G31A-01 INVITED
Transient slip episode in central and southwest Japan.
We have studied transient slip episodes in central and southwest Japan, where the Philippine Sea plate is subducting at the annual rate of ~ 6.5cm/yr. A slow thrust slip event occurred in Bungo Channel, in southwest Japan, and anomalous displacements were found in at multiple GPS time series for stations in Shikoku and Kyushu islands. We analyzed those GPS data with time-dependent inversion method, and found that the slip initiated at a deeper part (~ 40km) of the plate boundary. The acceleration and following decelelation took longer than one year. The second event is the 2000 Tokai slow slip event, which initialed in early 2000. We also applied the same time-dependent inversion method to infer the space-time distribution of slip and slip-rate at the plate boundary. The slip initiated at the deeper part of the plate boundary (~ 40km), and migrated upward to the depth of ~ 30km. This event lasted for longer than 5 years. Having cumulative slip at each epoch, we are able to calculate shear stress change with dislocation theory. We examined velocity-dependence of the shear stress change to compare with a numerical simulation based on rate-state friction. The observed stress-velocity path is similar to what obtained for high-speed rupture, suggesting that slow slip events occurred in velocity weakening friction. This makes a clear contrast to what we obtained for the afterslip for the 2003 Tokachi-oki earthquake.
G31A-02
Preliminary Evidence for Non-Volcanic Tremor in the Oaxaca Region, Mexico
A multi-year deployment of a large aperture seismic and GPS array in southern Mexico, for the purpose of studying the Oaxaca segment of the Middle America subduction zone, entered its second development phase during the summer of 2006. The Oaxaca segment is one of the few places on earth where land-based geophysical observations can be used to study both the locked, seismogenic area of a subduction interface and the region of deeper transitional slip, where episodic tremor and slip may originate. The newly expanded array, consisting of nine broad-band seismic and eleven continuous GPS stations, measures deformation and ground vibrations in the state of Oaxaca and adjacent areas. The seismic array has nominal 100-km spacing and extends from the coast to ~300 km inland. Preliminary analysis of seismic data indicates that the stations are uniformly characterized by low background noise in the non-volcanic tremor passband of 1-5 Hz, accomplishing an important goal of our site selection. Furthermore, we find several periods of non-volcanic tremor coincident on at least three of our seismic stations, including 3 short episodes during the summer of 2006. Based on relative amplitudes of the recordings, these episodes "hopped" from west (late June) to east (late July) of Oaxaca city and then back west again (mid August). The episodes are most prominent at stations ~100 km from the coast, which is consistent with non-volcanic tremor occurring near the zone of transitional motion downdip from the seismogenic zone. As of February 2007, there have been no prominent slip transients recorded by the GPS stations in the Oaxaca array since the seismic recordings began in early June 2006, so it may be that these are small episodes of non-volcanic tremor in which accompanying slow slip is below the current noise level of GPS. This hypothesis will be tested by the next prominent slow slip event in this region, when one would expect to see larger seismic amplitudes from non-volcanic tremor.
G31A-03
A History of Slow Slip Events in the Mexican Subduction Zone
Aseismic slow slip events (SSEs) are a common and important feature of the seismotectonic process in the Mexican subduction zone. Large SSEs have been recorded in the central part of Mexico (Guerrero and Oaxaca states) in 1972, 1979 by tide gauges and later on in 1996, 1998, 2002, and 2006 by GPS network and a long- baseline tiltmeter. Comparing long term mean sea level (MSL) rise, ~4.5 mm/yr, and a secular subsidence rate of the GPS, ~11 mm/yr, in Acapulco tide gauge site, we may conclude that at least 9 mm/yr of vertical deformation is recovering episodically by SSEs (assuming 2.5 mm/yr eustatic MSL). The 1972 SSE was probably the largest observed in Mexico. It produced the vertical uplift of ~140 mm in Acapulco City and a noticeable uplift along the Pacific coast down to the town of Salina Cruz (more than 500 km SW from Acapulco). As the permanent GPS network "SSN-Sismología-UNAM"' was established in 1997 we could reliably register transient slips starting from the 1998 SSE. The last two SSEs of 2001-2002 and 2006 are the best studied. Characteristic duration of these events is 6-12 months. The equivalent magnitude exceeded Mw7.5 in 2002. All SSEs, with the exception of the 1996 event, initiated almost simultaneously in the coast (the NW Guerrero seismic gap) and ~180 km inland. Then SSEs propagated laterally along the strike of the subduction zone. However, the propagation rate of ~2 km/day could be estimated reliably only for the 2002 event. The observations indicate that the areas affected by the 1972 and 2002 slow events may have been greater than ~250x500 km2. The shallow, subhorizontal configuration of the plate interface in Guerrero and partly in Oaxaca appears to be a controlling factor for the physical conditions favorable for such extensive SSEs. The entire partially coupled interplate zone in Guerrero is of ~160 km width (starting ~55 km inland from the trench) while the seismogenic, shallowest part of it is only ~40 km wide. Different models involving the elastic half space dislocation concept could not yet definitely resolve between the case of transient slip of the order of 10 cm propagating into the shallow seismogenic zone and the opposite case when the slip of 15-20 cm is totally develops on the deeper transition zone of the plate interface. The hazard estimate of the future large subduction thrust earthquake in Guerrero fairly depends on the answer to this query. Thus a denser GPS network in Guerrero is required before the next SSE.
G31A-04 INVITED
GPS constraints on 35+ slow slip events within the Cascadia subduction zone, 1997- February, 2007
Refinements to GPS analyses in which we factor geodetic time series to better estimate both reference frames
and transient deformation resolve 35 slow slip events (SSE) located throughout the Cascadia subduction zone
from 1997 through early 2007. Timing of transient onset is determined with wavelet-transformation of the
geodetic time series. 30 continuous GPS stations are included in this study up through 2005, and over 70
stations for the 2007 event. Events are analyzed that range from northern California to southwestern British
Columbia, with station density generally increasing towards the north. The improved analyses better resolves the
largest creep and also identifies many smaller events. At 48.5N latitude, the 14-month average recurrence interval
still applies, four events after first recognition. Elsewhere, such periodicity is not observed. Along central
Vancouver Island to the north (49N), a host of smaller events distinct from the 14-month periodicity occur with no
obvious periodicity. Sporadic smaller events also appear throughout the subduction zone to the south, including
some within the region of the 14-month periodicity of larger events. In southern Washington State, some of the
largest transient displacements are observed, but lack any obvious periodicity in their recurrence. Along central
Oregon, an 18-month recurrence is evident, while in northern California (Yreka) the 11-month periodicity
continues through 2005. To invert GPS offsets of the 12 best-recorded events for slip, we use a cross-validation
scheme to derive optimal smoothing of non-negative thrust faulting along a plate interface divided into 40 along
strike and 24 down-dip subfaults. Those events have equivalent moment magnitudes ranging from 6.3 (smallest
resolvable with GPS) to 6.8, and typically 2-3 cm of slip. The largest spatial extent of all events resolved to date
is just under 350 km along strike, with a maximum observed duration of seven weeks across the network; the
majority last less than one month and show typically half the spatial extent. Unlike other subduction zones like the
Middle America Trench, no longer-duration Cascadia events are observed, nor cumulative deformation greater
than 0.6 cm. The frequency, size and style of the many newly resolved smaller transient deformation events show
they occur frequently here, that GPS captures only the largest events, and that smaller SSE's routinely occur at
much greater occurrence rate at levels not detectable with GPS. Moreover, the location of the slow slip events,
and the general pattern of deformation, suggest that the characteristics of seismogenic locking beneath
Cascadia remains enigmatic.
http:www.geodesy.org/pnwdaily/GPS_Summary/gps_summary.html
G31A-05
Slow slip events, temperature, and interseismic coupling at the Hikurangi subduction zone, New Zealand
The Pacific Plate subducts beneath New Zealand's North Island at the Hikurangi subduction zone. Campaign GPS results show that large portions of the subduction thrust are locked during the interseismic period. However, the down-dip limit of locking beneath the southern North Island is deeper (35 to 50 km) than beneath the northern and central North Island where it terminates at 10 to 15 km depth. Continuous GPS in New Zealand operating since 2002 has revealed several slow slip events (SSE) on the Hikurangi subduction zone. We observe a rich variety of SSEs with widely varying durations, depths, magnitudes and recurrence intervals. Thus far, we observe that deeper slow slip events (25 to 60 km depth) have longer durations (about 1 to 2 years) and longer times between them, while shallower slow slip events (10 to 15 km) occur more rapidly (within weeks) and more frequently. Despite the along-strike variability in the maximum depth of the interseismic locking, the SSEs recorded to date have all occurred at or near the down-dip transition from interseismic locking to creep. This observation allows us to examine the role of temperature in controlling the transition zone. We will present the results of methods to extract SSEs from the GPS time series and estimate fault zone temperatures where the SSEs occur.
G31A-06
Slow slip events at the Alaska Subduction Zone
A large slow slip event (SSE) occurred at the Alaska subduction zone during 1998-2001 [Ohta et al., 2006, EPSL]. The SSE occurs downdip of the Prince William Sound asperity, on a section that accumulated some slip deficit before and after the SSE. During the SSE, >20,000 sq. km of the plate interface slipped >10 cm, for a cumulative moment magnitude of Mw=7.2. Slip during this event was accompanied by significant non-volcanic tremor, which located in the same area as the slip. The area of the SSE did not have significant slip in the 1964 earthquake, compared to the slip of the main asperity. Our previous work did not attempt to determine the time history of slip in the SSE, because of a lack of continuous GPS data, but a reanalysis of the data that includes a few new sites allows us to estimate the approximate time history. Based on a preliminary solution, it appears that the initiation time probably varied over the area of the SSE, implying some spatial propagation, but our ability to resolve this is very limited. The continuous GPS record suggests that that may have been at least one and possibly several smaller SSEs in the same area. These smaller events appear to last for several weeks, but because most appear at a similar time of year, it is possible that they represent an unmodeled seasonal signal instead. We will compare the displacement pattern and time history of the large SSE to the displacements to evaluate whether these smaller events actually represent SSEs on the same part of the plate interface.
G31A-07 INVITED
Aseismic events in Southern California: Detection with InSAR
Aseismic slow slip events are usually studied using data types that have a dense temporal sampling rate, such as continuous GPS or tremor analysis using seismic data. However, even the sparser temporal coverage of InSAR data can further our understanding of these events in three significant ways - First, in areas where aseismic transients have been detected on geodetic arrays, InSAR may be able to provide a spatially denser image of the extent and magnitude of deformation. Second, InSAR observations are complementary to GPS because of the differing sensitivities to horizontal and vertical motions. Thirdly, in areas with no ground-based geodetic instrumentation, InSAR can be used in survey mode to detect deformation signals that are not associated with any observed seismicity. The temporal constraints on such signals may not be tight enough to allow for dynamics models of how aseismic transients occur, but InSAR-only detections can improve our understanding of the spatial extent of these types of events and can also identify key areas for future instrumentation and observation. Here, I summarize some of the contributions of InSAR observations of slow slip events, including data spanning the 2005 Obsidian Buttes swam in the Salton Trough, CA, and InSAR time-series results for the Salton Trough using both traditional interferometry and the persistent scatterer method.
G31A-08
Posteseismic Deformation Near Yucca Mt., Nevada
Geodetically, postseismic deformation can be identified through the nonlinear curvature in the temporal time series of site position. The ability to identify postseismic deformation is therefore dependent upon geodetic accuracy, length of time series, the spatial pattern of the postseismic deformation, and the characteristics of the causative seismic event, including its distance from the GPS sites. In this talk, we discuss three possible sources of postseismic deformation in GPS time series acquired from sites in the Yucca Mountain region: the Mw 7.3 Landers event (28 Jun 1992), the M_s 5.6 Little Skull Mountain Earthquake (29 Jun 1992), and the Mw 7.1 Hector Mine event (16 Oct 1999). Yucca Mountain is in the far field (200 km or more) of the Landers and Hector Mine events, and in the near field of the Little Skull Mountain event. These differences, and the fact that the permanent BARGEN GPS network did not begin operation in this area until 1999, mean that the events are sensed quite differently by the Yucca Mountain GPS networks.