Géosciences Azur, CNRS, Université de Nice, Sophie Antinopolies, Valbonne, France

The uplift crisis of the 1982-1984 in the Campi Flegrei area underlined the importance of seismic surveillance for this volcanic caldera. One of the key elements for an effective seismic network is to make use of a reliable velocity model for earthquake location. In the present work we will discuss criteria for the construction and validation of a new 3D P-wave velocity model for earthquake location in the Campi Flegrei area built from the integration of two high-resolution 3D tomographic images of the region. The model is used for locating a group of earthquakes from the uplift event of the 1982-1984.

A new experiment called SERAPIS (SEismic Reflection/Refraction Acquisition Project for Imaging complex volcanic Structures) has been planned and carried out, based on off-shore seismic energization and data acquisition on land and on sea-bottom. The experiment was performed in September, 2001 during which the vessel NADIR of IFREMER (equipped with 12, 16-liters airgun) produced more than 5000 air gun shots recorded at a sea-bottom seismograph array of 72 OBS and 62 stations installed on-land. Active seismic refraction DSS (Deep Seismic Soundings) acquired during the surveys conducted in 1980 and 1985 were recovered jointly with seismic data acquired in the Campi Flegrei area in the framework of the MareVes97 (an experiment devoted to the definition of the structure of the Somma-Vesuvio complex) offshore survey. The data set acquired during the SERAPIS experiment has been successfully used to infer 3D images of the volcanic structures of Campi Flegrei and Neapolitan bay. Active seismic waveforms and related P-picks (more than 90000 data) from the SERAPIS experiment are also available in the project data server.

The Campi Flegrei (CF) caldera experiences dramatic ground deformations unsurpassed anywhere in the world. The source responsible for this phenomenon is still debated. With the aim of exploring the structure of the caldera as well as the role of hydrothermal fluids on velocity changes, a multidisciplinary approach dealing with 3-D delay-time tomography and rock physics characterization has been followed. Selected seismic data were modeled by using a tomographic method based on an accurate finite-difference travel-time computation which simultaneously inverts P-wave and S-wave first-arrival times for both velocity model parameters and hypocenter locations. The retrieved P-wave and S-wave velocity images as well as the deduced Vp/Vs images were interpreted by using experimental measurements of rock physical properties on CF samples, to take into account steam/water phase transition mechanisms affecting P-wave and S-wave velocities. Also, modelling of petrophysical properties for site-relevant rocks constrains the role of overpressured fluids on velocity. A flat and low Vp/Vs anomaly lies at 4 km depth under the city of Pozzuoli. Earthquakes are located at the top of this anomaly. This anomaly implies the presence of fractured over-pressured gas-bearing formations and excludes the presence of melted rocks. At shallow depth, a high Vp/Vs anomaly located at 1 km suggests the presence of rocks containing fluids in the liquid phase. Finally, maps of the Vp*Vs product show a high Vp*Vs horse-shoe shaped anomaly located at 2 km depth. It is consistent with gravity data and well data and might constitute the on-land remainder of the caldera rim, detected below sea level by tomography using active source seismic data. For a more exhaustive description of the utilized methodologies, of synthetic tests for spatial resolution and uncertainty assessment and, the interpretation of results, the reader may refer to the paper Vanorio et al. (2005).