The Mexico Earthquake of September 19, 1985—Model for Generation of Subduction Earthquakes

A mathematical model for rational decision-making should be based not only on statistical data but also on the salient features of the physical phenomena at play. The model developed here simulates the magnitudes and occurrence times of subduction earthquakes but it may also apply to other tectonic seisms. It uses linear springs and dashpots to idealize the convection drag and ridge push on tectonic plates and the latter's behavior. It divides the rupture area into many faultlets whose frictional strength is lognormally distributed; this strength drops upon breakage and is then gradually recovered. Without introducing ad-hoc features to fit the seismic data, the model produces essentially correct magnitude exceedance rates, quasiperiodic characteristic earthquakes, sequences of fore- and aftershocks, and background noise. The model can be expanded to include spatial correlation.