furthermore, Hawthorne and Rubin's Cascadia-wide tidal stress model (2010) allows us to look at the tremor-tide relationship in more detail than ever before. The ability to look at the tidal modulation of tremor amplitude in space as well as time will increase our understanding of this phenomenon and may provide information about the frictional properties of the plate interface. We focus on the August 2010 episodic tremor and slip (ETS) event recorded by the Array of Arrays, a seismic experiment on the Olympic Peninsula. The instrument response is deconvolved, seismograms band-pass filtered at 1.5-5.5 Hz and envelopes are made in 5-minute windows. An inverse problem compensates for site corrections and source-receiver distances to produce, for any given time, a single amplitude measurement at the source. Source locations are determined using an envelope waveform cross-correlation method. Then, we compare the amplitudes, catalog of tremor locations, and the tidal stress at the desired location and time. Amplitudes during the August 2010 ETS event are clearly modulated by tidal stresses. Viewed in the frequency domain, there are clear peaks in the tremor amplitude spectrum at several tidal periods, most prominently the 12.4 and 24 hour periods. Comparison with Hawthorne and Rubin's tidal stress model shows that higher amplitudes are associated with positive shear stress in the downdip direction and, less strongly, with more compressional normal stress.