January 6, 1997

Recent quakes don't appear to violate seismic gap hypothesis

The seismic gap hypothesis of earthquake recurrence, threatened by four large and seemingly "premature" quakes in the last two years, appears under closer scrutiny to remain valid.

The hypothesis maintains that once an earthquake ruptures a fault, the same region will not rupture again until enough time passes--usually many decades or centuries--for stress to rebuild. Researchers have used this hypothesis to look for "gaps" in fault zones that have a previous history of earthquakes but have not ruptured recently, and to identify those gaps as the most likely sites for quakes in the near future. This notion is commonly applied around the Pacific Rim, where most of the world's great earthquakes occur.

However, four events in the last two years seemed to cast doubt on this approach. In each case, quakes slightly smaller than magnitude 8 struck in areas where even larger earthquakes had occurred within the last 10 to 30 years. That's not nearly enough time in most cases, seismologists believe, for such levels of stress to reaccumulate in a fault zone.

Susan Schwartz, director of the W. M. Keck Seismological Laboratory at UCSC, decided to take a closer look at the details of each set of earthquakes. Since the early 1960s, global seismic records have allowed researchers to reconstruct how a quake occurs--where the earth slips along the fault and how that slip is distributed. "We've found that earthquake slip along a fault can vary dramatically," Schwartz says. "Some regions of the fault zone move a lot, while others may not move much at all, even if they're nearby."

In three of the four cases, Schwartz found, the recent quakes broke sections of the faults that had moved little in the original events. Those quake pairs were in the Kuril Islands north of Japan (M 8.5 in 1963, M 7.9 in 1995); the Solomon Islands east of New Guinea (M 8.0 in 1971, M 7.8 in 1995); and northern Honshu in Japan (M 8.2 in 1968, M 7.7 in 1994). "These are all nice examples of subsequent earthquakes filling in the gaps," Schwartz says. "They are not repeat events, even though the epicenters are close together and the aftershock zones have some overlap."

The fourth set of earthquakes, in the central Aleutian Islands, is harder to unravel. There, a M 8.6 quake in 1957 was followed by a M 8.2 in 1986 and a M 7.9 in 1996. The slip distribution for the 1957 quake is difficult to determine, Schwartz says, because it predated the global seismic network. At the very least, it appears that the 1986 and 1996 rupture zones were markedly different.

"The physical concept behind the seismic gap hypothesis is still correct," Schwartz says. "If an area slips in an earthquake, the same area should not slip again in the near future. But because the slip is not uniform over the whole rupture area, there is no 'characteristic' earthquake to identify from one cycle to the next." The need to tease out such levels of detail about each rupture, she says, may make it difficult for seismologists to apply the seismic gap hypothesis in a meaningful way.

--Robert Irion

Return to the Currents home page.

Go to UCSC's home page.