A reduced order approach for probabilistic inversions of 3D magnetotelluric data

Multi-observable probabilistic inversion (e.g. Afonso et al., 2013a,b, 2016b) is a recent framework specifically designed to provide insights into the physicochemical structure of the lithosphere and its complex interactions with the sublithospheric upper mantle. Of particular relevance is the inclusion of 3D magnetotelluric (MT) data, as it provides complementary information not only on the thermal structure but also on water content and fluid pathways; this is critical for understanding and imaging the complex fluid-rock interactions responsible for mineralization events and water-assisted tectonism. However, in order to isolate the effect of fluids from other potential compositional and thermal \background effects, MT data needs to be informed by other data sets such as seismic and gravity data. In order to include MT data into multi-observable probabilistic inversions, we first need to solve the problem of computational effciency when solving the MT equations in 3D. For this, we have combined probabilistic inversion methods, parallel MT solvers (Zyserman & Santos, 2000) and advanced reduced order modelling techniques to obtain fast, yet accurate, solutions to both the MT inversion and the full 3D joint inversion of MT and surface wave data. Such a probabilistic formalism offers a natural framework to assess non-uniqueness and uncertainties affecting the inversion, which are otherwise hard to quantify using traditional inversion methods. The outcomes of this thesis demonstrate the capabilities of the conceptual and numerical framework for 3D multi-observable probabilistic inversions and open up new exciting opportunities for integrated geophysical imaging of the Earth's interior.