This thesis has two interconnected aims: to develop geophysical modelling and inversion methods, and to apply them to mineral exploration through case studies from Norrbotten, Sweden. Mineral exploration increasingly targets deposits at greater depth, where surface observations alone do not constrain subsurface structure. Integrated geoscientific investigations, and geophysical methods in particular, are then one of the few ways to obtain information on physical properties at depth.

The five case studies comprise magnetotelluric, controlled-source electromagnetic, magnetic, and gravity methods, interpreted together with other geoscientific information, most commonly geological and petrophysical. The methodological work centres on the MR3DMod modelling and inversion framework developed for multi-resolution grids. In this thesis the framework was extensively used for electromagnetic data and further developed for 3-D inversion of potential-field data. Three studies, carried out in collaboration with LKAB, investigate iron mineralisation in active mining districts. The remaining two address regional exploration and an electromagnetic exploration workflow focused on a poorly characterised electrical-conductivity anomaly.

The studies present several 3-D geophysical models at local and regional scales and demonstrate their use in geological interpretation and mineral exploration. A central methodological contribution is a robust inversion scheme based on the Huber loss function and implemented through iteratively reweighted least squares. Applied to all magnetotelluric studies, it reduces the influence of observations that cannot be fitted consistently by standard inversion, providing a more quantitative complement to manual data selection and reducing the need to apply substantial error floors. The thesis also shows the value of integrating multiple geophysical models with petrophysical and geological information and presents a clustering approach to support that integration.

Beyond the individual case studies, the thesis identifies two mineral-prospectivity signatures. First, paired extreme positive and negative magnetic-susceptibility contrasts in 3-D models can form a recognisable signature of magnetic, commonly iron-rich, mineralisation. Although the negative contrasts may reflect inversion artefacts from the limitations of smooth inversion and unmodelled magnetic remanence, the physical volumes that produce such artefacts remain of prospectivity interest. Second, regional crustal-scale conductive anomalies can be refined by local magnetotelluric surveys into testable exploration targets. 

Taken together, the results demonstrate a practical multi-scale workflow: regional geophysical models identify prospective structures, and targeted local surveys resolve their 3-D geometry and exploration relevance. By combining improved inversion methods with geological and petrophysical information, 3-D geophysical models can strengthen geological understanding and support more focused mineral exploration workflows.