By E. M Russo, Susan L. Beck, Steve W. Roecker, Angela M. Reusch, Aaron Velasco, Carl Ebling, and Paul M. Bremner - Winter 2011
The global seismology community established a new model of international cooperation in response to the great Chile earthquake of February 27, 2010. The adaptable framework is geared to produce a high-quality, open-access data set of seismic waveforms collected in the wake of great earthquakes. Collection of the unprecedented Maule International Seismic Dataset (MISD) was supported by an NSF RAPID award and used fortuitously available EarthScope Flex Array (FA) sensors. IRIS PASSCAL personnel and volunteers from the IRIS Community deployed and maintained 58 broadband seismic stations from April-September 2010. The FA sensors constituted the IRIS CHAMP (for Chile RAMP) network, the largest equipment contribution to the MISD. Three CHAMP stations transmitted continuous data to PASSCAL in Socorro, NM – the first time data have been telemetered from a temporary deployment outside of North America.
When great earthquakes strike, will we be ready to exploit all they can potentially tell us about the processes occurring at the end of the seismic cycle? Although the great 2004 Sumatra earthquake caught Earth scientists largely unprepared, when the Mw 8.8 February 27, 2010, Maule earthquake struck – thanks in large part to a quick decision to use available EarthScope equipment, and to a flurry of negotiations between national groups responding to the crisis – the global seismic community did everything possible to collect unique seismic and geodetic data as they accrued following the event.The Maule earthquake damaged central Chile to a degree commensurate with its magnitude, mitigated by careful preparation by Chilean civil defense authorities and scientists. Rapid repair and restoration of critical infrastructure allowed seismologists from Germany, France, the UK, and the US, working with seismologists in Chile, to coordinate sensor deployment, thus capturing aftershocks and other signals associated with this significant earthquake. In all, 91 broadband, 48 short period (1 Hz), and 25 accelerometer stations were installed. The CHAMP sites were deployed by the end of March and decommissioned in late-September 2010 to return equipment to a scheduled FA project. The six-month CHAMP dataset – all freely available from the IRIS Data Management Center (network code XY) – provides an unprecedented view of end-seismic cycle processes, including a detailed look at seismic tremor in the wake of a great earthquake.
Open data: a new model for international collaboration in aftershock studies. One of the most important policies implemented at the inception of EarthScope established the immediate open access to data recorded at the EarthScope facilities. Recognition that open access to all data collected following the Maule earthquake would ensure both its maximum and timely use spurred IRIS leadership to negotiate an open data access agreement with other international groups deploying instruments. Cooperation allows for the most efficient use of equipment, obviating redundant network designs. The MISD will therefore contain data from a relatively uniform network covering virtually the entire on-land rupture region, some 90,000 km2 , at a nominal station spacing of ~30 km (Figure 1).
Uniform coverage of the Maule rupture region allows us to address some interesting questions. Episodic tremor, slow slip events, and low frequency earthquakes (LFEs) have been observed at the down-dip edge of megathrust rupture zones in Japan, Cascadia, Costa Rica, Mexico and southern Chile, revealing a strong correlation between tremor and slow, aseismic backslip of the forearc region that relieves stress on the deeper portions of the interplate interface. Improvements in tremor locations have shown that tremors occur on or very near the interplate interface in both Japan and Cascadia, and are part of a continuum of shear slip processes occurring at high pore-fluid pressures that include aseismic slip, slip producing tremor, LFEs and seismic slip. However the relationship between episodic tremor and megathust earthquakes is not understood. Is tremor excited by megathrust ruptures, or is it suppressed or forced to migrate elsewhere in the system? Does tremor occur only along the boundaries of the Maule slipped region, as has been suggested for other subduction zones? If nonvolcanic tremor in the Maule rupture zone region conforms to type, we expect to detail regions of slower slip that produce LFEs and tremor along the interplate interface. Identification of such regions may have implications for slip acceleration during rupture near the down-dip seismogenic limit.
Initial analysis of MISD data provides observations of LFEs (Figure 2), and indications of nonvolcanic tremor activity (Figure 3) in central Chile (~33-38°S). These are the first observations ever of tremor in the wake of a mega-thrust earthquake. A major challenge to detect tremor in the Maule rupture zone is the extreme frequency of high-amplitude “normal” aftershocks (Figure 3), which make identification of distinct tremor wave packets difficult. A preliminary search for tremor in the frequency domain shows significant power in the characteristic 2-10 Hz tremor frequency band. Signals are stronger on the horizontal components consistent with propagation of tremor as shear waves, as observed elsewhere. Tremor signals occur at nearly all the CHAMP stations (Figure 3) and have persisted during the entire recording period. Although we have not yet located these tremors and are therefore unable to say whether they occurred on the fault that slipped during the Maule earthquake, their existence indicates that tremor is not suppressed following great earthquakes. This is unexpected since seismic slip during the earthquake certainly relieves built-up interplate stress, which, in classic elastic-rebound models, must accumulate again before episodic slow slip and tremor again begin to operate. Given this observation, a more likely scenario is that tremor is currently generated during the slow, long-duration aseismic slip on the rupture interface that has been observed to follow other great earthquakes.
Ongoing and future studies of the MISD will certainly take us further towards resolving the presence of tremor and other issues concerning end-seismic cycle processes. Thanks to EarthScope, IRIS and PASSCAL, and to our international collaborators, the global seismological community will have the opportunity to attack these problems with an unprecedented data set in hand.