Connecting Local Features of Coseismic Rupture to Large-scale Properties of Seismicity

Since the dawn of earthquake science, a clear split has risen between theoretical seismology and statistical seismology. While the physics of coseismic rupture can be described by using a classical approach, the statistical patterns of seismicity are described by power laws produced by collective processes generated by many-scales nonlinear cooperative effects. We set up a simple model (Zaccagnino et al., 2022) for describing the size-frequency scaling and the temporal evolution of seismicity starting from local physical properties of faults. The key paradigm of our model is that seismicity is driven by the optimization of energy needed to mobilize crustal volumes given some mechanical constraints. An initial perturbation (Ep) interacts with a segment of the fault interface at the level of a certain patch (Figure 1). Based on its internal stability state (U0), represented by the initial internal energy, the interface breaks down if the perturbation increases its energy beyond the breakdown level (Ub). In that case, the slip occurs and the fracture spreads rapidly; not only that, since the fault zone is in an unstable and frustrated state, meanwhile the fracture propagates, the surrounding rock volumes move towards a more stable energy level, amplifying energy release by a factor k. The active role of rock volumes is justified by the belonging of seismogenic crust to the reign of disordered, self-organized systems where many-scales components play a role in producing collective catastrophic behaviors. At large spatial and temporal scales, our model implies that a relationship between fracturing regimes, “efficiency” of the seismic process, duration of the seismic sequences and geodynamic setting exists, with potential impact on seismic hazard. The parameter describing how the number of earthquakes decreases after a major seismic event, p, turns out to be positively correlated to the exponent of the frequency-size distribution of seismicity, b. A collection of one hundred seismic sequences data suggests that p≈2/3(b+1).