Donald Turcotte is a researcher who has carved a distinguished career in the geosciences by applying mathematical principles to geophysical topics such as volcanism, mountain building and forest fires. In this exclusive interview with Physics World, Turcotte, who is based at the University of California, Davis, turns his focus to the science of earthquake prediction. He explains his belief that the distribution of earthquakes can be described using the basic scaling laws of mathematics.

Specifically, Turcotte explains how the magnitude and spread of earthquakes relates to a fractal distribution – that is, a pattern that repeats across a range of scales. It implies that, on average, any seismic region should experience predictable numbers of smaller and larger earthquakes. "In general, you can expect that where you have 10 magnitude-five earthquakes in a period of time, you expect to have one magnitude six," he says. Turcotte explains that this mathematics has been used to establish a risk basis globally.

Turcotte concedes, however, that the underlying physics behind this distribution is still not fully understood. And, tellingly, he warns that the approach could not have been used to forecast the Japanese earthquake of 2011 because it still deals in generalities. In other words, there are still great uncertainties when trying to predict specific earthquake events.

This interview was filmed in San Francisco at the 2011 Fall Meeting of the American Geophysical Union.