Ridgecrest earthquake shut down cross-fault aftershocks

By Shinji Toda, Ph.D., IRIDeS, Tohoku University, and Ross Stein, Ph.D., Temblor, Inc.
 
The Magnitude 7.1 earthquake abruptly halted aftershocks on the M 6.4 cross fault. It also produced a far-flung aftershock sequence that touched the San Andreas, but refused to cross the Garlock Fault.
 
Citation: Shinji Toda, and Ross Stein (2019), Ridgecrest earthquake shut down cross-fault aftershocks, Temblor, http://doi.org/10.32858/temblor.043
 

Shutdown

 

A remarkable illustration of how stress controls seismicity was handed to us by the Ridgecrest earthquake sequence. Two perpendicular faults ruptured in the M 6.4 earthquake on July 4 (left panel below), both lighting up in aftershocks. But when the M 7.1 struck 34 hours later, seismicity along the cross fault abruptly stopped (middle panel below). Why?

 

Maps of seismicity and calculated Coulomb stress change for the Ridgecrest sequence. (The calculation is for ‘optimally-oriented’ strike-slip faults with friction of 0.4 at 10 km depth.)
Maps of seismicity and calculated Coulomb stress change for the Ridgecrest sequence. (The calculation is for ‘optimally-oriented’ strike-slip faults with friction of 0.4 at 10 km depth.)

 

The answer seems to be that the M 7.1 rupture cast a ‘stress shadow’ over the cross fault, inhibiting failure by several bars, and so shutting aftershocks off. In our calculation of the stress imparted by the M 7.1 (right panel above), the cross fault is ‘shadowed,’ (blue zones) while the stress increased to the northwest (red zones), to southeast extending across the Garlock Fault. The isolated side lobes also experienced aftershocks to the M 7.1 event. The seismicity shutdown along the cross fault is dramatic when viewed as a time series for M≥1 shocks (below), with the seismicity rate dropping by a factor of 10.

 

In this time series of seismicity for M≥1 earthquakes, each shock is a green ‘lollipop,’ and the cumulative number of shocks is shown by the blue line. Following the M 7.1, aftershocks on the cross fault didn’t get smaller, they just occurred less frequently.
In this time series of seismicity for M≥1 earthquakes, each shock is a green ‘lollipop,’ and the cumulative number of shocks is shown by the blue line. Following the M 7.1, aftershocks on the cross fault didn’t get smaller, they just occurred less frequently.

 

A few bars are not a lot of stress. In the photo below, Ross puts a half a bar of Coulomb stress across the base of his palms by pushing as hard as he can.

 

Ross Stein demoing Coulomb stress. (Photo by Dr. Bob Sabol, United Airlines passenger)
Ross Stein demoing Coulomb stress. (Photo by Dr. Bob Sabol, United Airlines passenger)

 

Stress Shadow

 

Seismic shaking can promote seismicity but not inhibit it, so the shutdown illustrates that small decreases in Coulomb stress can inhibit earthquakes, a phenomenon first discovered by Ken Hudnut, Leonardo Seeber and Javier Pacheco (Hudnut et al., 1989). They found that the 1987 Superstition Hills earthquakes, a M 6.2 followed by a M 6.6 some 11 hours later, also shut down seismicity on a cross fault. This phenomenon was seen again in 1997 Kagoshima couplet in Japan, adjacent M 6 shocks 48 days apart (Toda and Stein, 2003), and in the 1992 M 7.3 Landers shock, which shadowed part of the M 6.1 Joshua Tree aftershock zone that had struck 66 days beforehand (Toda et al., 2012). The Joshua Tree shutdown began several days after the Landers shock, the Kagoshima shutdown began within about a day, but at Ridgecrest, it was immediate. Stress shadows were named by Ruth Harris and Robert Simpson of the USGS, which they argued explained the paucity of large shocks in the century or so after the great 1857 in southern California (Harris and Simpson, 1996), and after the great 1906 quake in northern California (Harris and Simpson, 1998). Stress shadows were also hypothesized to explain the shutdown of the southern Hayward Fault after the 1989 M 6.9 Loma Prieta shock (Simpson and Reasenberg, 1994; Stein, 1999).

 

Remote Shocks

 

If one steps back from the cross fault, one sees what appear to be aftershocks as much as 150 km (90 mi) from the M 7.1 epicenter in three of the four stress trigger (red and yellow) lobes in the panel at left below, and these appear to correspond to the stress trigger lobes. But are these aftershocks, or just shocks that would have occurred in those remote locations anyway?

 

Calculated Coulomb stress change (along optimally-oriented strike slip faults at 10 km depth with friction of 0.4) is shown at left. Seismicity rate change after the M 7.1 quake is shown at right, in which warm colors indicate rate increases (aftershocks), and turquoise colors indicate rate decreases. Rate decreases are difficult to detect in the weeks after the mainshock.
Calculated Coulomb stress change (along optimally-oriented strike slip faults at 10 km depth with friction of 0.4) is shown at left. Seismicity rate change after the M 7.1 quake is shown at right, in which warm colors indicate rate increases (aftershocks), and turquoise colors indicate rate decreases. Rate decreases are difficult to detect in the weeks after the mainshock.

 

We can answer that question by plotting the change in the seismicity rate, comparing the first 11 days after the M 7.1 to the preceding year (above right). Here, the warm colors are places where the seismicity rate increased after the M 7.1, light blue areas are where it decreased. Now, all four stress trigger lobes appear to correspond to seismicity rate increases (e.g., aftershocks), and they extend all the way to the San Andreas Fault at Tejon Pass and Cajon Pass.

 

San Andreas in play?

 

What do the subtle, or perhaps apparent, seismicity increases at the two big bends on the San Andreas Fault mean for future great quakes? We do not know. We can say only that the effects, while promoting failure, are very small, about the same as the twice-daily tidal stresses.

 

The Garlock Wall

 

But if the shutdown and remote shocks can be explained by stress transfer, another feature of the aftershocks is mysterious. If you glance again at the Coulomb stress lobes, you can see that the stress trigger zone extends well to the south of the Garlock Fault, but the M 7.1 aftershocks do not. That is even clearer in the map below: The Garlock is a wall.

 

Neither the preceding decade of quakes (in blue) nor the Ridgecrest aftershocks (in red) cross the Garlock Fault, but aftershocks extend outward in most other directions.
Neither the preceding decade of quakes (in blue) nor the Ridgecrest aftershocks (in red) cross the Garlock Fault, but aftershocks extend outward in most other directions.

 

Why would a fault be a barrier to aftershocks? We can think of two explanations, but we can’t prove either. The first is that Coulomb stress changes amplify the background seismicity rate (this follows from the theory of rate/state friction developed by Jim Dieterich at the USGS; Dieterich, 1994). If the background rate is high (those are where there are many blue shocks above), then the seismicity will be very responsive to the stress changes. That seems to be what happened at the Coso Volcanic Field, which was active before the M 7.1, was stressed by the M 7.1, and produced abundant aftershocks. But if its near-zero, as it appears to be south of the Garlock, the stress changes have almost no effect.

 

Even if that explanation were correct, that would not explain why the northern Mojave Desert, south of the Garlock Fault, was so quiet in the first place? Perhaps because it is composed of different rocks than the material to the north. The Garlock has slipped a total of ~100 km (60 mi), juxtaposing different chunks of crust. To the north lies mostly granitic rocks, and to the south lies mostly volcanic rocks (see the map below). Somehow, rock type might modulate seismicity.

 

Geologic map of the Ridgecrest area (Jennings et al., 1962), labeled by type of bedrock. The yellow veneer of sediments blanket some of the bedrock. Aftershocks of the Ridgecrest sequence are red; the preceding decade of quakes are blue.
Geologic map of the Ridgecrest area (Jennings et al., 1962), labeled by type of bedrock. The yellow veneer of sediments blanket some of the bedrock. Aftershocks of the Ridgecrest sequence are red; the preceding decade of quakes are blue.

 

So, what did we learn?

 

The Ridgecrest sequence demonstrates how Coulomb stress controls seismicity. Ridgecrest has also produced an exceptionally distributed aftershock sequence that just kisses the San Andreas, but refuses to cross the Garlock Fault.

 

Ridgecrest fault rupture, with a main rupture and distributed faulting over 20 m (70 ft). Photo by USGS. To see a series of fault rupture images, see Stewart et al. (2019).
Ridgecrest fault rupture, with a main rupture and distributed faulting over 20 m (70 ft). Photo by USGS. To see a series of fault rupture images, see Stewart et al. (2019).

 

References

 

James Dieterich (1994), A constitutive law for rate of earthquake production and its application to earthquake clustering, J. Geophys. Res., 99, 2601-2618, doi.org/10.1029/93JB02581.

Ruth A. Harris and Robert W. Simpson (1996), In the shadow of 1857—the effect of the great Ft. Tejon earthquake on subsequent earthquakes in southern California. Geophys. Res. Lett. 23, 229–232, doi.org/10.1029/96GL00015

Ruth A. Harris and Robert W. Simpson (1998), Suppression of large earthquakes by stress shadows: A comparison of Coulomb and rate-and-state failure, J. Geophys. Res., 103, 24,439-24,451, doi.org/10.1029/98JB00793.

Kenneth W. Hudnut, Leonardo Seeber, and Javier Pacheco (1989), Cross‐fault triggering in the November 1987 Superstition Hills Earthquake Sequence, southern California, Geophys. Res. Letts., 16, 199-202, doi.org/10.1029/GL016i002p00199.

Charles W. Jennings, John L. Burnett, and Bernie W. Troxel (1962), Geologic Map of California, Trona Sheet, Scale 1:250,000, California Geological Survey, Sacramento.

Robert W. Simpson and Paul A. Reasenberg (1994), Earthquake-induced static-stress changes on central California faults, in The Loma Prieta, California, Earthquake of October 17, 1989—Tectonic Processes and Models, Robert W. Simpson, Editor, U.S. Geological Survey Professional Paper 1550F, p. F55-F89.

Ross S. Stein (1999), The role of stress transfer in earthquake occurrence, Nature, 402, 605–609, doi.org/10.1038/45144.

Jonathan P. Stewart, Editor (2019), Preliminary Report on Engineering and Geological Effects of the July 2019 Ridgecrest Earthquake Sequence, Report GEER-064, issued July 19, 2019, 69 p., http://geerassociation.org/administrator/components/com_geer_reports/geerfiles/GEER_Ridgecrest_report_ver1.pdf

Shinji Toda and Ross S. Stein (2003), Toggling of seismicity by the 1997 Kagoshima earthquake couplet: A demonstration of time-dependent stress transfer, J. Geophys. Res., 108, 2567, doi:10.1029/2003JB002527.

Shinji Toda, Ross S. Stein, Gregory C. Beroza and David Marsan (2012), Aftershocks halted by static stress shadows, Nature Geoscience, 5, 410–413, doi.org/10.1038/ngeo1465

  • Ross Stein

    From Nano Seeber, Lamont Doherty Earth Observatory: “Impressive fault interaction! The Garlock Wall is dramatic and mysterious. The branching of the MS rupture as it approaches the Garlock fault is consistent with a fault ending against a buttress.”

  • Ross Stein

    From Ken Hudnut, U.S. Geological Survey: “Nice article. I like the array of interesting points you’ve made about the Ridgecrest sequence. Yes, as Nano said, the fault interaction is impressive & remarkable for sure! At GSA Annual Meeting, I think Brian Olson & Jaime Delano will be speaking about Ridgecrest, and at SCEC, Sue Hough will give a keynote and Katherine Kendrick will present a very important poster on the masterful and ambitious mega-compilation of surface rupture observations.”

  • Ross Stein

    From David Marsan, University of Savoie, France: “Very nice, and impressive result. I guess that, if what you call the ‘cross fault’ was located at the tip of the 7.1 rupture, then the shadow would be absent, so it’s lucky it is more in the middle.

    Given the abundant seismicity before, perhaps a detailed analysis can be done to see how far one has to move away from the main fault to see this shadowing effect – this distance is expected to be related to the spatial heterogeneity of the stress change.

    In any case, very convincing. By the way, this also cast some doubt on the afterslip hypothesis (that aftershocks are due to afterslip).

  • Ross Stein

    Your buttressing comment makes sense. Another mystery is why the Garlock moves at all.
    It is horribly mis-oriented for the regional stress, so it must be as slippery as a watermelon seed.

  • Ross Stein

    You guys published on this 30 years ago!

  • Ross Stein

    These opportunities are rare, as there must be a very high seismicity rate to see the sudden decline, and as you say, you need a geometry that throws an active rupture into a stress shadow. Interestingly, when one looks at M≥2 seismicity, the shutdown is hard to see. So one also needs a very good catalog. Why there is no delay is also interesting. Perhaps there are fewer faults there of many orientations.

    The Garlock Fault behaved strangely. There is minor left-lateral creep where it was stressed by the M 7.1 (evident in the ARIA interferogram), but the burst of post-M 7.1 seismicity increase on or near the Garlock occurs farther west, not where we would expect it to be unless these shocks are not actually on the Garlock.