Università di Bologna "Alma Mater Studiorum", Dottorato di Ricerca in Geofisica XVIII Ciclo, 2006
Claudio Satriano
An effective approach to the problem of seismic risk mitigation must take advantage of all the recent progresses in seismology, computer science and telecommunications. Modern real-time seismic information systems are capable to provide rapid notifications on earthquake parameters and expected ground shaking at target sites, trough fast telemetry and processing of data from dense seismic networks. Such systems can significantly shorten the time necessary for emergency response and recovering of critical facilities like roads or communication lines. If earthquake parameters can be estimated before the arrival of strong ground shaking, then a procedure called seismic early warning can be established. The physical basis for early warning is well understood. Namely, destructive S and surface waves travel at about half speed of the P-waves, and seismic waves travel at much slower speed than signals transmitted by communication systems. By processing the first few seconds of ground motion, recorded in the source area, it is possible to determine the size of the event and its location and evaluate the impact of the incoming earthquake on a potential distant target. The issue of a warning can trigger individual measures, like looking for repair or getting away from critical facilities, or automatic actions, such as slowing down trains, preventing planes from landing or deactivating gas pipelines. Designing and developing a real-time seismic information system is, at the same time, a scientific and a technological challenge. In this work both these aspects will be covered, with a particular attention to the development of the Irpinia Seismic Network (ISNet), a prototype seismic alert management system under development in the Campania Region (southern Italy). After a general introduction to the concepts of real-time seismology and their application to the ISNet (chapter 1), a tool for real-time estimation of hypocentral location (RTLoc) is presented (chapter 2). The technique is based on the equal differential time formulation and on a complete probabilistic approach to the problem of earthquake location. The method follows an evolutionary approach, since, at each time step, the location is updated, using information coming from recorded arrivals and not yet triggered stations. Accurate determinations of earthquake parameters, like hypocentral location, require the use of realistic three-dimensional velocity models, since flat, layered media often give poor results in terms of reliable estimates of earthquake source parameters when complex tectonic or volcanic environments are considered. In chapter 3 the construction of two 3D models for the Campi Flegrei Caldera and the Irpinia Region is described. Additionally, a technique for automatic, iterative travel time tomography (ITAAP) is presented. Starting from a coarse reference P-pick, adjusted travel times are calculated trough waveform cross-correlation and used to determine the velocity model. Theoretical travel times from the resulting tomographic image are used as new reference pick, and the process is iterated. Among the potential applications of the method, the automatic update of reference velocity models for permanent seismic networks is discussed. The last chapter (4) covers the technological aspects of the Irpinia Seismic Network. ISNet is designed as a system where data acquisition, processing and diffusion is distributed over its nodes. This formulation is the basis for the Distributed Event Management System (DEMaS), introduced and discussed in this chapter.