Fig. 1. The above graph represents a pair of similar events in China filtered from 0.5 to 5 Hz, with y axes normalized to unit amplitude. The two lines representing the wavelengths, one blue and one cyan, are so similar that they seem to blur into one line. (the differences are more easily discernable in the enlarged image). The lower panels are magnified segments of the topmost panel. The predicted P-wave arrives at 143 s, the S-wave arrives at 256 s, and the Lg-wave arrives at 315 s.

Finding the epicenter of earthquakes has not changed in principle since the 1930s -- after closely examining seismograms from different widely-spaced listening stations, researchers decide on the arrival times of various seismic waves and calculate an approximation. In practice this can result in errors of several miles.

But a pair of seismologists at Columbia University's Lamont-Doherty Earth Observatory, part of the Earth Institute, has developed a method that increases scientists' ability to pinpoint an earthquake's epicenter. In the process of applying the technique to seismograms for earthquakes in and near China that occurred between 1985-2000, seismologists David P. Schaff and Paul G. Richards also discovered that 10 percent or more occurred not more than one kilometer from each other. The results of their work, which appeared in the February 20 issue of the journal Science, could greatly improve understanding of how and why earthquakes occur and how one earthquake might influence later events.

"This is a big step forward. It gives a whole new level of understanding of the earthquake process when we can interpret each earthquake with accurate knowledge of the location of its neighbor," said Richards, the Mellon Professor of Natural Sciences at Columbia.

The newer method, known as cross correlation, permitted the Columbia scientists to establish the relative arrival times of seismic waves with errors reduced to about a hundredth of a second. Previously, seismologists often make errors of a second or more, which can translate into location errors of five or ten miles, and sometimes more.

1301 events (9% of the Annual Bulletin of Chinese Earthquakes), consisting of 950 doublet pairs satisfying the criteria for repeating events, are well-distributed throughout China (circles). Recording stations are denoted by solid triangles.

This fact of spatial and temporal clustering, says Schaff, a post-doctoral research scientist at the Lamont-Doherty Earth Observatory and lead author on the study, suggests some mechanism is involved where one earthquake triggers another.

"Demonstrating the existence and abundance of repeating events shows that earthquakes do not occur randomly in time and space," said Schaff.

Their research was supported by the Defense Threat Reduction Agency, as part of a program of research and development to improve monitoring compliance with the Comprehensive Nuclear-Test-Ban Treaty.

The Lamont-Doherty Earth Observatory, a member of the Earth Institute at Columbia University, is one of the world's leading research centers examining the planet from its core to its atmosphere, across every continent and every ocean. From global climate change to earthquakes, volcanoes, environmental hazards and beyond, Observatory scientists provide the basic knowledge of Earth systems needed to inform the future health and habitability of our planet.

The Earth Institute at Columbia University is the world's leading academic center for the integrated study of Earth, its environment, and society. The Earth Institute builds upon excellence in the core disciplines — earth sciences, biological sciences, engineering sciences, social sciences and health sciences — and stresses cross-disciplinary approaches to complex problems. Through its research training and global partnerships, it mobilizes science and technology to advance sustainable development, while placing special emphasis on the needs of the world's poor.