By Ross Stein, Ph.D, Temblor
Check your hazard rank
In September 2016, Wenyuan Fan and Peter Shearer, from the Scripps Institution of Oceanography at U.C. San Diego, published an astonishing discovery in Science. Fan and Shearer detected nearly 50 previously unidentified M≥5.5 aftershocks up to 300 kilometers (200 mi) away from their M=7-8 mainshocks during the brief passage of the seismic ‘surface’ waves from the mainshock, or within 3 minutes. The authors concluded that remote dynamic triggering commonly exists and fundamentally promotes aftershock occurrence throughout the globe.
How could these large aftershocks have been overlooked?
There are no large, immediate, and remote aftershocks in any seismic catalog, and so most of us had concluded that this must be exceedingly rare, or is non-existent. That’s because today, any M≥4.5 shock anywhere on Earth can be reliably detected.
But, during the first few minutes after a large mainshock, its seismic wave train could obscure aftershocks, and so during this period, large shocks could conceivably have been hidden. To overcome this obstacle, Fan and Shearer used a technique in which a continent’s worth of seismometers are retroactively trained on the site of a single large earthquake halfway around the globe, and then used to track where the seismic energy was released in time. The method, called ‘beam back-projection,’ was introduced by Miaki Ishii, Peter Shearer, Heidi Houston and John Vidale in 2005 (Ishii et al., 2005). Although remote dynamic triggering of tiny aftershocks is well known (Velasco et al., 2008; Parsons et al., 2014), there are only a few examples of M≥5.5 aftershocks (Johnson et al., 2015), the most impressive of which was the 2012 M=8.6 Indian Ocean shock, which triggered large aftershocks all over the globe (Pollitz et al., 2012). But those aftershocks struck over several days—not minutes—long after the seismic waves had vanished.
Implications of the Fan-Shearer hypothesis
If they are correct, the hazard after a large mainshock would be more widespread than understood today, and the first several minutes after a large mainshock are more dangerous than we currently assume. But there is another, equally important, implication: For historical quakes, whose magnitudes and locations are interpreted from contemporary intensity reports (descriptions of shaking), we might be overestimating their magnitudes and blurring their locations, because widespread shaking in aftershocks would be misconstrued as caused by the mainshock.
A Challenge by Yue et al.
In October 2017, Han Yue, from Beijing University, Jorge C. Castellanos, Chunquan Yu, and Lingsen Meng from UCLA and Zhongwen Zhan from Caltech published a rebuttal of the Fan-Shearer hypothesis in Geophysical Research Letters. In a nutshell, Yue et al. argue that the seismic signals interpreted by Fan and Shearer as aftershocks are actually water reverberations in near-shore regions. The reverberations are triggered by the seismic waves launched by the mainshock. Fan and Shearer had raised this possibility in their paper, but ultimately dismissed it. Yue et al. present a series of falsification tests, but I am going to focus on what I consider the two most persuasive.
In the figure above, energy from a M=7.2 mainshock southwest of the trench is imaged by P waves. The energy is spread over about 100 km because this is roughly the rupture area of the shock. There are also strong energy pulses landward of the trench, near the ‘10 s resonance contour.’ These are the pulses identified by Fan and Shearer as aftershocks. But Yue et al. point out that if these were aftershocks, they should also appear when using S waves. But they are absent in the right-hand panel above. If, instead, they were water reverberations, they should appear in the P wave panel but not in the S wave panel, because S waves do not transmit in water. So, this would seem to be a very strong test, which the Fan-Shearer hypothesis does not pass.
Singing seismograms
In a second falsification test, Yue et al lined up seismograms of the M=7.2 mainshock recorded throughout the hemisphere. It takes about 25 s for a M=7.2 earthquake to rupture, and in those first 20-30 seconds, one sees the somewhat chaotic signature of the rupture. But beginning at 61 s (and perhaps at 51 s) one can see a coherent pulse on all the records (the red ‘+’ signs in the figure below). This pattern repeats at least three times at 10 s intervals (green, blue, and cyan ‘+’ signs in the figure below).
Yue et al argue that earthquakes do not produce such simple and periodic pulses. When Yue et al. located the source of the pulses, they land next to the beam back-projection energy pulses that Fan and Shearer identified as aftershocks. In addition, the pulses are very close to the seafloor depth contour that would produce the observed 10 s resonance. So, it would be hard to argue that water reverberation was not occurring, and occurring right where Fan and Shearer identified aftershocks.
Dueling posters at the Fall Meeting of the American Geophysical Union
Fortunately, Wenyuan Fan (now a Post-Doctoral Scholar at Woods Hole Oceanographic Institution) and Han Yue presented side-by-side posters at AGU Meeting in New Orleans two weeks ago. This gave everyone the chance to see both sides of the story, and it also enabled me to pose questions to each author based on the arguments and rebuttals of the other.
Fan and Shearer believe that with more tuning, an aftershock energy pulse might emerge in the S wave analysis. They now concede that water reverberations are evident in the signals, but they argue that these are water reverberations from the remote aftershock, not the mainshock. So, while remote dynamic aftershocks might be less common than they originally proposed, it still occurs. Han Yue says that he cannot (yet) eliminate this possibility, and so the debate continues.
So, who’s right?
In addition to the falsification tests, two other factors lead me to believe that Yue et al. are likely correct, and that few if any of the signals are actually aftershocks. Why would different types of mainshocks (thrust, extensional, and strike-slip) all trigger aftershocks at about the same water depth between the trench and the coast? This just seems very unlikely. Beyond that, if there is a simpler, quotidian explanation for a phenomenon (water reverberation), then it should be favored over a more exotic interpretation (heretofore unseen aftershocks).
With that said, debates like this are essential to science, which only advances when bold new ideas are promulgated, and promulgated in a manner that can be unambiguously tested. And for that we can thank Fan and Shearer. All we can really do in science is falsify hypotheses; proving something right is extremely difficult.
Here is a video of seismic wave propagation through rock and water by Yue et al., 2017
The first ring-like wave launched from the 20-km deep hypocenter is the P wave, traveling at about 7 km/sec; the second is S wave, traveling at about half that speed. The video is moving at about realtime. The thick black line is the seafloor. At the site of the epicenter, the seafloor is about 5 km (3 mi) deep. The camera moves with the advancing waves toward the coast. Water reverberations become most pronounced when the seafloor shallows to about 2 km deep, at a distance of about 220 km. The P waves bounce back and forth every 10 s or so in the water.
References
Wenyuan Fan and Peter M. Shearer (2016), Local near instantaneously dynamically triggered aftershocks of large earthquakes, Science, 353, 1133-1136, DOI: 10.1126/science.aag0013.
Miaki Ishii, Peter M. Shearer, Heidi Houston, and John E. Vidale, Extent, duration and speed of the 2004 Sumatra-Andaman earthquake imaged by the HI-Net array (2005), Nature, DOI: 10.1038/nature03675.
Christopher W. Johnson, R. Bürgmann, and F. F. Pollitz (2015), Rare dynamic triggering of remote M≥ 5.5 earthquakes from global catalog analysis, J. Geophys. Res., 120, 1748–1761, doi:10.1002/ 2014JB011788.
Tom Parsons, Margaret Segou, Warner Marzocchi (2014), The global aftershock zone, Tectonophysics, 618, 1–34, .doi.org/10.1016/j.tecto.2014.01.038
Fred F. Pollitz, Ross S. Stein, Volkan Sevilgen, and Roland Bürgmann (2012), The 11 April 2012 east Indian Ocean earthquake triggered large aftershocks worldwide, Nature, 490, 250–253, DOI:10.1038/nature11504.
Aaron A. Velasco, S. Hernandez, T. Parsons, and K. Pankow (2008), Global ubiquity of dynamic earthquake triggering, Nature Geoscience, 1, 375–379, doi:10.1038/ngeo204
Han Yue, Jorge C. Castellanos, Chunquan Yu, Lingsen Meng, and Zhongwen Zhan (2017), Localized water reverberation phases and its impact on backprojection images, Geophys. Res. Letts., DOI: 10.1002/2017GL073254.