Time series of GPS data from Guerrero, southern Mexico, show brief (34 month) reversals of motion totalling anywhere from 1 to 10 cm. These events occur about once per year and indicate slow aseismic thrust slip occurring on the plate boundary. Three large-displacement events have been observed in campaign data around 1996 [ Larson et al., 2004], at campaign sites and three permanent GPS sites in 1998 [Lowry et al., 2001; Larson et al., 2004] and at eight permanent sites in 2002 [ Kostoglodov et al., 2003; Iglesias et al., 2004; Yoshioka et al., 2004]. Smaller but nevertheless significant displacements also occurred in 1999, 2000, 2001, 2003 and 2004 [Lowry et al., 2006]. Taken together, these eight events suggest that slow thrust slip recurs on the Guerrero subduction plate boundary at intervals of about one year.
Modeling of the Guerrero slow slip events (also referred to as "silent earthquakes") indicates that they occur in the transition from velocity weakening friction (i.e., the seismogenic zone) to velocity strengthening friction at greater depths [Lowry et al., 2006]. Interestingly, these events are made possible by unexpectedly strong frictional coupling of the plate boundary within the transition zone during inter-event periods. The minimum magnitude equivalents of moment release range from Mw=6.1 for the smallest events to 6.8 for the largest, and actual moment release is probably larger because this does not include deformation from slip outside the region of GPS measurements [Lowry et al., 2006].
Similar (albeit generally smaller) aseismic fault slip events have been observed in continuous GPS data from Cascadia, Alaska, Japan and New Zealand. All of these events appear to have activated areas of the subduction megathrust immediately down-dip of the seismogenic portion of the fault, and events in Cascadia and Japan also have a similar (approximately annual) periodicity.
In collaboration with Kristine Larson and Roger Bilham at the University of Colorado, and Vladimir Kostoglodov and other colleagues at UNAM, I am investigating several questions posed by these exciting new observations, including:
The large 200102 fault slip event began within weeks after a Mw=6.1 intraplate normal faulting event (the Coyuca earthquake) near the center of the region of early aseismic slip. GPS campaign data suggest that significant slip also occurred throughout Guerrero immediately after the 1995 Mw=7.3 Copala earthquake. It is possible that static stress change during the Coyuca and Copala earthquakes excited stress waves which propagated the aseismic events. I am currently developing InSAR interferograms to more closely examine the connection between the Copala earthquake and the 199596 Guerrero slip event, and in collaboration with John Anderson at U. Nevada-Reno, improving estimates of locations and other parameters of seismicity for the period in which we have GPS measurements. These will be combined in a numerical model of rate- and state-dependent friction to examine the relationship of geodetically observed slip to seismic events. Interestingly, there are no significant earthquakes anywhere on the Cocos-North Americ plate boundary in the months preceding the 1998 slow slip event, so I am also examining other possible triggering mechanisms (including resonant response to climate-driven and other mass loading processes that vary stress on the fault at the 6- to 25-month periods of slow slip recurrence).
Aseismic moment release requires a slip deficit in the accommodation of relative motion across the fault. I have developed inverse models of GPS and other geodetic data in Guerrero to estimate simultaneously the steady-state slip rate and transient slip. The model indicates that the aseismic slow slip events release a slip deficit accumulated during the interval between events. The model can also be used to evaluate seismic hazard. If one assumes that slip during the 1992-2005 period of GPS measurement is representative of the entire (1911-2005) aseismic period, the Guerrero gap segment of the subduction megathrust has accumulated sufficient potential energy to generate a Mw=7.9 earthquake. I also find that, if oblique Cocos-North America plate motion is partitioned into upper-plate strike-slip faulting, most of the ~8-10 mm yr-1 strike-slip movement is accommodated on the Chapala-Oaxaca fault zone with a small contribution of 0-3 mm yr-1 across the Atoyac fault.
Slip during aseismic events loads the up-dip, seismogenic portion of the fault by increasing the static shear stress. An important question that I hope to resolve via numerical modeling of dynamic friction relates to the conditions necessary for an aseismic slip event to generate a major earthquake rupture (instead of healing as the Guerrero events have done thus far). A second issue relates to why large fault slip events have been seen in Guerrero, Cascadia, Alaska and Japan but not in other subduction zones or on faults in other tectonic regimes. These locations are unusual in that they are all characterized by near-horizontal slabs at depths of 35-60 km: I.e., at transitional temperatures between pure stick-slip behavior and low-stress stable sliding. In Guerrero, nearly 100 km of down-dip fault surface is in the transitional slip regime, resulting in the potential for slip deficit over a broad area, and correspondingly, very large moment release. Faults that are more nearly vertical will likely have less than 10 km of surface in the transitional regime, so potential moment release would be proportionately smaller, and possibly too small to be seen with GPS (but still large enough to be observed with more sensitive instruments such as borehole strainmeters and long baseline tiltmeters).
|This material is based upon work supported by the National Science Foundation under grant numbers 9725712, 0125618, and 0207820. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.|