New Work Reinforces Megaquake's Harsh Lessons in Geoscience

1. Richard A. Kerr

High-tech analyses of Japan's March earthquake overturn long-held views of fault behavior and warn that another disaster may be looming.

The moment the Tohoku-Oki earthquake struck off northern Japan on 11 March, many researchers knew their expectations had been shattered. The great offshore fault could not be counted on to behave at all predictably. And using onshore observations to gauge whether an offshore fault is building toward failure has grave limitations.

Now three papers (http://scim.ag/MSimons, http://scim.ag/S-Ide, and http://scim.ag/M-Sato) published online this week in Science help show why the inevitable release of seismic energy failed to play out as expected and why monitoring from afar fell short. The papers also point to a possible huge quake to the south, closer to Tokyo. Seismologists are concerned, says Mark Simons of the California Institute of Technology in Pasadena, but they are also now acutely aware of their limitations. “We have no idea what's going on” to the south, he says, but they're anxious to find out.

A game of ring toss.

March's huge quake (yellow contours) and past smaller quakes (colored loops) have left a patch of threatening fault (question mark).

CREDIT: ADAPTED FROM MARK SIMONS ET AL., SCIENCE (2011)

Many seismologists had thought that the offshore fault north of Tokyo was fairly simple and uniform. The ocean plate diving beneath Japan, the thinking went, should stick and slowly build up enough stress to rupture the fault. And the fault should fail segment by segment in large but not huge earthquakes. That's how the fault seemed to have behaved in recent centuries, with quakes of magnitude 7 to 8 or so popping off on any one segment every few decades or few centuries.

But it turns out that the fault has more than one mode of operation. The three Science papers gauge where and by how much the fault slipped in the 11 March magnitude-9.0 quake. Simons and his colleagues combined seismic data recorded around the world, crustal movements on Japan recorded by GPS, and tsunami waves recorded at buoys at sea. Satoshi Ide of the University of Tokyo and colleagues compared the seismic signature of the magnitude 9 with that of its largest foreshock, a magnitude 7.3. And Mariko Sato of the Japan Coast Guard in Tokyo and colleagues actually measured the motion of the sea floor before and after subsea GPS observations. “You can't have a better recorded earthquake,” says David Wald of the U.S. Geological Survey in Golden, Colorado, who was not involved in any of the studies.

The three studies and unpublished estimates by other groups suggest that during the quake, the descending ocean plate and the overlying plate carrying Japan slipped past each other by as much as 50 to 60 meters. “Those are enormous slips,” Wald says, running about two to three times the maximum slip reported for the magnitude-8.8 Maule, Chile, quake of last year. But the pattern of slip is equally striking. Five contiguous segments of the fault spanning more than 600 kilometers broke at once in the quake, rather than one or at most two, as scientists had assumed. But only two central segments experienced extreme slip, and that high slip was concentrated far offshore on the shallower part of the fault (within the figure's yellow contours). Historic quakes had broken short segments of the deeper part of the fault nearer land (the loops of various colors).

Obviously, the fault is more complicated than most researchers had assumed. Simons and his colleagues suggest that some irregularity on the fault is to blame. Something—perhaps a seamount on the sinking plate—pinned the high-slip patch of fault in place for 500 or 1000 years, they argue, while patches around it failed repeatedly in smaller quakes. The apparent absence of quakes in the stuck patch led many seismologists to assume that the fault there could be slowly but steadily slipping without building up any strain. And their only means of monitoring the buildup of strain on the fault—GPS measurements of slow ground movement on land—was greatly handicapped by having the stuck patch 150 kilometers offshore. With such a limited perspective on the past release and the current buildup of strain, a magnitude-9 quake caught researchers by surprise.

Learning that most of the March megaquake's slip was concentrated on two segments makes scientists more worried about other faults around the Pacific. “If you can get a 9 that is this compact,” Wald says, “it increases the number of places you can [fit in] a 9 where you may not have expected one.”

All eyes are now on the southern portion of the length of fault that broke in the Tohoku quake. Neither historical quakes nor the Tohoku quake has broken the offshore, shallow portion of the fault there. And the Tohoku rupture transferred stress southward along the fault, abruptly increasing the stress there. As had been the case to the north, researchers can't say for sure whether that portion of the fault (marked by the question mark) has been freely slipping without generating quakes or locked and building strain toward a quake.

If the offshore southern portion is indeed stuck, Simons and colleagues see “the possibility of a sibling to the 2011 event” that could be “similar to what just occurred offshore,” but half as far from Tokyo. So researchers are anxious to find out whether the stress transferred southward from the 9 has accelerated slow slip on the fault and thus defused the threat of a quake. If the fault isn't slipping, another quake would be in the works. Speed is of the essence: A magnitude-8.7 sibling quake followed the 2004 Indian Ocean megaquake by 3 months.