A new approach to the interpretation of magnetic anomalies generated by geological structures resembling simple geometrical bodies is presented. The method is based on the Genetic-Price hybrid Algorithm (GPA) recently proposed by the authors for the inversion of potential-field data, specifically of self-potential signals. In this paper, the effectiveness of the proposed algorithm is tested on magnetic data for retrieving the parameters of the anomaly causative sources. First, the testing is performed on synthetic noise-free and noisy signals due to magnetized sphere-, dike- and fault-like models, then the analysis is extended to field magnetic anomalies. Concerning the synthetic data analyses, a very good agreement is obtained between assumed and retrieved source parameters. Specifically, the error between true and inverted parameter sets was found to be no higher than 9% even by adding 15% of Gaussian white random noise to the initial dataset. As for the study of field data, the values of depth, horizontal position, effective magnetization intensity and angle provided by the proposed GPA method compare well with those obtained by other interpretative approaches. Finally, the results of the GPA application to the inversion of magnetic data measured in the Mt. Somma-Vesuvius volcanic area are reported. In particular, the interpretation of the magnetic anomalies along a SW-NE profile in terms of multiple dikes provides information about depth and location of buried volcanic structures that match well those from other geophysical and geological analyses.
Read full abstract