Advanced data analysis techniques in regional scale exploration for orthomagmatic ore deposits in glaciated terrains

Global demand for mineral resources is rapidly increasing due to factors such as population growth, increasing living standards and expanding requirements of modern technologies. In glaciated terrains like Finland, mineral exploration is challenging as mineral deposits are buried under transported cover, mainly till. Thus, till geochemistry is one of the effective tools for mineral exploration in such terrains. This study aims to improve the usability of the existing high resolution till geochemistry dataset, collected by the Geological Survey of Finland (GTK) from 1971 to 1983, and the cleansed dataset is used to develop new sustainable mineral exploration approaches for critical raw materials. The dataset is applied to the problem of identifying highly prospective areas for orthomagmatic ore deposits that host Co, Cu, Ni, V, Ti, and platinum group metals (PGM). The focus of the first paper was on preliminary data analyses and data preprocessing for till profiles. Compositional data analysis identified mismatches between neighbouring map sheets, which may be due to issues related to till sampling, analytical methods, and levelling. In addition, analysing centred log-ratio (clr)-transformed data may give a better insight into the data, and principal component analysis (PCA) effectively revealed the important spatial patterns. The second paper was focused on stratigraphical analyses of the till profile aimed to identify the most favourable range of till depth to use when generating predictive maps, while preserving the orthomagmatic geochemical signal. Two approaches were used: a) fixed depth interval method and b) statistical method. Fuzzy overlay maps, validated using ROC curves and P–A plots, suggested several potential areas for undiscovered orthomagmatic mineralisation. The third paper considers the application of outlier detection methods in mineral exploration. Mineral deposits are generally distinguished by elevated concentrations of the targeted elements compared to surrounding samples. Local outlier detection methods were applied to mineral exploration using geochemical datasets that vary in scale, sampling density, and data quality to predict the favourable areas for new orthomagmatic deposits, addressing the growing demand for developing environmentally responsible exploration methods applicable for both regional and target-scale exploration.