Three-dimensional kinematic depth migration of converted waves: application to the 2002 Molise aftershock sequence (southern Italy)

International audienceMigration techniques, currently used in seismic exploration, are still scarcely applied in earthquake seismology due to the poor source knowledge and sparse, irregular acquisition geometries. At the crustal scale, classical seismological studies often perform inversions based on the arrival time of primary phases (P- and S-waves) but seldom exploit other information included in seismic records. Here we show how migration techniques can be adapted to earthquake seismology for converted wave analysis. As an example, we used data recorded by a dense local seismic network during the 2002 Molise aftershock sequence. In October and November 2002, two moderate magnitude earthquakes struck the Molise region (southern Italy), followed by an aftershock sequence lasting for about one month. Local earthquake tomography has provided earthquake hypocenter locations and three-dimensional models of P and S velocity fields. Strong secondary signals have been detected between first-arrivals of P- and S-waves and identified as SP transmitted waves. In order to analyse these waves, we apply a prestack depth migration scheme based on the Kirchhoff summation technique. Since source parameters are unknown, seismograms are equalized and only kinematic aspects of the migration process are considered. Converted wave traveltimes are calculated in the three-dimensional (3D) tomographic models using a finite-difference eikonal solver and back ray tracing. In the migrated images, the area of dominant energy conversion corresponds to a strong seismic horizon that we interpreted as the top of the Apulia Carbonate Platform and whose geometry and position at depth is consistent with current structural models from existing commercial seismic profiles, gravimetric and well dat