Skip to content

Resolving traction changes on fractures in a 3D heterogeneous crust using fictitious domains

Dabaghi Farshid1, Valérie Cayol2 and Bodart Olivier1

  • Affiliations: 1 The Lyon University, Université Jean Monnet, Institut Camille Jordan, F-42023 Saint-Etienne, France. 2 Laboratoire Magmas et Volcans, Université Clermont Auvergne, CNRS, IRD, OPGC, Clermont-Ferrand F-63038

  • Presentation type: Poster

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

  • Poster Board Number: 228

  • Programme No: 2.4.18

  • Theme 2 > Session 4

Abstract

To track magma transport and to understand the seismic cycle, deformation data are analyzed relying on forward deformation models combined with inversions. When the geometry of the fracture is known, it is discretized with dislocations, and a least square cost function subject to a regularization constraint on the dislocation amplitudes is minimized. Most of the times, the dislocation solutions assume the earth is an elastic, homogeneous half-space, which may lead to inaccurate inversion results. To address this issue and to invert for a more physical quantity (stress vector rather than dislocation), we present a new method based on a fictitious domain approach inspired by extended finite element methods developed for cracks that do not follow volume meshes. The cost function involves the misfit between the solution of the physical problem and the observed data together with regularization terms. An algorithm is presented which relies on the forward and the adjoint problem. Synthetic tests, relying on the determination of the regularization parameters by L-curves, show that an initial solution is reliably determined. The method is then reformulated to be applied to InSAR data, which correspond to an incomplete knowledge of the displacement field.  Synthetic tests demonstrate the efficiency and robustness of the method for one to four Line-of-Sight observations, with missing data and variable noise. The method was then applied to the May 2016 Piton de la Fournaise intrusion, and we found that our results are consistent with a previous analysis reying on a different approach.