MIT scientists have discovered strong evidence that a 1994 earthquake in the mid-Atlantic off South America was initiated by a 100-second episode of slow slip along the fault, and that such precursors may be typical for large earthquakes located on the ocean-ridge transform system.
The discovery, which was reported in the October 4 issue of Science, gives clues about how some large earthquakes start and could shed light on the long-standing question of which types of earthquakes, if any, are predictable.
"What we learn about quakes in the oceanic ridge-transform system may shed light on earthquakes in general," said Professor Thomas H. Jordan, head of the Department of Earth, Atmospheric and Planetary Sciences. Professor Jordan, Jeffrey J. McGuire, a graduate student in EAPS, and Pierre F. Ihmle, a former MIT student now at the Institut de Physique du Globe in Paris, are co-authors of the study.
The scientists used a battery of sophisticated analytical techniques to look at data collected from a worldwide network of seismic stations in different ways. They specifically looked at low- frequency energy releases that seismologists typically don't study. The slow precursors are not easy to spot on seismometers because there is no obvious isolated burst of energy, but rather anomalies in the size and arrival times of low-frequency waves.
"Earthquakes in the middle of the ocean have a slow, smooth deformation that is difficult to detect. These earthquakes preferentially radiate energy at low frequencies," said Dr. Jordan, the Robert R. Shrock Professor of Earth and Planetary Sciences.
In 1993, Drs. Ihmle and Jordan and another former student, Paolo Harabaglia, said that the 1989 Macquarie Ridge earthquake, which occurred in an oceanic fault south of New Zealand, was initiated by a precursor that released energy slowly and smoothly beginning about six minutes before the main rupture. That large earthquake had a magnitude of 8.2.
The new paper in Science reports that the 1994 earthquake near the equator in the Romanche Transform in the mid-Atlantic ridge-transform system off South America was preceded by a similar episode of slow and smooth energy release that started 100 seconds before the main earthquake. That earthquake had a magnitude of 7.0. Earthquakes of this magnitude can cause major damage if they occur on land.
By exploring the spectral characteristics of the earthquake, or the energy of the seismic waves emitted from it at various frequencies, the scientists discovered that the slow precursor of the Romanche earthquake grew for at least 100 seconds before it triggered a fast, main rupture. The total amount of fault slippage that occurred during the precursor was about 15 percent of the main shock's slippage.
"Due to the low noise levels of the modern global network of seismic stations, this event marks the first time where a slow precursor's energy release can be seen as the first energy arriving at seismometers, providing more convincing evidence for the precursor's existance than in the past," Mr. McGuire said. He added that studying slow precursors may shed light on how big earthquakes get started.
This type of precursor does not occur in California, a highly active earthquake region. One reason may be that the fracture characteristics of continental rocks differ from those of the oceanic crust.
"We hope to learn something fundamental about earthquakes. You need to look at the full complexity of the earthquakes on the planet to understand earthquakes as a whole," Professor Jordan said.
To aid in further study of the precursor phenomenon, he and colleagues have started a global survey that uses very low-frequency waves to study earthquakes of magnitude 6.0 or greater that have occurred since 1995.
Funding for the research came from the National Science Foundation and NASA.
A version of this article appeared in MIT Tech Talk on October 9, 1996.