Simulated Annealing for Volcano Muography

Abstract

Muography or muon radiography is a non-invasive emerging image technology relying on high energy atmospheric muons, which complements other standard geophysical tools to understand the Earth's subsurface. This work discusses a geophysical inversion methodology for volcanic muography, based on the Simulated Annealing algorithm, using a semi-empirical model of the muon flux to reach the volcano topography and a framework for the energy loss of muons in rock. The Metropolis-Simulated-Annealing algorithm starts from an 'observed' muon flux and obtains the best associated inner density distribution function inside a synthetic model of the Cerro Machin Volcano (Tolima-Colombia). The estimated initial density model was obtained with GEOMODELER, adapted to the volcano topography. We improved this model by including rock densities from samples taken from the crater, the dome and the areas associated with fumaroles. In this paper, we determined the minimum muon energy (a function of the arrival direction) needed to cross the volcanic building, the emerging integrated flux of muons, and the density profile inside a model of Cerro Machin. The present inversion correctly reconstructed the density differences inside the Machin, within a 1 percent error concerning our initial simulation model, giving a remarkable density contrast between the volcanic duct, the encasing rock and the fumaroles area.