T11F-01 INVITED
The complex evolution of transient slip revealed by precise tremor locations in western Shikoku, Japan
Transient slow slip events are increasingly being recognized as important components of strain release on faults and may substantially impact the earthquake cycle. Surface-based geodetic instruments provide estimates of the overall slip distribution in larger transients but are unable to capture the detailed evolution of such slip, either in time or space. Accompanying some of these slip transients is a relatively weak, extended duration seismic signal, known as non-volcanic tremor, which has recently been shown to be generated by a sequence of shear failures occurring as part of the slip event. By precisely locating the tremor, we can track some features of slip evolution with unprecedented resolution. Here, we analyze two weeklong episodes of tremor and slow slip in western Shikoku, Japan. We find that these slip transients do not evolve in a smooth and steady fashion but contain numerous sub-events of smaller size and shorter duration. In addition to along-strike migration rates of about 10 km/day observed previously, much faster migration also occurs, usually in the slab dip direction, at rates of 25-150 km/hour over distances of up to 20 km. We observe such migration episodes in both the up-dip and down-dip directions. These episodes may be most common on certain portions of the plate boundary that generate strong tremor in intermittent bursts. The surrounding regions of the fault may slip more continuously, driving these stronger patches to repeated failures. Tremor activity has a strong tidal periodicity, possibly reflecting the modulation of slow slip velocity by tidal stresses.
T11F-02 INVITED
35 Cascadia Episodic Tremor and Slip events observed on GPS, seismic, and strain/tiltmeter arrays
Several rapidly expanding GPS networks along the greater Cascadia forearc have enabled identification of 36 isolated Episodic Tremor and Slip (ETS) events since 1997, including two in 2007. ETS events are observed throughout the forearc, from northern California to southwestern British Columbia, with station density generally increasing towards the north. Events located in well-instrumented regions can be tracked as they migrate laterally north-south along the plate boundary, but increasing station density has resolved many smaller transients that could not previously be confidently identified. At the specific latitude of the northern Washington State and southwestern British Columbia, the 14-month average recurrence interval still holds true, 5 events after first recognition. Elsewhere, this periodicity is not observed. Along central Vancouver Island, a host of smaller events distinct from the 14-month recurrence occur with an aperiodic fashion. Sporadic smaller events also appear throughout the subduction zone to the south, including within the region known for the 14-month periodicity. 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 previously documented still holds true. We invert estimated GPS offsets for the largest 14 events using non-negative thrust faulting along a plate interface divided into roughly 500 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 the events resolved to date is just under 500 km along strike, and maximum duration is seven weeks, which lies in marked contrast to other subduction zones. Averaged over many ETS events, the upper limit of transient slip in the vicinity of Seattle, WA lies just west of the heavily urbanized Puget Sound region, suggesting that the lower limit of megathrust seismic rupture may extend much closer to this area than previously thought. During the time window of 2005-2007.2, a systematic quantification of seismic tremor observed both during and outside of ETS events has demonstrated a very linear relationship between tremor duration and inverted GPS equivalent moment. Moreover, no GPS transients during this time period are observed in the absence of tremor. This suggests that the GPS-measured transient deformation may be the integrated near-field offsets resulting from many small, discrete seismic slip events (meaning aseismic, transient creep is not needed to explain the GPS signal). This hypothesis is supported from inferences about tremor source characteristics drawn from borehole seismic spectral features and tremor locations determined from surface arrays: the observed quantity, timing and location of tremor largely explains the ETS signal as observed on surface and borehole strain- and tiltmeter arrays. Together, these data suggest that Cascadia, and presumably other subduction zone ETS events, are comprised largely of a form of microseismic clustering unique to subduction zones and may in fact have little aseismic creep associated with them.
T11F-03
Non-Volcanic Tremor Observed in Guerrero, Mexico
Non-volcanic tremor (NVT) activity is clearly identified as episodes of higher spectral amplitude in the range of 1-8 Hz in daily spectrograms from the continuous records at broad band seismic stations in the Guerrero state, Mexico. The analyzed data cover a period of 2001-2007 when in 2001-2002 a large slow slip event (SSE) had occurred in the Guerrero-Oaxaca region, and then followed by a steady-state interseismic epoch of 2003-2005 and a new large SSE occurred in 2006. The tremor signals in Mexico are very similar to those obtained in Cascadia subduction zone. An average tremor burst remains for 10-60 min and is dominated by S-waves. More than 100 strong NVT bursts were recorded by the most of the Meso-American Subduction Experiment (MASE 2005-2007) seismic stations with the majority of epicenters clustered in the narrow band of ~40 x 150 km2 to the south of Iguala city and in parallel to the coastline. Depths of NVT hypocenters are poorly constrained but mostly scattered in the continental crust between 5 and 45 km depth. Tremor activity is significantly higher during the 2001-2002 and 2006 SSE compared with that for the "quiet" period of 2003-2005. Very low NVT occurrence for the period of about 2 months right after the SSE is apparent in 2002 and 2006. Similar to the Japan and Cascadia subduction zones, the main tremor activity in Mexico develops in the area close to the mantle wedge. In Mexico it is located at approximately 200 km inland from the trench. While conductivity pattern obtained from the magnetotelluric profile in Guerrero does not correlate directly with the NVT distribution, gravity and magnetic anomalies modeling favors a hypothesis that the NVT is apparently related to the dehydration and serpentinization processes.
T11F-04 INVITED
Triggered Microearthquakes but no Tremor Accompanying Slow Slip in the Hikurangi Subduction Zone, New Zealand
This is the first study to systematically examine continuous broadband seismic data spanning slow slip events in New Zealand. We have reviewed a total of 20 weeks of seismic data spanning three slow slip events: two in the shallow region of the Hikurangi subduction zone near Gisborne (in 2004 and 2006) and one deeper, ~18 month-long event beneath the Manawatu region (early 2004 to mid-2005). We have not detected seismic tremor similar to that seen elsewhere during any of these slow slip events. This observation does not appear to be the result of network limitations: we deployed five additional seismographs during the latter stages of the 2006 Gisborne event to augment the permanent network and still did not detect seismic tremor. However, our analysis has revealed a pronounced increase in ML 1--2 microseismicity during the 2004 Gisborne event that is spatially,/em> restricted to a region of the subducting plate downdip from the slow slip patch inferred from GPS observations and temporally restricted to the period of slow slip. This increased rate of local seismicity is not evident in the catalog produced by routine analysis and was only detected by a methodical review of the continuous waveform data. These subduction zone microearthquakes have similar spatiotemporal characteristics to the "coshocks" reported by Segall et al. (2006) in an intraplate setting.
T11F-05
Observing Episodic Slow Slip with PBO Borehole Strainmeters along the Northern Cascadia Margin
Episodic non-volcanic low-frequency seismic tremor and slow slip (ETS) downdip of the megathrust seismogenic zone are now routinely observed at the Nankai and Cascadia subduction zones. These phenomena yield important information about the mechanism of the downdip termination of the seismogenic zone and raise intriguing questions regarding the process of strain accumulation and release during time periods between great megathrust earthquakes. Borehole strainmeters (BSM"s), newly installed along the Cascadia margin under the U.S. Plate Boundary Observatory (PBO) initiative, have enabled us to study ETS with much improved detection sensitivity. Here we report preliminary results to demonstrate the value of the BSM's in helping us define the temporal character of prolonged ETS events and the spatial character of shorter-lived ETS events. In the region from Puget Sound to southern Vancouver Island, rather regularly occurring, prolonged (> 6 days) ETS events affect large areas, spanning up to a few hundred kilometres along the strike of the subduction zone. The ETS event of September 2005, the first one to be monitored with BSM observations as well as GPS and tremor observations, started from northern Puget Sound, migrated northwest along the margin, and stopped in southern Vancouver Island. BSM's in north-western Washington detected significant changes in strain components during this event and their temporal character is consistent with a dislocation model of a slow slip on the deeper plate interface propagating northwest at a constant rate of about 7 km/day. Between these regularly occurring prolonged ETS events, we have also observed sporadic tremor activity with shorter durations. Whether these small tremor events are also accompanied by slow slip and the characteristics of such slip, if present, are largely unknown because of their small geodetic signal. A 4-day episode of seismic tremor in central Vancouver Island in November 2006 resulted in ill-defined surface displacements at the closest continuous GPS stations. However, strain data from a borehole strain meter (B012) at Ucluelet, located on the west coast of central Vancouver Island, are consistent with slip on a small patch of the deeper subducting plate interface to the northeast of this site, the region where the tremors were observed. Simple elastic dislocation models with 3 to 5 cm of slip constrained to the subducting plate interface can replicate the observed strain signal as well as the limited surface displacements.
T11F-06
A Time-Dependent Inversion Analysis for Short-Term Slow Slip Events in Southwest Japan Based on the NIED Hi-net Tiltmeter Data
In the deeper part of the southwest Japan subduction zone, a wide variety of "slow earthquakes," including low-frequency tremors (Obara, 2002), short-term slow slip events (SSE; Obara et al., 2004), and very low-frequency earthquakes (Ito et al., 2007), have occurred coincident with each other. Since SSEs have the maximum size (Mw~~ 6.0) and duration (several days) among them, the SSEs may control the overall characteristics of the coincident episodes, such as a recurrence rate, duration, and a migration velocity of the source region. In this context, it is important to know the source processes of the SSEs in order to understand not only the repeating nature of the coincident episodes, but also the generation cycles of megathrust earthquakes. We have applied a simple model that assumes a single rectangular fault to model tilt deformations due to the SSEs observed by the NIED Hi-net tiltmeter network (e.g., Hirose and Obara, 2005). This analysis is intended for estimating an approximate position and size of each event. The typical area of the estimated SSE faults has been 50~× 30~km2. In practice, however, the observed tilt data show different time variations at different stations, or even for different directions at a particular station, during an episode. These facts indicate that the tilt records are capable of determining more detailed source processes for the SSEs. In this study, we developed a time-dependent inversion code that is applicable to the tiltmeter data, based on the Kalman filter approach (e.g., Segall and Matthews, 1997). We demonstrate that this method can resolve the slip histories for a number of SSEs in the western Shikoku region, southwest Japan. For the April 2006 western Shikoku episode, we assumed 5~× 3 subfaults with an area of 20~× 20~km2 each, and the slip history was resolved as follows: the slip started in the base of the Sadamisaki Peninsula on April 16th, and the slip area extended to the east direction for the following four days. The temporal change in the estimated slip area corresponds roughly to the spatio-temporal distribution of the tremors in this episode. This result shows a closer spatial correspondence between the tremors and the SSE than ever indicated, suggesting that they have occurred at almost the same position.
T11F-07
Spatio-temporal distribution of energy radiation from low frequency tremor
Recent fine-scale hypocenter locations of low frequency tremors (LFTs) estimated by cross-correlation technique (Shelly et al. 2006; Maeda et al. 2006) and new finding of very low frequency earthquake (Ito et al. 2007) suggest that these slow events occur at the plate boundary associated with slow slip events (Obara and Hirose, 2006). However, the number of tremor detected by above technique is limited since continuous tremor waveforms are too complicated. Although an envelope correlation method (ECM) (Obara, 2002) enables us to locate epicenters of LFT without arrival time picks, however, ECM fails to locate LFTs precisely especially on the most active stage of tremor activity because of the low-correlation of envelope amplitude. To reveal total energy release of LFT, here we propose a new method for estimating the location of LFTs together with radiated energy from the tremor source by using envelope amplitude. The tremor amplitude observed at NIED Hi-net stations in western Shikoku simply decays in proportion to the reciprocal of the source-receiver distance after the correction of site- amplification factor even though the phases of the tremor are very complicated. So, we model the observed mean square envelope amplitude by time-dependent energy radiation with geometrical spreading factor. In the model, we do not have origin time of the tremor since we assume that the source of the tremor continuously radiates the energy. Travel-time differences between stations estimated by the ECM technique also incorporated in our locating algorithm together with the amplitude information. Three-component 1-hour Hi-net velocity continuous waveforms with a pass-band of 2-10 Hz are used for the inversion after the correction of site amplification factors at each station estimated by coda normalization method (Takahashi et al. 2005) applied to normal earthquakes in the region. The source location and energy are estimated by applying least square inversion to the 1-min window iteratively. As a first application of our method, we estimated the spatio-temporal distribution of energy radiation for 2006 May episodic tremor and slip event occurred in western Shikoku, Japan, region. Tremor location and their radiated energy are estimated for every 1 minute. We counted the number of located LFTs and summed up their total energy at each grid having 0.05-degree spacing at each day to figure out the spatio-temporal distribution of energy release of tremors. The resultant spatial distribution of radiated energy is concentrated at a specific region. Additionally, we see the daily change of released energy, both of location and amount, which corresponds to the migration of tremor activity. The spatio-temporal distribution of energy radiation of tremors is in good agreement with a spatio-temporal slip distribution of slow slip event estimated from Hi-net tiltmeter record (Hirose et al. 2007). This suggests that small continuous tremors occur associated with a rupture process of slow slip.
T11F-08
Non-Volcanic Tremor in South-Central Alaska and its Relation to the 1998-2000 Slow-Slip Event
The discovery, less than a decade ago, of non-volcanic tremor and slow slip events in the Cascadia subduction zone and in the Nankai trough, has led to an investigation of the phenomena at many tectonic plate margins. It now looks likely that slow-slip events and non-volcanic tremor are present in many, if not all subduction zones, in addition to other types of faults. Slow-slip events most likely play an important role in the seismogenic processes at subduction zones, by either loading locked portions of the interface, and/or perhaps in triggering major earthquakes on the plate boundary. Non-volcanic tremor appears to be a closely related phenomenon, which might be used as an indicator of aseismic creep, or changes in fluid or pore pressure, and perhaps shed light on the processes active in the deeper interplate transition zone. Here we examine non-volcanic tremor in south- central Alaska occurring simultaneously with a large slow-slip event that occurred in the region between 1998- 2000. We tabulate the observed non-volcanic tremor events in terms of durations, locations (when possible), and maximum amplitudes for three months of each year between 1998 and 2001. The majority of the non-volcanic tremor signals observed after the onset of the slow-slip event in March 1998 are bursts lasting between 10 and 15 minutes with frequencies ranging from 1-10 Hz. Similar signals lasting up to a day are also observed, but less frequently. By the summer of 2001 the quantity of non-volcanic tremor decreases dramatically from the previous summer. Episodes rarely last up to 10 minutes and maximum amplitudes are lower than previous years. The decrease in non-volcanic tremor activity in the summer following the end of the slow-slip event is the first evidence that the relationship between slow-slip events and non-volcanic tremor seen in Cascadia and southwest Japan may also exist in central Alaska.