Co-seismic displacements associated to the Molise (Southern Italy) earthquake sequence of October-November 2002 inferred from GPS measurements

The 2002 earthquake sequence of October 31 and November 1 (main shocks Mw = 5.7) struck an area of the Molise region in Southern Italy. In this paper we analyzed the co-seismic deformation related to the Molise seismic sequence, inferred from GPS data collected before and after the earthquake, that ruptured a rather deep portion of crust releasing a moderate amount of seismic energy with no surface rupture. The GPS data have been reduced using two different processing strategies and softwares (Bernese and GIPSY) to have an increased control over the result accuracy, since the expected surface displacements induced by the Molise earthquake are in the order of the GPS reliability. The surface deformations obtained from the two approaches are statistically equivalent and show a displacement field consistent with the expected deformation mechanism and with no rupture at the surface. In order to relate this observation with the seismic source, an elastic modeling of fault dislocation rupture has been performed using seismological parameters as constraints to the model input and comparing calculated surface displacements with the observed ones. The sum of the seismic moments (8.9 x 10(17) Nm) of the two main events have been used as a constraint for the size and amount of slip on the model fault while its geometry has been constrained using the focal mechanisms and aftershocks locations. Since the two main shocks exhibit the same fault parameters (strike of the plane, dip and co-seismic slip), we modelled a single square fault, size of 15 km x 15 km, assumed to accommodate the whole rupture of both events of the seismic sequence. A vertical E-W trending fault (strike=266 degrees) has been modeled, with a horizontal slip of 120 mm. Sensitivity tests have been performed to infer the slip distribution at depth. The comparison between GPS observations and displacement vectors predicted by the dislocation model is consistent with a source fault placed between 5 and 20 kin of depth with a constant pure right-lateral strike-slip in agreement with fault slip distribution analyses using seismological information. The GPS strain field obtained doesn't require a geodetic moment release larger than the one inferred from the seismological information ruling out significant post-seismic deformation or geodetic deformation released at frequencies not detectable by seismic instruments. The Molise sequence has a critical seismotectonic significance because it occurred in an area where no historical seismicity or seismogenic faults are reported. The focal location of the sequence and the strike-slip kinematics of main shocks allow to distinguish it from the shallower and extensional seismicity of the southern Apennines being more likely related to the decoupling of the southern Adriatic block from the northern one. (c) 2006 Elsevier B.V. All rights reserved.