Skip to content

A trans-dimensional inversion algorithm to model deformation sources with unconstrained shape in finite element domains

Erica De Paolo 1, Nicola Piana Agostinetti2, Elisa Trasatti3

  • Affiliations: 1 Istituto Nazionale di Geofisica e Vulcanologia Osservatorio Etneo, Catania, Italy. 2 Università degli Studi di Milano Bicocca, Milano, Italy. 3 Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy

  • Presentation type: Poster

  • Presentation time: Monday 16:30 - 18:30, Room Poster Hall

  • Poster Board Number: 283

  • Programme No: 2.4.51

  • Theme 2 > Session 4

Abstract

Geodetic measurements provide insights on deep processes occurring in active volcanic areas. Inverse modeling is employed to interpret these signals and estimate the parameters that describe magma chambers or dike intrusions. Despite the robustness of the available techniques, realistic source representations are still challenging. Analytical models offer rapid solutions that are useful for facing urgent requests during unrest but assume an isotropic elastic crust and define a-priori simple source geometries. Numerical models implementing finite element methods (FEMs) account for the topographic load and crust heterogeneity effects but require longer computational time. In this study, we present a new, efficient, FEM-based modeling approach for full data-driven source shape definition. We approximate a source of potentially any shape aggregating cubic elements of a FE-mesh loaded with a stress tensor. The deformation field is obtained by the linear combination of each unit contribution, scaled with factors depending on the source geometry. With this strategy, we avoid re-meshing, reproducing the deformation patterns of pressurized cavities in a fixed and continuum domain. A library of pre-computed surface displacement, related to each unit in a user-defined volume, is managed by an original trans-dimensional inversion algorithm. This code shapes the source that best fits the observations, aggregating these units and combining their responses. We employ two competing sets of 3D Voronoi cells to sample the model domain, defining which unit belongs to the source and the inactive crust. This software uses parallel computing for optimal performance and is designed to be adaptable to any volcano.