A 665 year record of Coulomb stress changes on active faults in the central Apennines, Italy.

Abstract:

Active extension in the central Apennines is accommodated on numerous 20-30km long normal faults. Over multiple earthquake cycles fault slip is controlled by viscous flow in narrow shear zones, which are below the brittle seismogenic crust and are driven by upwelling mantle beneath the central Apennines. However, on short timescales, there is evidence for clustering along strike on the north eastern set of faults in the region, with the south western faults comparatively quiet during the period of reliable historical earthquake records (since 1349 AD). In contrast, 15±3ka strain rates show no evidence of skewness towards the north eastern faults. This suggests that on short timescales, elastic loading and fault interaction may be controlling the location of earthquakes and the seismic hazard, as opposed to the view that fault activity has permanently migrated from the south west flank of the central Apennines to the north east flank. We used Coulomb stress modelling to test whether the sequence of historical earthquakes can be explained by stress triggering and elastic loading.

Palaeoseismic and historical records were used to reconstruct the co-seismic static Coulomb stress changes for 27 earthquakes in central Italy from 1349-2009. 15±3ka throws measured across faults in the area were used as an analogue for the slip distributions, with the slip direction constrained by field measurements of frictional wear striae on exposed bedrock fault scarps. Interseismic loading was modelled using a shear zone rheology below the seismogenic zone of each fault; slip rates measured at the surface were used to control the rate of loading. The sensitivity of the model was explored by iterating varying slip distributions, fault kinematics and earthquake locations.

We show that for sequences of clustered earthquakes that occurred on timescales of days to weeks, co-seismic static Coulomb stress transfer can explain the pattern of faulting with stress changes of 0.001-0.1 MPa transferred onto faults preceding rupture. Over hundreds of years, the relationship between long-term stress loading and earthquake occurrence is less clear and highlights the need for a greater understanding of the time elapsed since the last earthquake on active faults in the central Apennines, so as to constrain the earthquake cycle on all faults in the area.