A 3-DIMENSIONAL GRAVITY INVERSION APPLIED TO SAO-MIGUEL-ISLAND (AZORES)

Gravimetric studies are becoming more and more widely acknowledged as a useful tool for studying and modeling the distributions of subsurfac e masses that are associated with volcanic activity. In this paper, ne w gravimetric data for the volcanic island of Sao Miguel (Azores) were analyzed and interpreted by a stabilized linear inversion methodology . An inversion model of higher resolution was calculated for the Calde ra of Furnas, which has a larger density of data. In order to filter o ut the noncorrelatable anomalies, least squares prediction was used, r esulting in a correlated gravimetric signal model with an accuracy of the order of 0.9 mGal. The gravimetric inversion technique is based on the adjustment of a three-dimensional (3-D) model of cubes of unknown density that represents the island's subsurface. The problem of non-u niqueness is solved by minimization with appropriate covariance matric es of the data (resulting from the least squares prediction) and of th e unknowns. We also propose a criterion for choosing a balance between the data fit (which in this case corresponds to residues with rms of the order of 0.6 mGal) and the smoothness of the solution. The global model of the island includes a low-density zone in a WNW-ESE direction and a depth of the order of 20 km, associated with the Terceira rift spreading center. The minimums located at a depth of 4 km may be assoc iated with shallow magmatic chambers beneath the main volcanoes of the island. The main high-density area is related to the Nordeste basalti c shield. With regard to the Caldera Furnas, in addition to the minimu m that can be associated with a magmatic chamber, there are other shal low minimums that correspond to eruptive processes.