[S10-1] Development, testing and application of earthquake forecasting models

[S10-1-06] Break of slope in earthquake size distribution and aseismic deformation rate

^○Peter Shebalin¹, Inessa Vorovieva¹, Clement Narteau², Sergey Baranov³ (1.Institute of Earthquake Prediction Theory and Mathematical Geophysics, Moscow, Russia, 2.Equipe de Dynamique des Fluides Geologiques, Institut de Physique du Globe de Paris, Sorbonne Paris Cite, Paris Diderot University, Paris, France, 3.Kola Branch of the Geophysical Survey, Russian Academy of Sciences, Apatity, Russia)

Crustal faults accommodate slip either by a succession of earthquakes or continuous slip, and in most instances, both these seismic and aseismic processes coexist. Recorded seismicity and geodetic measurements are therefore two complementary data sets that together document ongoing deformation along active tectonic structures. We show that creep along the San Andreas Fault is responsible for a break of slope in the earthquake size distribution. This slope increases with an increasing creep rate for larger magnitude ranges, whereas it shows no systematic dependence on creep rate for smaller magnitude ranges. This is interpreted as a deficit of large events under conditions of faster creep where seismic ruptures are less likely to propagate. These results suggest that the earthquake size distribution does not only depend on the level of stress but also on the type of deformation.
Significant part of postseismic deformations is known to be aseismic, and the aftershock process consists of both seismic events and aseismic relaxation of stress. Accordingly, we suppose that magnitude-frequency relation for aftershocks may experience a break of slope. We test this hypothesis on the example of several aftershock sequences of major earthquakes and show that in most cases a two-slopes model statistically is preferable. The practical issue is that forecasting models of aftershocks, hypothesizing a single-slope magnitude distribution of earthquakes, may significantly overstate (or sometimes understate) the expected rates of large aftershocks. We demonstrate it on examples. The research was partially supported by Russian Science Foundation (Project N 16-17-00093).