LAUSR

We performed a tectonophysical analysis of earthquake frequency–size relationship types for large Central Asian earthquakes in the regions of dynamical influence due to major earthquake-generating faults based on data for the last 100 years. We identified four types of frequency–size curves, depending on the presence/absence of characteristic earthquakes and the presence or absence of a downward bend in the tail of the curve. This classification by the shape of the tail in frequency–size relationships correlates well with the values of the maximum observed magnitude. Thus, faults of the first type (there are characteristic earthquakes, but no downward bend) with Mmax ≥ 8.0 are classified as posing the highest seismic hazard; faults with characteristic earthquakes and a bend, and with Mmax = 7.5–7.9, are treated as rather hazardous; faults of the third type with Mmax = 7.1–7.5 are treated as posing potential hazard; and lastly, faults with a bend, without characteristic earthquakes, and with a typical magnitude Mmax ≤ 7.0, are classified as involving little hazard. The tail types in frequency–size curves are interpreted using the model of a nonlinear multiplicative cascade. The model can be used to treat different tail types as corresponding to the occurrence/nonoccurrence of nonlinear positive and negative feedback in earthquake rupture zones, with this feedback being responsible for the occurrence of earthquakes with different magnitudes. This interpretation and clustering of earthquake-generating faults by the behavior the tail of the relevant frequency–size plot shows raises the question about the physical mechanisms that underlie this behavior. We think that the occurrence of great earthquakes is related to a decrease in effective strength (viscosity) in the interblock space of faults at a scale appropriate to the rupture zone size.