2009 Portland GSA Annual Meeting (18-21 October 2009)

Paper No. 4
Presentation Time: 9:00 AM

TREMOR PATCHES AND STREAKS IN CASCADIA


GHOSH, Abhijit1, VIDALE, John E.2, SWEET, Justin R.1, CREAGER, Kenneth C.1 and WECH, Aaron G.1, (1)Earth & Space Sciences, University of Washington, Seattle, WA 98195-1310, (2)Earth & Space Sciences, University of Washington, Johnson Hall Rm-070 Box 351310, 4000 15th Avenue NE, Seattle, WA 98195, aghosh.earth@gmail.com

We installed an 84-element temporary seismic array in northern Washington over the migration path of non-volcanic tremor (NVT). The array captured a good part of the episodic tremor and slip (ETS) event in May 2008, and a preceding much smaller event two months earlier. We apply a beamforming technique to track NVT in unprecedented detail. It detected up to 4 times more duration of tremor than the conventional envelope cross-correlation method during the period of weak NVT activity. We locate tremor assuming it is at the subduction interface, estimate relative tremor moment released by each detected tremor window, and map it on the interface. This procedure reveals that NVT activity is not uniformly distributed across the ETS zone, and unveils several distinct patches that released much of the tremor moment during 2008 ETS event. The moment patches appear to be devoid of ordinary earthquakes, which may indicate that the mode of stress release is controlled by the heterogeneity in fault strength along the plate interface.

The high resolution of our method also allows us to observe the details of NVT migration on the time scale of hours. For the first time in Cascadia, we found convergence-parallel transient streaks of tremor migrating at much faster velocities (several tens of km/hr), with movement in both up- and down-dip directions. This is in contrast to the long-term along-strike marching of tremor at 12 km/day. These streaks tend to propagate steadily, often repeating the same track on the interface multiple times. They light up the moment patches by a combination of increased amplitude and/or longer residence time within the patches. Physical mechanism causing the streaks, however, is not entirely clear. Perhaps surges of fluid released at the megathrust transmit stress pulses causing failure and/or opening cracks along the linear corrugated weak features on the fault.