Hair trigger earthquakes
One of the more perplexing problems in earth science is the connection between earthquakes. 40 years ago, if asked, seismologists would always assert that one earthquake following another that was far away was just a coincidence. But how far away is “far away”? And within what time window? A recent paper in Nature Geoscience argues that some earthquakes cataloged as single events on single faults are in fact multiple events representing rupture on distinctly separate faults.(A press release is here if you can’t get the original paper).
The earthquake that really reset a lot of thinking was the 1992 Landers, California earthquake. It was found that some areas far away (100s to ~1000 km) had swarms of small earthquakes start up just as the surface wave from the Landers quake passed through. Several decent sized events cropped up at distances kind of far from the mainshock–the M6.7 Big Bear earthquake some 20 km away and minutes later and the M5.7 Little Skull Mountain earthquake on the next day and a few hundred kilometers away. The combination of seeing small earthquakes triggered by the dynamic stresses of a passing seismic wave with the increasing sense that large earthquakes were simply small earthquakes that sort of failed to stop made it more palatable for scientists to connect distant events.
Unfortunately it still isn’t clear how you would trigger, say, the Little Skull Mountain event a full day later and much farther away than any static or dynamic stress drop would seem to suggest. And there are plenty of older earthquake sequences with similar questions. The 1954 Dixie Valley-Fairview Peak-Rainbow Mountain sequence started with a six month lag from the Rainbow Mtn events to the Fairview Peak quake, but that then was followed all of four minutes later by the Dixie Valley quake. Although there is literature arguing that the static stresses might be enough to explain some of the connections, there are enough marginal numbers to suspect we don’t fully understand this sequence either. It almost seems like you can have a sort of Rube Goldberg machine getting tripped by one earthquake, which might launch some smaller events or change fluid pressures that trips some more events that eventually leads to a second big event.
So what? Well, as the current paper points out, we have to be careful in estimating a maximum magnitude event: surface faults that seem disconnected might all rupture together, and over distances rather greater than we thought (something similar happened at a smaller scale in the Landers event, where some faults thought to be separate all ruptured together). This is particularly a concern in southern California, where a rats nest of faults on similar trends holds out the possibility for connected ruptures (one nightmare would be if the Malibu Coast fault links through Santa Monica, Beverly Hills and Hollywood and on to the Raymond Hill Fault and possibly on to the frontal faults on the south side of the San Gabriel Mountains; such an event would be devastating and presently is not considered plausible. It also passes through some of the highest priced real estate on the West Coast). In some ways, the 2011 Tohoko might be similar as more sections of the subduction zone ruptured than many scientists expected. On the other extreme, detailed mapping of the slip distribution on faults from analysis of seismic and geodetic records has shown that slip along a fault can be quite variable. Basically, our old notion that the penny-shaped crack could provide all the insight we needed is pretty much dead.
So for now, we are warned that things might be worse than we currently guess. We don’t have a full handle on the interaction of different earthquakes and fault system; this is where considerable effort is focused so that earthquake forecasts might better reflect what really could happen.