The rupture process of the 1980 Irpinia, Italy, earthquake from the inversion of strong motion waveforms

Abstract

The distribution of slip and the behaviour of rupture propagation during the 23rd November, 1980, Irpinia earthquake ( M S = 6.9) was investigated by analysing strong motion accelerograms. The rupture was assumed to develop on a two-dimensional, purely normal fault whose geometry was constrained by geologic and geodetic data and by the aftershock space distribution. Theoretical seismograms have been computed using the isochron approach with a spatially variable slip intensity. The asymptotic ray theory is used to evaluate the far-field Green functions. The rupture model was updated following a trial-and-error approach. The fit of observed waveforms with synthetic seismograms allowed several peculiarities in the rupture history to be pointed out. The source model resulting from the solution of the forward waveform modelling was then used as starting model for the inversion of ground velocities. Therefore, we performed a linearized inversion for slip velocity and rupture times on the fault by iteratively perturbing the assumed starting model. The inverse problem is solved by means of a tomographic back-projection technique. The inferred source model shows a bilateral rupture which does not propagate uniformly away from the hypocentre. The rupture velocity and the slip distribution on the fault plane are rather variable. The solution confirms the shallow dislocation gap in the central part of the fault suggested by the behaviour of the surface rupture. The results indicate that the rupture propagated northwestward along a further fault fragment for which no surface evidence is observed. The aftershocks occurred mostly inside of or near to the edges of the source areas of maximum co-seismic slip.