Why the Next Large Earthquake is Likely to be a Big Surprise

Destructive earthquakes occur unexpectedly: although evident everywhere to everybody and essentially tied to earthquake inherent unpredictability, this issue appears difficult to formalize. One formalization is developed here in terms of event recurrence. The problem is cast in statistical terms using reverse queue theory, and the solution applied to a uniquely detailed database of seismogenic sources: the DISS Italian database. As a result, the series of the recurrence times estimated from palaeoseismology and geology is found to be statistically incompatible with the series of the time intervals elapsed since last events. Such an incompatibility stands for a number of active seismogenic sources one order of magnitude larger than the number of sources which produced historic and instrumental earthquakes. In practice, this implies that only 1 of the 10 future Mw ≥5:5 earthquakes will be a recurrence of an event reported in the Italian seismic catalogs. This has radical consequences on seismic-hazard estimates.