One to many seismogenic sources: from single earthquakes to seismic sequences

One of the major challenges in seismology is the lack of a unified view of earthquake dynamics embracing all the spatial and temporal scales at which it takes place: while several models describe successfully how single seismic events develop and statistical laws have been set up to characterize the distribution of seismicity in magnitude, space and time, at least as a consequence of mainshocks, such dichotomic information is rarely put into communication to better understand how seismogenesis occurs. Although a comprehensive, cross-scale theory is intrinsically impossible because of the different levels of physical complexity involved in the seismological processes, it is possible to derive several well-known links, as well as interesting new ones, between coseismic properties and long-term statistical patterns. We introduce a simple model based on optimization criteria to explain such mathematical relationships. Given an initial energy perturbation localised on a fault segment, the interface breaks down if the perturbation increases its energy beyond the breakdown level. The slip occurs and the fracture spreads rapidly. Not only that (which is not enough to explain how the fault system will behave at large scales), since the fault zone is in an unstable and frustrated state, i.e., a configuration forced by tectonic stress: meanwhile the fracture propagates, the adjoining interfaces and volumes move towards a more stable energy level, amplifying energy release by a certain amount. Then, the latter can be interpreted as a measure of the triggering power of fracture and applied to connect local and collective dynamics. We focus on the role of tectonic setting and the differences between in-fault and off-fault seismicity. Our model can reproduce several features of seismicity, such as the dependence of the b-value of the Gutenberg-Richter law on the tectonic setting, the correlation between b- and p-value of the Omori-Utsu law, the fractal dimension of hypocentral time series, duration of seismic sequences and the efficiency of the seismogenic process. We utilise the same framework to analyse the composition of moment tensor solutions of global and regional shallow seismicity in terms of double-couple vs non-double-couple contributions. We find that thrust events are characterized by higher double-couples with respect to normal and strike-slip fault earthquakes. Our results are also coherent with the broadly studied stress dependence of the scaling exponent b-value of the Gutenberg-Richter law, which turns out to be anticorrelated to the double-couple contribution. Our work suggests that the structural and tectonic complexity of the seismogenic source marked by the roughness of faults and the width of the dislocation zones has a significant impact on coseismic dynamics, which should be considered in the routinely applied observational procedures to avoid systematic biases in the estimation of the parameters of the seismic source, e.g., the seismic moment.