Earthquakes in Oklahoma? Rodgers and Hammerstein didn't see that one coming.
Indeed, this week's temblors in the Sooner State highlight the challenge scientists face as they try to improve earthquake-hazard assessments in the central and eastern United States, particularly across the lower half of the country.
In regions known for relatively frequent, large quakes, such as the west coasts of North and South America or deep in the heart of Turkey, sources of stress on faults are well known. And the faults themselves are increasingly well-studied, allowing scientists to estimate repeat rates for major temblors along these shifting cracks in Earth's crust.
In the middle of the continental US, however, research over the past decade suggests that trying to estimate future quake activity may be more like a high-stakes game of "Whac-A-Mole."
One fault system might generate a cluster of quakes over a period of a few years, then it delivers diminishing set of aftershocks for centuries while stress migrates to a new fault system, explains Seth Stein, a geophysicist at Northwestern University in Evanston, Ill.
The past of any one fault system may not be a prologue to its future. Trying to divine that future may represent "an exercise in closing the barn door after the horse has gone," he says.
Oklahoma sits squarely in the portion of the country that appears most susceptible to this wandering seismic activity, some studies suggest.
That's likely to be of little comfort to people rattled by this week's quakes. The first shock struck central Oklahoma early Saturday morning, with a magnitude of 4.7. This turned out to be a fore shock in advance of a magnitude 5.6 quake that struck the same fault Sunday night. The area has experienced a series of aftershocks, including another magnitude 4.7 quake Monday night. A magnitude 3.6 aftershock hit early Tuesday afternoon.
Throughout 2009 and 2010, the state has experienced unusually high earthquake activity, although the activity may be typical for the state when viewed over long periods of time, according to the state geological-survey office in Leonard.
No one has a good handle on why the activity has increased. This week's quakes may well be part of this "swarm," Dr. Stein says.
Some residents are asking whether oil and gas extraction in the state may have triggered the quakes, particularly via the practice of hydraulic fracturing, or "fracking." This involves pumping fluids into the ground to force more oil or gas out. Fluids can in effect lubricate faults, reducing the friction that may be holding back an earthquake.
Several studies in the US and overseas have found evidence that fracking and other techniques for injecting fluids into oil and gas formations have triggered small quakes, with varying degrees of certainty.
In August, for instance, researchers with Oklahoma's geological survey looked into complaints that fracking had triggered a series of small earthquakes ranging in magnitude from 1 to 2.8 south of Elmore City.
After sorting through the evidence, the survey concluded that while it was possible fracking triggered the quakes, the data available weren't good enough to support "a high degree of certainty" that fracking was the culprit.
Studies of fracking activities in England have come to stronger conclusions, but the magnitude of the quakes were similarly small, although strong enough for people to feel them.
Stein says this week's fore shock and main shock were too strong to attribute them to oil and gas extraction. Others note that this week's quakes have appeared at depths significantly deeper than deposits where fracking takes place.
The area hit by this week's quakes doesn't show up as vulnerable on the US Geological Survey's hazard map for the state. The vulnerable area sits in southwestern Oklahoma, along the Meers fault, the only fault in the state for which geologists have found evidence at the surface. Its last significant break occurred about 1,300 years ago in a quake estimated at magnitude 6.5 or 7.
Another case of migration? If so, it could fit a pattern Stein and other researchers posit after studying the New Madrid fault zone, centered under southeastern Missouri. That fault was responsible for a set of four intense quakes between 1811 and 1812. Each quake has been estimated to have reached at least magnitude 7, and perhaps as high as magnitude 8.
It has been shuddering at weaker levels ever since.
Stein and University of Missouri geophysicist Mian Liu have studied aftershock patterns from earthquakes along plate boundaries, in the wider regions around such boundaries, and in mid-continent.
Quakes along plate boundaries appear to have the shortest periods of aftershocks, on the order of 15 years. Mid-continent quakes appear to have the longest period of aftershocks – 200 years and counting.
Meanwhile, nearly a decade of GPS measurements around the region show that any "reloading" of the New Madrid fault is happening at a glacial pace. The crust is deforming at a pace of roughly half the width of a human hair per year. That's all that plate tectonics seems to be contributing to the stresses on the fault, compared with tens of millimeters a year along a plate boundary such as the San Andreas fault.
On the basis of those measurements, it's hard to see how the New Madrid fault stands a 25 to 40 percent chance of a magnitude 6 quake or stronger over the next 50 years, the researchers say.
These observations have led Stein, Purdue University geophysicist Eric Calais, and colleagues to propose that many quakes taking place in mid-continent may be aftershocks of large quakes that happened hundreds of years ago.
And if the contributions of plate tectonics to a rise in stress are as small as has been measured in the New Madrid area, geophysicists may have to look for other sources of stress.
In the northeastern quarter of the US, for instance, the crust is still heaving a sigh of geophysical relief at losing a two-mile-think layer of ice that covered Canada and the US northern tier during the last Ice Age. That seems to be more influential in setting up earthquake activity in that region than tectonics alone.
South and west of an arc stretching from Virginia into Illinois and the Dakotas, the stress buildup slows dramatically. This leads the team to suggest that the general motions of the continental plate alone don't account for the quake activity in the rest of the mid-continent, and that other factors need to be sought.