MIT scientists have found that the 1989 Macquarie Ridge earthquake, the largest seismic event in 14 years, was initiated by a precursor that released its energy very slowly and smoothly beginning some six minutes before the main rupture.
The discovery, which was reported in the July 9 issue of Science, gives clues to how some large quakes start, and could lead to new strategies for their prediction.
The work should also "generate a healthy debate" within the seismological community, said Professor Thomas H. Jordan, one of the authors of the Science paper, because he and his colleagues found the slow precursor by analyzing signals "that aren't very obvious." Professor Jordan is head of the Department of Earth, Atmospheric and Planetary Sciences (EAPS).
Key to this analysis was a battery of sophisticated analytical techniques the researchers have developed over the last few years. These techniques "allow us to do things that other groups can't," Professor Jordan said. "We're looking at part of the seismic signal that has not been analyzed for slow precursors."
Using data collected from seismic stations in a worldwide network, the team applied these techniques to the Macquarie Ridge earthquake, a temblor of magnitude 8.2 that occurred on an oceanic fault south of New Zealand.
Specifically, they used the techniques to explore the spectral characteristics of the quake, or the frequencies of the seismic waves emitted from it. "Imagine the Earth as a giant bell that rings at certain frequencies. We know how to see if there's an anomaly in that ringing, and we can analyze that," explained Pierre F. Ihmle, lead author of the paper and a graduate student in EAPS. (A third author, Paolo Harabaglia, graduated from MIT this year with an SM degree in EAPS.)
The analytical techniques allowed the scientists to study very low-frequency waves emitted from the quake, and in this case the team found an anomaly in these waves. "The [frequency] spectrum didn't look like the spectrum for a normal earthquake," Mr. Ihmle said.
After laboriously ruling out other effects that could have accounted for the anomalous low-frequency waves, the scientists concluded that the only explanation was a precursor to the main quake that occurred in a region below that rupture and released its energy very slowly and smoothly. (Even so, the scientists say that the precursor had a total magnitude of 7.6 over the six minutes in which it occurred.)
They were disappointed, however, to find no evidence of the precursor on seismometers which record quake vibrations. "The precursor was so large that we thought we might be able to see it on the seismometers," Professor Jordan said.
Why didn't they? "We think because it was so slow, and progressed so smoothly, that the seismometers couldn't detect it as an isolated pulse of energy arriving before the main shock," Professor Jordan said. ( In contrast, the main quake happened very quickly and violently, with an initial burst of energy that lasted about 20-30 seconds, and generated huge waves on the seismometers.)
Slow precursors have been detected for other earthquakes, but this is the first to be discovered on a strike-slip fault, where two of the earth's plates move past each other horizontally. (The other precursors occurred at faults in subduction zones, where one plate slips under another.)
The San Andreas fault in California, like the Macquarie Ridge fault, is also strike-slip. However, slow precursors have never been found for continental quakes. (The absence of such precursors on land is evidently due to the fracture characteristics of continental rocks, which are different from those of the oceanic crust, Mr. Ihmle said.)
Nevertheless, the Macquarie Ridge precursor is exciting because it sheds more light on "slow" earthquakes, or those that release their energy over unusually long amounts of time. "The rupture in a normal earthquake propagates at a speed of over two kilometers per second, whereas the rupture in a slow earthquake travels at less than one kilometer per second," Professor Jordan said. Slow quakes, which are most common in the middle of the ocean (the Macquarie Ridge quake is a good example), "are not well understood," Mr. Ihmle said.
They are also the focus of the MIT scientists' research. And so far, "our work shows that at least some of these slow quakes are compound events consisting of a slow precursor followed by a normal earthquake," Professor Jordan said.
He concluded: "The fact that we've found one class of earthquakes [with slow precursors] gives us some optimism that maybe we can learn something about the processes that occur before a quake, which could lead to viable strategies for short-term earthquake prediction. And that's a psychological boost for seismologists."
This research was sponsored by the NSF and NASA.
WORLD-WIDE NETWORK-This map shows the location of the 1989 Macquarie Ridge earthquake (white cross) and the world-wide network of seismic stations that recorded it. The shaded squares, circles and triangles indicate stations used in the analysis of low-frequency seismic waves; the small black dots indicate stations used in the analysis of high-frequency waves. Map provided by Pierre Ihmle and published in the July 9, 1993 Science (Vol. 261, pg. 177).
A version of this article appeared in the July 14, 1993 issue of MIT Tech Talk (Volume 38, Number 1).