We consider a model of a fault containing two regions (asperities) whose slip is associated with earthquakes on the fault. The stress field generated by seismic events is relaxed in the following post-seismic interval, due to the properties of rocks in the upper mantle. The sequence of asperity slips in an earthquake is controlled by several elements of the model, e.g. the intensity of seismic waves radiation, the coupling between the asperities and the difference in their frictional strengths.

We devise a model of a fault system and study the conditions under which it can generate a sequence of seismic events. We also point out the conditions determining the order of the events in a sequence. A significant conclusion of our analysis is that the features of seismic sequences originated by a given fault system (order of events, duration of the sequence, duration of interseismic intervals) typically change from one sequence to another, due to the complexity of fault interaction.

The paper presents new analytical solutions for both the coseismic slip and the interseismic evolution of a fault with two asperities of different strengths. It enlightens the relationship between the state of the fault before a seismic event and the number and sequence of slipping modes in the event. It shows that the knowledge of the source function of a seismic event constrains the subsequent evolution of the system. The model is applied to the fault that generated the 1964 Alaska earthquake.