Principal component analysis and K-means clustering as tools during exploration for Zn skarn deposits and industrial carbonates, Sala area, Sweden

This contribution presents an application of principal component analysis (PCA) and K-means clustering as tools for data dimension reduction grouping multivariate, whole-rock lithogeochemical data. The study dataset consists 64 geochemical variables measurements spectrophotometric brightness determined from 181 dolomite marble samples, collected at various distance two contrasting types mineral deposits, 1) stratabound, marble- skarn-hosted Zn-Pb-Ag sulphide deposits 2) industrial deposits. Clustering PCA outputs are assessed based on spatial distribution relative to known interpretability using geological domain knowledge, test if the methods can provide a non-biased classification samples which is useful exploration vectoring. illustrate that three principle components derived centered log-ratio transformed account 79.69% variance. unsupervised division into different groups reflecting contents detrital (siliciclastic-volcaniclastic), biogenic hydrothermal in protoliths. Spatial clusters reveal systematic patterns thus providing guide. most prospective divided ‘halo dolomite’ exhibiting elevated Fe Mn, ‘ore also showing Zn, Pb, Ag, Sb, Hg. be reconciled with magnetite Mn-bearing Mg-silicates carbonates alteration haloes, proximal enrichment sulphides (galena, pyrrhotite, pyrite, sphalerite). Samples these returned low brightness, resulting oxides grinding dark powders during sample preparation, significantly lowering powdered marble, even when occurring concentrations. Conversely, ‘clean group characterized by elements above, high Ca, Mg, Sr total carbon, magnetic susceptibility spatially coincide An additional ‘detrital-rich distinct other content field strength Al, intermediate brightness. variety represent containing higher co-settled volcaniclastic-siliciclastic material precursor. Assessment clustered relation same show Halo Ore differentiated geomagnetic methods, hence proxy their indirect detection geophysical surveys.