PhD Position in ML-Based Interpretation of Geophysical Legacy Data for Mineral Exploration

Doctorat / PhD Type de contrat : TEMPORARY Date limite de candidature : 2026-11-17T23:59:59+01:00

Develop and test new methods for deep exploration of sediment-hosted copper systems by integrating legacy seismic and well-log data with basin modelling, geochemistry and machine learning. Job description We are looking for a PhD candidate who is keen to explore how legacy seismic and well-log data acquired for petroleum exploration can be transformed into a new resource for deep mineral exploration. The candidate will combine quantitative geophysical interpretation, basin modelling and machine learning to investigate a challenging target: the thin Kupferschiefer horizon buried beneath the Netherlands and the North Sea. The Dutch subsurface offers an exceptional dataset of more than 6,000 wells, including approximately 1,700 in which the Kupferschiefer has been identified, together with extensive high-quality 2D and 3D seismic reflection data and available cores. Yet these data were acquired and interpreted for hydrocarbon exploration. Recognizing and mapping a horizon a few metres thick at kilometers depth, and identifying features associated with its mineral potential, requires methods that can extract information across datasets with very different resolution, coverage and physical meaning. The goal of this PhD project is to develop an integrated methodology for characterizing the Kupferschiefer and its geological context. You will analyse well logs to identify the target horizon and extrapolate it to wells where it has not yet been recognized. You will use seismic data for cross-well correlation and improved spatial characterization, and basin analysis to investigate the roles of stratigraphy, faulting and basin evolution. Geological, petrophysical, geochemical and mineralogical information will be integrated to identify characteristics and fingerprints associated with copper mineralization. Machine learning will be an important part of this methodological development. You will establish robust and interpretable baseline models for well-log analysis and investigate approaches designed for sequential data, including convolutional and transformer-based models. You will also contribute to multimodal methods that combine seismic, well-log, geological and geochemical information while incorporating geological knowledge and physical constraints. A postdoctoral researcher will work alongside you for one year, focusing primarily on advanced multimodal and physics-informed machine-learning developments. The research is part of the EU-funded INTELLIGENT project. Results from the Dutch basin will be combined with complementary studies by other groups working on the Kupferschiefer and related systems to develop a regional-scale mineral-system model and transferable exploration workflows. Your multidisciplinary research work will be carried out within the Applied Geophysics and Petrophysics Section and the Resource Engineering Section of GSE at TU Delft.…