Reassessment of the source of the 1976 Friuli, NE Italy, earthquake sequence from the joint inversion of high-precision levelling and triangulation data

Abstract

In this study, we revisit the mechanism of the 1976 Friuli (NE Italy) earthquake sequence (main shocks M w 6.4, 5.9 and 6.0). We present a new source model that simultaneously fits all the available geodetic measurements of the observed deformation. We integrate triangulation measurements, which have never been previously used in the source modelling of this sequence, with high-precision levelling that covers the epicentral area. We adopt a mixed linear/non-linear optimization scheme, in which we iteratively search for the best-fitting solution by performing several linear slip inversions while varying fault location using a grid search method. Our preferred solution consists of a shallow north-dipping fault plane with assumed azimuth of 282◦ and accommodating a reverse dextral slip of about 1 m. The estimated geodetic moment is 6.6 × 1018 Nm (M w 6.5), in agreement with seismological estimates. Yet, our preferred model shows that the geodetic solution is consistent with the activation of a single fault system during the entire sequence, the surface expression of which could be associated with the Buia blind thrust, supporting the hypothesis that the main activity of the Eastern Alps occurs close to the relief margin, as observed in other mountain belts. The retrieved slip pattern consists of a main coseismic patch located 3–5 km depth, in good agreement with the distribution of the main shocks. Additional slip is required in the shallower portions of the fault to reproduce the local uplift observed in the region characterized by Quaternary active folding. We tentatively interpret this patch as postseismic deformation (afterslip) occurring at the edge of the main coseismic patch. Finally, our rupture plane spatially correlates with the area of the locked fault determined from interseismic measurements, supporting the hypothesis that interseismic slip on the creeping dislocation causes strain to accumulate on the shallow (above ∼10 km depth) locked section. Assuming that all the long-term accommodation between Adria and Eurasia is seismically released, a time span of 500–700 years of strain-accumulating plate motion would result in a 1976-like earthquake.