Maxime Godano

Bayesian inversion of seismic spectral ratio for source scaling: Application to a persistent multiplet in the western Corinth rift

We propose a method to precisely estimate earthquake source parameters as magnitude, size of rupture, stress-drop and coseismic slip, and their uncertainties. This method, that relies on a Bayesian approach, allows the determination of the scalar seismic moment, corner frequency (fc) and their associated uncertainties, by inverting ratios between seismic displacement spectra of nearby located earthquakes.We apply this method to a large earthquake multiplet (56 events) located under the northern coast of the Corinth gulf at 8 km depth. This multiplet, is regularly active between 2001 and 2007. Results show fcP /fcS ratios globally between 1.0 and 1.5 which is compatible with the values predicted by Madariagas circular rupture model. In details, 6 earthquakes however exhibit corner frequency variations as a function of the station azimuth compatible with linear rupture propagation. Magnitude ranges 1.08 and 2.80 with a b-value of 1.04. Source rupture length globally ranges between 40 and 170 m for stress drop between 1 and 100 MPa. We show that the number of ruptures and the cumulated coseismic slip are maximal at the center of the multiplet: this suggests that Multiplet-866 could be seen as a weak seismogenic patch surrounded by a locked fault. How-ever, the large value of the maximum coseismic slip cumulated over the period 2000–2008 (10 cm) rather suggests creep allowing rapid stress reloading and repeated earthquakes with short delays. We therefore propose that Multiplet-866 is surrounded by a heterogeneous fault surface with both locked and creeping areas.

Inferring fault mechanical conditions from the source parameters of a complex microseismic multiplet in the Corinth rift, Greece

We develop a mechanical model of tight clusters of coplanar seismic asperities, to investigate a particular microearthquake swarm located at 8 km depth in the Corinth rift in Greece, which was active between 2001 and 2007. Although it is classified as a multiplet based on waveform similarity, this seismic sequence is much more complex than a repeating earthquake sequence, and cannot be interpreted as the regular failure of a single asperity forced by surrounding aseismic creep. Here, we suggest that such complex sequences could be generated by the failure of a set of coplanar asperities interacting in a small region of a fault segment. We show that in order to reproduce the dynamics of the observed sequence and the characteristics of the events, the cluster of asperities has to be located very close to an aseismically slipping fault segment, which could be an updip extension of the deep detachment zone in the rift, creeping at 1.5 cm/yr. For more general cases of coplanar clustered asperities, we show that the shape of the cumulative coseismic displacement pattern associated with the repeated failures of the asperities is strongly controlled by the behavior of the fault area surrounding the asperity cluster. In particular, if the cluster is part of a locked fault area, the resulting long-term cumulative displacement is maximum at the center of the cluster. In contrast, an asperity cluster surrounded by aseismic creep leads to a uniform cumulative coseismic slip pattern. The ratio between cumulative slip at the center of the seismogenic patch and cumulative slip at its periphery could therefore be an indicator of the mechanical conditions prevailing on the fault. A systematic study of the source parameters of complex microseismic sequences could therefore provide insights into the mechanical state of active faults continuously generating microseismicity.