Possible origin of memory in earthquakes: Real catalogs and an epidemic-type aftershock sequence model

Earthquakes are one of the most devastating natural disasters that plague society. Skilled, reliable earthquake<br> forecasting remains the ultimate goal for seismologists. Using the detrended fluctuation analysis (DFA) and<br> conditional probability (CP) methods we find that memory exists not only in inter-occurrence seismic records,<br> but also in released energy as well as in the series of the number of events per unit time. Analysis of a stan-<br> dard Epidemic Type Aftershock Sequences (ETAS) earthquake model indicates that the empirically observed<br> earthquake memory can be reproduced only for a narrow range of model’s parameters. This finding, therefore<br> provides tight constraints on the model’s parameters and can serve as a testbed for existing earthquake forecast-<br> ing models. Furthermore, we show that by implementing DFA and CP results, the ETAS model can significantly<br> improve the short term forecasting rate for the real (Italian) earthquake catalog.