Detailed Tremor Migration Styles in Guerrero, Mexico Imaged with Cross-station Cross-correlations

Abstract:

Tremor occurred downdip of the area that slipped the most during the 2006 slow slip event (SSE) in Guerrero, Mexico, as opposed to Cascadia, where tremor locations and rupture zones of SSEs largely overlap. Here we obtain high resolution tremor locations by applying cross-station cross-correlations [Armbruster et al., 2014] to seismic data from the Meso-America Subduction Experiment deployment. A few 3-station detectors are adopted to capture detailed deformation styles in the tremor “transient zone” and the downdip “sweet spot” as defined in Frank et al., 2014. Similar to Cascadia, tremor activities in our study region were comprised mostly of short tremor bursts lasting minutes to hours. Many of these bursts show clear migration patterns with propagation velocities of hundreds of km/day, comparable to those in Cascadia. However, the propagation of the main tremor front was often not in a simple unilateral fashion. Before the 2006 SSE, we observe 4 large tremor episodes during which both the transient zone and the sweet spot participated, consistent with previous findings [Frank et al., 2014]. The transient zone usually became active a few days after the sweet spot. We find many along-dip migrations with recurrence intervals of about a half day within a region about 10 km along strike and 35 km along dip in the sweet spot, suggesting possible tidal modulation, after the main front moved beyond this region. These migrations appear not to originate at the main front, in contrast to tremor migrations from a few km to tens of km across observed in Cascadia [Rubin and Armbruster, 2013; Peng et al., 2015; Peng and Rubin, submitted], but possibly similar to Shikoku, Japan [Shelly et al., 2007]. We do not observe obvious half-day periodicity for the migrations farther downdip within the sweet spot. During the SSE, the recurrence interval of tremor episodes decreased significantly in both the transient zone and the sweet spot, with that of the former being much shorter. Within the sweet spot, the number of tremor events and the rupture area of each episode also decreased relative to the large pre-SSE episodes. The updip portion of the sweet spot often became active before the downdip portion, consistent with loading from farther updip where the slow slip was concentrated. In contrast, the pre-SSE episodes generally originated farther downdip.