GNGTS 2013 - Atti del 32° Convegno Nazionale

Three dimensional seismic imaging and earthquake locations in a complex, normal faulting region of southern Apennines (Italy) O. Amoroso 1* , A. Ascione 2 , S. Mazzoli 2 , J. Virieux 3 , A. Zollo 1 1 Department of Physics, University ‘Federico II’, Naples, Italy 2 Department of Earth Sciences, Environment and Georesources, University ‘Federico II’, Naples, Italy 3 ISTerre, CNRS, Université de Grenoble I, France * Now at Istituto Nazionale di Geofisica e Vulcanologia, Osservatorio Vesuviano, Naples, Italy Introduction. Delay time tomography based on local earthquakes can provide a detailed three-dimensional image of seismic velocity structure in areas which are expected to be affected by strong earthquakes. Refined earthquake locations in a reliable velocity model, allow to detect and characterize crustal structures, and embedded active faults with a high seismogenic potential (Eberhart-Philipps, 1993). On the other hand, the availability of a velocity model and the seismicity distribution allow to study the relationship between the geometry and mechanical behavior of a fault or faults system and the physical properties of the host environment (Michael and Eberhart-Phillips, 1991). The accuracy of earthquake locations is strongly controlled by several factors, among which the geometry of the network, the number of available phases especially the capability of identifying both P and S phases, the accuracy of arrival time readings and the knowledge of the crustal structure (Pavlis, 1986). In addition, the use of 1D layered velocity models can introduce systematic errors in the estimation of P- and S-travel times due to the presence of large-scale three-dimensional heterogeneities in the propagation medium (Matrullo et al. , 2013). In order to reduce the effect of using a 1D velocity model, relative locations and double differences techniques can be used. With these apporaches, the effect of a poor knowledge of the structure can be cancelled out for two events, very close in space, recorded at the same station travel as they travel following nearly the same path except nearby the source zone (Waldhauser and Ellsworth, 2000). However, Michelini and Lomax (2004) emphasized that systematic errors on earthquake location using double-differences may also be caused if the velocity model is not accurate. In this respect, a joint inversion of hypocenter and velocity parameters could allow to overcome the simplistic assumptions of the location methods mentioned above. The methods used today for the joint tomographic inversion have made substantial progress with respect to the basic theory originally developed by Aki and Lee (1990). Recent methods include efficient techniques for 3D ray tracing calculation using the eikonal equation (Vidale, 1990) for first- arrival traveltimes estimation on a given finite-difference grid in which the precision of travel- time calculation can be significantly improved by a successive integration of the slowness along the ray (e.g, Latorre et al. , 2004). High-resolution imaging of the sub-surface with local earthquake data requires the use of large and consistent data sets of first arrival times. The quality and resolution of the medium image depends not only on the source/receiver coverage of the target region but also on the accuracy of the travel time measurements. The common procedure of reading the arrival time of a phase (picking) involves the manual measurement of P- and S-arrival times on recordings of a single event at a time. Systematic errors can be introduced due to inadequate working procedures such as the interaction between the process of picking and the result of the location. The inconsistency of the data can remain unnoticed when the events are analyzed independently from each other, but it may clearly appear when performing a joint determination of the hypocentral and velocity model parameters and reducing the inconsistency would require a complete picking revision. The knoweledge of the S-wave add important constraints to the earthquake location problem. The S-phase is important to derive physical parameters of the subsurfaces. The correct reading of the arrival times of these waves can be complicated by various factors, such as the superposition of the tail of the P-wave, the presence of converted waves generated at different interfaces, and the splitting of S-waves caused by seismic velocity anisotropy 5 GNGTS 2013 S essione 1.1