2009-11-08

Models and data suggest that mid-continent earthquakes may well have been aftershocks from earthquakes that occurred more than 100 years earlier.

I have a deep and rather personal interest in earthquake and volcano prediction. This comes from spending most of my youth within a few kilometers of an active fault line and less than 100km from a volcano that has, in the past, left a layer of ash over most of the surface of the Earth. In fact, events in just the last year (nevermind the last decade) have convinced me that accurate earthquake and volcano prediction would probably be a bigger lifesaver than any other single scientific development.

So it was with interest that I read a recent Nature paper reporting that scientists might have been misinterpreting some aftershocks as earthquakes, leading them to overestimate the risk on some faults and underestimate the risk on others.

At present, earthquake prediction—and I use that word in its loosest possible sense—consists of looking at the frequency of small earthquakes along a fault of interest. These smaller earthquakes indicate that it is on the move and that, somewhere along the fault, the pressure is probably building up for a big one. Unfortunately, it is currently impossible to determine where the pressure point is—if, indeed, it does exist—and when it will give.

To make matters worse, events after an earthquake look remarkably similar to events before. After a major quake, the pressure point has relaxed, and there are a whole lot of small earthquakes as all the smaller pressure points get released. On short time scales, the aftershocks are pretty easy to distinguish from the small shocks that may presage an earthquake. The complicating factor is that the Earth really doesn't care about our timescales.

Some fault lines are very inactive—the boundary has basically fused and there is very little movement. These are located in places like New Madrid in Missouri, central China (that one should ring a bell), and Quebec. These places are considered pretty geologically inactive and yet, occasionally, they get hit by an earthquake—usually a big one. Typically, these earthquakes are not presaged by a series of small quakes, so they are even more unexpected.

Seth Stein from Northwestern University and Mian Liu from University of Missouri have modeled the aftershock behavior for faults with varying degrees of movement. They find that faults that have nearly fused have a very long aftershock interval, with large earthquakes being followed by smaller earthquakes over a period that can last a hundred years or more. In fact, as the rate of loading decreases—the loading is the pressure from the movement of the plates against each other—the aftershock interval and the period over which aftershocks occur increases.

They also point out that this aftershock period is strongly dependent on the physics of specific plate boundaries, making it next to impossible to draw more general and useful conclusions about fault and aftershock behavior without studying each fault individually. Despite this, even the incomplete seismic record allows them to conclude that these large earthquakes that seem to come out of nowhere are likely to be aftershocks from earlier, even larger earthquakes—a truly scary thought.

In addition, their work holds an interesting warning. It is quite possible for the aftershock interval to overlap the build up interval for an earthquake, further complicating the business of earthquake prediction.

Where do we go from here? This doesn't add a lot of information for researchers who focus on places like the Ring of Fire, where everything is moving at a tectonic sprint. On the other hand, it represents some long-term hope for regions like central China, where the faults are slowly fusing—the aftershocks will continue, but the area might well become geologically stable. For instance, researchers have looked at the GPS data for the New Madrid region, and found no measurable deformation: could the 1811-12 earthquakes have been its swan song?

  Email Article

  Discuss in the Forum

Back