Dynamic modelling of stress accumulation in Central Italy: role of structural heterogeneities and rheology

Recent efforts to relate the style of seismicity in the Central Apennines t o the stress build-up induced by active geodynamic processes by means of si mple viscoelastic dynamic models have shown that seismicity in this region is likely to be controlled by two major processes: the westward underthrust ing of the Adriatic plate under the Apennines and the asthenospheric upwell ing underneath Tuscany, In this work, using these results as a starting poi nt, we evaluate, by means of 2-D dynamic models, the effects of structural heterogeneities (major crustal-scale faults: this improvement is made possi ble by the recent publication of the results of the CROP-03 seismic line) a nd of temperature-dependent rheologies within the crust and the lithospheri c mantle on the stress field induced by the two processes mentioned above. Modelling results are compared with detailed seismicity data. We show that, for all models, the predicted stress fields are controlled? a t large wavelengths, by an upward state of flexure in Tuscany and a downwar d state of flexure in the Apenninic foredeep! in agreement with previous st udies. When activating crustal-scale discontinuities, however, strong local rotations of the major axes of the stress tensor occur in the proximity of the faults, generating complex local stress fields characterized by the co existence of compressive and tensional styles. This could explain the coexi stence, as observed within certain regions of central Italy, of compressive and extensional events. When detailing the rheology using plastic behaviou r for the lithospheric portions of the models, a complex profile of yield s tresses is considered, the materials characterized by higher strength being the sedimentary cover and the lithospheric mantle. Model-predicted stress concentrates in these materials. The lateral variability of the thickness o f the sedimentary cover induces strong local rotations, at shallow depths, of the stress eigenvectors. All of the models predict a stress field roughl y compatible with the distribution and style of seismicity of the region. T he use of progressively more realistic rheologies and geometries, however, gives slightly better results when compared to seismicity data. This paper shows that sensible results may be obtained by combining active and natural source seismicity with finite element modelling.