An extended marine, active seismic survey has been performed on September, 2001 in the gulfs of Naples and\ud Pozzuoli by recording about 5000 shots at a network of 62 sea bottom and 72 on shore seismographs. 3-D images of the\ud shallow caldera structure are obtained from the tomographic inversion of about 77000 first P arrival times using the Benz\ud et al. [1996] tomographic technique. The buried rim of the Campi Flegrei caldera is clearly detected at about 800 –\ud 2000 m depth, as an anular high P-velocity and high density body. It has a diameter of about 8– 12 km and a height of\ud 1 – 2 km. According to stratigraphic and sonic log data from deep boreholes and tomographic P velocities, the rim is\ud likely formed by solidified lavas and/or tuffs with interbedded lava. This study confirms the existence for a\ud depressed limestone basement beneath the caldera at less than 4 km depth, while no evidence are found for shallower\ud magmatic bodies

We present a singular value decomposition (SVD) based algorithm for polarization filtering of triaxial seismic recordings based on the assumption that the particle motion trajectory is essentially 2-D (elliptical polarization). The filter is the sum of the first two eigenimages of the SVD on the signal matrix. Weighing functions, which are strictly dependent on the intensity (linearity and planarity) of the polarization, are applied. The efficiency of the filter is tested on synthetic traces and on real data, and found to be superior to solely covariance‐based filter algorithms. Although SVD and covariance‐based methods have similar theoretical approach to the solution of the eigenvalue problem, SVD does not require any further rotation along the polarization ellipsoid principal axes. The algorithm presented here is a robust and fast filter that properly reproduces polarization attributes, amplitude, and phase of the original signal. A major novelty is the enhancement of both elliptical and linear polarized signals. Moreover as SVD preserves the amplitude ratios across the triaxial recordings, the particle motion ellipse before and after filtering retains a correct orientation, overcoming a typical artifact of the covariance‐based methods.

.[1] A new image of Mount Vesuvius and the surrounding area is recovered from the tomographic inversion of 693 first P wave arrivals recorded by 314 receivers deployed along five profiles which intersect the crater, and gravity data collected in 17,598 stations on land and offshore. The final three-dimensional (3-D) velocity model presented here is determined by interpolation of five 2-D velocity sections obtained from sequential integrated inversion (SII) of seismic and gravity data. The inversion procedure adopts the ''maximum likelihood'' scheme in order to jointly optimize seismic velocities and densities. In this way we recover velocity and density models both consistent with seismic and gravity data information. The model parameterization of these 2-D models is chosen in order to keep the diagonal elements of the seismic resolution matrix in the order of 0.2-0.8. The highest values of resolution are detected under the volcano edifice. The imaged 6-km-thick crustal volume underlies a 25 Â 45 km 2 area. The interpolation is performed by choosing the right grid for a smoothing algorithm which prepares optimum models for asymptotic ray theory methods. Hence this model can be used as a reference model for a 3-D tomographic inversion of seismic data. The 3-D gravity modeling is straightforward. The results of this study clearly image the continuous structure of the Mesozoic carbonate basement top and the connection of the volcano conduit structure to two shallow depressions, which in terms of hazard prevention are the regions through which magma may more easily flow toward the surface and cause possible eruptions.

SUMMARY A tomographic study has been carried out in Costa Rica utilizing teleseismic events recorded by the national seismic network OVSICORI‐UNA. The data consist of 130 earthquakes which occurred at epicentral distances of 23d̀‐100d̀ (P phases) and 110d̀‐180d̀ (PKP phases). 1378 IASP91 traveltime residuals were computed and inverted using an iterative linearized inversion procedure which incorporates 3‐D minimum‐traveltime ray tracing. The 3‐D velocity structure of the upper mantle, as derived from teleseismic tomography, shows a northeastward‐dipping high‐velocity body beneath northern Costa Rica, interpreted as the subducting Cocos plate. Although less evident, a similar pattern of lateral heterogeneities characterizes the deep structure of central Costa Rica. There is no evidence for deep high‐velocity bodies beneath southern Costa Rica. The 3‐D reconstruction of the subducting Cocos Plate shows a steeply dipping slab (subvertical) beneath northern Costa Rica. Here the oceanic lithosphere is older and denser than in central Costa Rica, where the slab dips at about 60d̀. Southern Costa Rica does not show a dipping slab; this is in accordance with the subduction/collision of the thick and buoyant Cocos Ridge. The geometry of the slab, the maximum velocity anomalies and the sudden change in the local seismicity depth pattern suggest the existence of a tear in the subducted plate, located between the northern and central Costa Rican subduction zones. The strain field in the overriding plate indicates a strong structural control induced by the along‐trench variations in subduction. Northern Costa Rica shows slight tensile deformation, while southern Costa Rica displays moderate to strong compressive deformation. The maximum horizontal gradient of the strain field is located in central Costa Rica where a sinistral transcurrent fault system crosses the country from the Pacific to the Atlantic coast. This system is likely to be generated by the along‐trench variations in subduction.