(back to Carl Tape home)

Earthquake nucleation and triggering on an optimally oriented fault (pdf, supplement, report)
Earth and Planetary Science Letters, v. 363, p. 231-241, 2013
Carl Tape, Michael West, Vipul Silwal, Natasha Ruppert

Seismic surface waves from large, distant earthquakes commonly trigger smaller earthquakes. However, delay times of hours to days between the surface waves and the triggered earthquakes weaken the causal connection. Furthermore, when there is no delay, the triggered earthquakes are typically too small or too obscured to obtain reliable source mechanisms. We present observations of instantaneous triggering of a strike-slip earthquake in central Alaska. Shear strain from the optimally-aligned teleseismic Love wave induced a 24 s exponential foreshock signal leading to the triggered earthquake. This nucleation phase, and the alignment of the triggered earthquake source mechanism with the teleseismic stress field, reveal the behavior of an existing fault under well-calibrated strain conditions. The Alaska earthquake provides the first observation of combined nucleation and triggering, and it suggests that transient stresses during nucleation may influence the subsequent earthquake rupture. Laboratory and theoretical studies of nucleation and triggering may help discriminate between different interpretations for the Alaska earthquake.

This material is based upon work supported by the National Science Foundation under Grant Number (NSF EAR 1215959).

Love waves in Alaska from the April 11, 2012 Mw 8.6 Sumatra earthquake. (a) Propagation path from Sumatra to central Alaska, at an azimuthal angle of  22.3 deg. The beachball symbol depicts the predominantly strike-slip rupture of the Sumatra earthquake. (b) Horizontal displacement field in Alaska at the origin time of the Mw 3.9 Nenana earthquake in central Alaska. The epicenter of this event is marked by the white dot at the center of the thick gray lines. The large gray arrow denotes the wave propagation direction from Sumatra. (c) Expanded view of (b) in the region of the triggered event. Major active faults are labeled as Denali, Tintina, and MFSZ, the Minto Flats seismic zone. (d) Shear strain computed from the displacement field in (b).

The occurrence of the Mw 3.9 earthquake near Nenana, Alaska, coincident in time and space with 4 cm horizontal displacement Love wave from the Sumatra earthquake.
(a) Representative three-component ground motion of the Sumatra earthquake in Alaska. This recording at MDM is the closest station (32 km) to the triggered Nenana earthquake. The Love wave is the dominant waveform throughout Alaska. (b) High-frequency filtered version of the same seismogram in (a) showing the Mw 3.9 Nenana earthquake. (c)-(d) Transverse displacement (c) and corresponding shear strain (d) at the Nenana epicenter, computed by stacking waveforms from the 13 closest stations. The dashed black lines denote the positive shear strain pulses A and C. The dashed red line denotes the emergence of the foreshock signal from the noise at t = -24 s.

Source inversion for the Nenana earthquake, showing a subset of observed three-component seismograms (black) compared with synthetic seismograms (red), both filtered between 0.3 Hz and 0.6 Hz.

Earthquake source mechanism for the 2012-04-11 Mw 3.9 Nenana earthquake in comparison with the orientation of Sumatra Love waves. The comparison is made at the Nenana earthquake epicenter at the origin time of the Nenana earthquake. The Sumatra Love waves propagate along the great-circle path from Sumatra (solid cyan line) and impart horizontal ground motion in the perpendicular direction (dashed cyan line). The principal axes of the associated strain tensor are shown as cyan arrows. The principal axes for the Nenana source mechanism are plotted as red arrows and white arrows. Inward-pointing arrowheads for P-axis directions are hidden beneath the beachball. The 1995-10-06 Mw 6.0 earthquake is plotted for comparison. Both earthquakes are consistent with left-lateral strike slip fault motion on the western of the two faults comprising the Minto Flats seismic zone.

Record section of the 24 s foreshock signal prior to the Nenana earthquake. The seismograms are the vertical component of velocity, filtered 2-8 Hz, ordered by distance from the Nenana epicenter, and aligned on the P onset of the mainshock. The red line at -24 s is the time at which the foreshock signal rises above the pre-Sumatra background noise level. The alignment of this pulse, in addition to the decrease in amplitude with distance, indicate that the foreshock signal is in the same region as the mainshock.

Envelope of velocity seismograms for two representative stations. The amplitudes are plotted on a log scale so that the noise level, foreshock signal, and mainshock are all visible. The linear slope of the foreshock signal represents an exponential growth in amplitude. MLY.AK is 86 km northwest of the Nenana epicenter; COLA.IU is 51 km east of the Nenana epicenter.
Comparison between the 2012-04-11 Mw 3.9 and 2004-11-17 Mw 3.6earthquakes, recorded at COLA.IU. The 2004 event has similar hypocenter, magnitude, and source mechanism to the 2012 event. The striking difference between these two records is the 2012 foreshock signal, interpreted as a nucleation phase.

Guide to interpretation scenarios for the Nenana earthquake. (bottom) The example time series is the MLY record (see above). The vertical line at t = -72.9 s is the peak positive shear stress perturbation from the Sumatra Love wave. (top) Shear stress perturbation due to the Sumatra Love waves, in kilopascals.

Key times and durations in this study. Station MDM is the closest station (32 km) to the triggered Nenana earthquake. By "at Nenana," we mean at the epicenter of the Nenana earthquake.

Tectonic Setting: Minto Flats seismic zone, central Alaska

Figure S4 from Tape et al. (2013). Figure S4: Inferring two faults from seismicity in the Minto Flats seismic zone. (a) Seismicity from the AEIC catalog, 19902010, M > 0. (b) Seismicity density from (a), plotted on a logarithmic color scale, which reveals the dominant two fault strands that we digitized. Other faults plotted are from (Plafker et al., 1994). The Minto Flats seismic zone (MFSZ) has previously been described as a single, diffuse seismic zone that extends approximately 200 km from SW to NE (Pulpan, 1986; Biswas and Tytgat , 1988; Page et al., 1995; Ratchkovski and Hansen, 2002). Despite the prominence of the two seismic lineaments, neither lineament appears as an active or inactive fault in the fault map of Plafker et al. (1994). The eastern lineament coincides with the Minto fault mapped by Pewe et al. (1966), but later interpreted as geomorphic feature (Page et al., 1991; Plafker et al., 1994). Here we suggest that the lineaments be identified as two active faults. The two faults of the MFSZ appear to bound the gravity-low signature of the Nenana sedimentary basin (Barnes, 1961; Van Kooten et al., 2012).

Text files containing the longitude and latitude coordinates for the two MFSZ fault strands: delimited (west only, east only).