New Evidence For Why Earthquake Swarms Occur Offers Hope For Better Earthquake Forecasting

"We believe the total energy released during the Izu islands earthquake swarm is the highest ever recorded," said Stein. "The energy released by this swarm was almost ten times greater than the 1980-81 Long Valley, Calif., swarm; the rate of energy release is 100 times higher."

Earthquake swarms, Stein noted, do not exhibit "typical" earthquake behavior in which a mainshock occurs, followed by a series of aftershocks that become less frequent with time. Instead, in a swarm there is a sustained high rate of earthquakes. Eventually, the seismic action stops or peters out. Swarms are most common in volcanic areas, such as in Hawaii, the Pacific Northwest, Alaska, Yellowstone, and parts of northern and southern California. However, such swarms also may occur on tectonic faults, including California's famed San Andreas. Why swarms occur has always been a mystery.

The focus of the Toda et al paper was to test a hypothesis of USGS scientist Jim Dieterich about why earthquake swarms occur at such remote distances from the volcanic or fault source. If scientists could validate this theory, the information could then be used, with proper instrumentation, to improve forecasting of large swarm shocks and earthquake hazards in all settings and countries, Stein said.

Dieterich's theory predicted that the rate of earthquakes increases in direct proportion to the rate at which the crust is stressed. It also predicted that the duration of aftershocks of the magnitude-6 mainshocks during a swarm would shorten as the rate of stress on the earth's crust increased.

The data from Toda and Stein's research demonstrated a good agreement with the theory. "We found that the aftershocks of the magnitude-6 earthquakes at Izu normally last for a year; during the swarm, they lasted a day, which fits Dieterich's theory," Stein said.

To test the theory of how such earthquake swarms occur, the scientists used data recorded from a series of seismometers and Global Positioning System sensors that the Japanese had placed on the islands and, in the case of seismometers, on the ocean floor.

The results of their investigation enabled the researchers to infer, with highly accurate data, that the Izu swarm was caused by a blade-like injection of molten rock into the earth's crust over a distance of about 10 miles long and 10 miles deep. Although the magma did not reach the earth's surface, the blade -- or dike -- was forced open by the magma pressure a massive 65 feet. In comparison, the amount of magma injected beneath Long Valley caldera in 1980-81 was one-tenth this amount, Stein noted.

The Izu islands, said Stein, moved 3 feet apart as molten rock was forced into a vertical conduit beneath Miyake volcano. Then the magma burst a wall of the conduit like a ruptured artery, and magma was injected into the crust in a thin, blade-like shape, extending 10 miles to the northwest in one week.

During the next 7 weeks, the blade of magma swelled, like a hot-water bladder filling up, until the walls confining the magma had pulled apart by 65 feet about 8 miles below the seafloor.

"We were trying to understand this odd dog-bone configuration of earthquake events," Stein said. Because the Japanese had such a good system of GPS stations and an excellent record of the seismic activity of the swarm, the researchers were able to determine that it is likely that the dog-bone configuration is a result of how stress from the dike opening is transmitted to the surrounding crust.

"Unlike a 'standard' earthquake, in seismic swarms, the seismic clock is rolling along at a thousand times its normal rate, triggering earthquakes as far away as 25 miles from the dike," Stein said. "The distribution of earthquakes in a swarm is related to the stress of the dike opening and forced expansion by magma."

The benefit of this research, said Toda, is that if continuous GPS and seismic data are available during a volcanic crisis, and if scientists have a good understanding of the geometry of the intruding magma, the rate and likely distribution of damaging magnitude 6 or greater earthquakes can be forecast.

Such a continuous dense network of GPS receivers now exists in Japan, resulting in evacuation of the volcanic island at the center of the 2000 swarm. In the United States, the Plate Boundary Observatory proposal to Congress by the National Science Foundation and NASA and supported by USGS would place such GPS receivers across the entire western United States. Right now, only a few sites in the United States have such monitoring capabilities, including the Southern California Integrated GPS network (SCIGN) and the Bay Area Regional Deformation network (BARD).