Inverse Modelling of Volcanomagnetic fields using a genetic algorithm techniques

The inversion problem deals with the identification of the parameters of a volcanic source that causes observable changes in magnetic data recorded in volcanic areas. To study the inverse problem, synthetic data were generated by considering different volcanic sources: the traditional Mogi model, used for describing the inflation/deflation of the magma reservoir, and the Okada source that seems more realistic to model the opening of eruptive fractures. The main geophysical mechanisms leading to magnetic anomalies in volcanically active areas were considered, such as thermomagnetic, piezomagnetic and electrokinetic effects. The inversion problem was formulated following an optimization approach based on the use of genetic algorithms. Firstly, a number of tests were carried out on synthetically generated data to assess the performance of the inversion procedure. An appropriate index was defined to statistically evaluate the accuracy for each parameter of the sources. The results obtained show that both the Mogi and Okada models can be inverted unambiguously provided that a sufficient number of measuring points are considered. Secondly, two real case studies, concerning the thermomagnetic anomaly observed during the 1989 Mt Etna eruption and the piezomagnetic field detected during the 2000 Miyakejima volcano eruption, are reported together with their sensitivity analyses of the inverse solutions obtained. Application of the GA technique to data taken on Mt Etna and Miyakejima volcano allows for estimating the optimum parameters of volcanomagnetic sources.