Predicting earthquakes — how good are we?!

Interpreting intracontinental earthquakes.

Interpreting intracontinental earthquakes. Our historical record of seismic activity is very short, by geological time scales. So extrapolating that record to predict future earthquakes can lead to nasty surprises, such as 2008’s devastating earthquake in Sichuan, China, which occurred on a fault that had seen little recent activity. Large earthquakes are typically followed byaftershocks whose frequency decays to some background level of seismicity, following an empirical relation known as Omori’s law. But determining the time scale of the decay and the baseline activity can be difficult. A new model by Seth Stein of Northwestern University and Mian Liu of the University of Missouri–Columbia posits an inverse relationship between the aftershock-sequence durations and the slip rates along faults. Large earthquakes are most common along the boundaries of tectonic plates, and the occurrences of aftershocks tend to decay quickly—within a decade or so—to a relatively high background. The relative plate motion at such boundaries can be rapid, faster than 10 mm/yr. Continental interiors, far away from plate boundaries, deform much more slowly, typically less than 1 mm/yr. And thanks to that slower rate of fault loading, aftershocks can last hundreds of years or longer, as shown in the figure. Thus, warn the researchers, interpreting continental earthquakes as steady-state seismicity can overestimate the hazard in presently active areas and underestimate it elsewhere. (S. Stein, M. Liu, Nature 462, 87, 2009.)   —RJF

From Physics Today, the "Physics Update" section, Jan. 2010

Click here to see the full table of contents for the latest issue.