McCausland, W. A.

We describe a new method to estimate the S‐P time of tremor‐like signals and its application to the nonvolcanic tremor recorded in July 2004 by three dense arrays in Cascadia. The cross correlation between vertical and horizontal components indicates that very often the high‐amplitude tremor signal contains sequences of P and S waves characterized by constant S‐P times (TS‐P) in the range 3.5–7 s. A detailed observation of the three component seismograms stacked over the array stations confirms the presence of P and S wave sequences. The knowledge of the TS‐P poses a strong constrain on the source‐array distance, which dramatically reduces the uncertainty on source locations when used with more traditional array processing techniques. Data were analyzed using the zero lag cross‐correlation technique (ZLCC) to estimate the propagation properties of the most correlated phases in the wavefield. Detailed polarization analyses were computed using the covariance matrix method in the time domain. Polarization parameters, joint with the results of ZLCC, allows for the discrimination between P and S coherent waves. Results show that the tremor wavefield is composed mostly by shear waves, although a consistent amount of coherent P waves is often observable. The comparison of the back azimuth at the three arrays indicate that the source of deep tremor migrates over a wide area, and often many independent sources located far from each other are active at the same time. The tremor source was located by a probabilistic method that uses the results of ZLCC, given a velocity model. When available, the inclusion of the TS‐P time in the location procedure strongly reduces the depth range, with a distribution of hypocenters very near the subduction interface. This result, significantly different compared with previous less precise locations, makes the Cascadia nonvolcanic tremor more similar to the nonvolcanic tremor recorded in Japan, at least in cases of measurable TS‐P. The polarization azimuth aligned with the slow slip direction and the source located on the plate interface indicate that deep tremor and slow slip are two different manifestations of a common phenomenon related with the subduction dynamics.

We have here analyzed local and regional earthquakes using array techniques with the double aim of quantifying the errors associated with the estimation of propagation parameters of seismic signals and testing the suitability of a probabilistic location method for the analysis of nonimpulsive signals.We have applied the zero-lag cross-correlation method to earthquakes recorded by three dense arrays in Puget Sound and Vancouver Island to estimate the slowness and back azimuth of direct P waves and S waves. The results are compared with the slowness and back azimuth computed from the source location obtained by the analysis of data recorded by the Pacific Northwest seismic network (PNSN). This comparison has allowed a quantification of the errors associated with the estimation of slowness and back azimuth obtained through the analysis of array data. The statistical analysis gives σBP 10° and σBS 8° as standard deviations for the back azimuth and σSP 0:021 sec= km and σSS 0:033 sec =km for the slowness results of the P and S phases, respectively. These values are consistent with the theoretical relationship between slowness and back azimuth and their uncertainties. We have tested a probabilistic source location method on the local earthquakes based on the use of the slowness estimated for two or three arrays without taking into account travel-time information. Then we applied the probabilistic method to the deep, nonvolcanic tremor recorded by the arrays during July 2004. The results of the tremor location using the probabilistic method are in good agreement with those obtained by other techniques. The wide depth range, of between 10 and 70 km, and the source migration with time are evident in our results. The method is useful for locating the source of signals characterized by the absence of pickable seismic phases.

Tectonic tremor has been recorded at many subduction zones, including the Nankai, Cascadia, Mexican, and Alaskan subduction zones. This study, the first to use small aperture seismic arrays to track tremor, deployed three small aperture seismic arrays along the Cascadia subduction zone during a tremor and slow slip episode in July 2004. The tremor was active during virtually all (up to 99%) minutes of the analyzed tremor episode using 5 min sample windows. Individual wave phases were tracked across the arrays and used to derive slowness vectors. These were compared with slowness vectors computed from a standard layered Earth model to derive tremor locations. Locations were stable within a volume roughly 250 km2 in epicenter and 20 km in depth for hours to days before moving to a new volume. The migration between volumes was not smooth, and the movement of the sources within the volume followed no specific pattern. Overall migration speeds along the strike of the subduction zone were between 5 and 15 km/d; smaller scale migration speeds between volumes reached speeds up to 2 km/min. Uncertainties in the best locations were 5 km in epicenter and 10 km in depth. For this data set and processing methodology, tremor does not locate predominately on the primary subduction interface. Our favored model for the generation of tectonic tremor signals is that the tremor is triggered by stress and fluid pressure changes caused by slow slip and is composed, at least in part, of low‐frequency earthquakes broadly distributed in location

Preliminary analysis of deep tremor recorded during July, 2004, in the Cascadia Subduction zone shows that the use of small aperture arrays can resolve the slowness and back azimuth of seismic waves with a useful resolution. Data were collected by three dense arrays of short-period seismometers specifically deployed in the Puget Sound area under an US-Italy-Canada cooperative effort. Slowness analyses at the three arrays indicate that the 2-4 Hz tremor wave-field is composed by waves propagating with apparent velocities higher than 4 km/s. Combining this with polarisation analysis show these waves to be transverse (SH) waves. However, P-waves, though smaller in amplitude, can be detected by different slowness values obtained for the radial and transverse components. The intersection of wave vectors determined by the back azimuth and slowness values measured at the three arrays provides a preliminary estimate of source location for a sample of the recorded deep tremor.