Bayesian rupture imaging in a complex medium: The 29 May 2012 Emilia, Northern Italy, earthquake

Abstract

We develop a new approach to image earthquake rupture from strong motion data. We use a large data set of aftershock waveforms, interpolated over the seismic fault to obtain Green’s function approximations. Next we deploy a Bayesian inversion method to characterize the slip distribution, the rupture velocity, the slip duration, and their uncertainties induced by errors in the Green’s functions. The method is applied to the 29 May 2012 Mw 6 Emilia earthquake, which ruptured a fault buried below the Po Plain sediments (Northern Italy). Despite the particularly complex wave propagation, the near-field strong motion observations are well reproduced with 15 rupture parameters. The rupture and slip velocities were notably slow (~0.5 Vs and <0.5 m/s, respectively), implying that the fault was difficult to break. This method opens some perspectives for earthquake rupture studies in areas where numerical simulations suffer from imprecise knowledge of the velocity structure.