First-Pass prospectivity mapping for Au–Ag mineralization in Sikhote–Alin Superterrane, Southeast Russia through field sampling, image enhancement on ASTER data, and MaxEnt modeling

The largest Au–Ag deposits in the eastern Asian region occur along the borders of cratons and superterranes. Meanwhile, intra-terranean areas host numerous small deposits, the prospecting and exploration of which are complicated by natural factors. The target areas of the Kema Terrane and the Kema Metallogenic Belt (Sikhote–Alin Superterrane, Southeast Russia) stretch along the coast of the Japan Sea to the north from latitude 45 as a part of the Sikhote–Alin Superterrane. The Kema Terrane is an Early Cretaceous back-arc basin, and it is overlain by a volcanic–sedimentary cover that is intruded by now partially eroded Late Cretaceous and Paleogene acidic plutons. This terrane includes epithermal occurrences of Au–Ag mineralization that are related to volcanic uplifts and associated with metasomatically and hydrothermally altered volcanic rocks. Field mapping in this terrane is extremely complicated by dissected relief and dense forest cover. The novelty of this research includes overcoming difficulties in delineating new prospective areas with machine learning and decreasing vegetation effects for prospecting epithermal Au–Ag deposits. The model uses fragmented and incomplete data in combination with field sampling and laboratory analysis of alteration minerals, which accompany Au–Ag mineralization. “Mineral” images derived from directed principal component (DPC) analysis of ASTER bands were used as evidence to complement information on mineral assemblages derived from field sampling. The DPC analysis helped in partial removal of the influence of sparse and moderately dense vegetation on ASTER multispectral data. A mineral occurrence probability map (MOPM) was generated by implementation of the maximum entropy (MaxEnt) method on the DPC-derived “mineral” images. This map delineates discovered deposits not used in model training and reveals previously unknown areas with high mineral occurrence probability.