Marie Curie PhD Fellow

The Role

We invite applications for a PhD Fellow, to work with Dr Andrew Valentine on the topic of “Physics-informed super-resolution imaging of the Earth’s interior”. The successful candidate will become an employee of Durham University and will simultaneously be registered as a PhD student. They will become part of the MCSA Doctoral Network “EarthSafe”, and benefit from a range of training and networking activities – as well as opportunities for secondments at other partner organisations.

EarthSafe is funded by the European Commission and UK Research and Innovation. It will support 10 PhD researchers, based in Barcelona, Durham, Trieste and Twente, and benefits from an international network of partner organisations drawn from academia and industry. It aims to support the rapid transition to green energy and low-carbon economies in Europe and beyond by creating transformational data-fusion platforms to inform and enhance global exploration frameworks for deep geothermal resources and critical minerals to support green technologies. It will achieve this by harnessing:

1. the complementary strengths and resolutions of multiple satellite and land-acquired geo-datasets;
2. a novel combination of Artificial Intelligence, probabilistic inverse theory and state-of-the-art computational modelling; and
3. a new social innovation component to support the creation and adoption of socially-responsible and sustainable practices.

The project

Robust, interpretable imaging of Earth’s interior is an essential tool in the transition to a green economy, supporting efficient use of geothermal energy resources and the discovery of critical metal deposits. A variety of geophysical observables provide information about the subsurface, including seismic, gravity, and electromagnetic datasets. However, none can provide a perfect image: for both physical and practical reasons, results are always a ‘blurred’ version of reality.

Recent advances in the fields of machine learning and artificial intelligence have yielded new techniques for image processing, including tools for image enhancement, upscaling and inpainting. This project will explore how these techniques can be combined with fundamental geophysical theory to provide high-resolution constraints on subsurface properties, and support interpretation of results. It will identify, translate, adapt and apply relevant results from a number of strands of research, including:

1. AI-driven image processing and ‘super-resolution’;
2. physics-informed machine learning;
3. homogenization theory; and
4. inverse theory.

This project will primarily focus on theoretical investigations, code development, and computational experiments. A secondment to the University of Twente (The Netherlands) will provide training and experience in applied geophysical imaging and multi-observable inversion, and this project will integrate closely with EarthSafe Project 1. It will suit a candidate with a strong interest and background in mathematical and computational topics. Previous exposure to geophysics is not necessarily required, as full training can be given, but candidates should have a clear interest in contributing to the growth of a sustainable geo-resource infrastructure.

The successful candidate will collaborate closely with researchers at the University of Twente, as well as with partners Geoscience Australia, Minerals Target International and BHP. They will be expected to contribute to the general activities of the research group and participate actively in the intellectual life of the department and university. There may be opportunities to teach and demonstrate for labs and field classes, depending on the candidate’s skills and interests.

Advert information