Predictability of the Rate of Seismic Energy in North America

The predictability of the rate of seismicity and strong earthquakes of North America is assessed based on the catalog of the U.S. Geological Survey for 1900–2016. A second-order nonlinear differential equation is used as the mathematical model of the study; the algorithms for optimization and predictability assessment are developed by the author. The conducted estimates show high predictability of the rate of seismic energy. Among 1422 analyzed strong earthquakes, the foreshock predictability is revealed for 120 earthquakes (215 thousand determinations) and the aftershock predictability for 1410 earthquakes (more than 3 million determinations). The predictability related to the strong earthquakes appears at small (1.5–3 km) radius of the samples of hypocenters and increases in terms of the number of the predicted earthquakes with the increase in the radius of the samples. The forecast distances in time are, on average, tens of days for foreshock predictability and thousands of days for aftershock predictability. The obtained results demonstrate a very promising potential of the approximation-extrapolation approach for forecasting both the strong earthquakes themselves and the subsequent aftershock decay of seismic activity.