PRINCIPAL COMPONENT ANALYSIS FOR GEOCHEMICAL DATA ANALYSIS AND CHEMICAL ELEMENTS ASSOCIATIONS OF PROSPECTIVE KOLUMBE AREA (KEMA TERRAIN, SIKHOTE-ALIN FOLDED BELT)

Native manifestations and deposits of precious metals of folded belts of surrounding of the North Asian and Sino-Korean cratons are of lesser resources and areas; that requires growth of research and technological potential for discoveries of the mineralizations. In Sikhote-Alin folded belt, which is typical example of boundary folded belts developed over a subduction zone, terranes of different age are described. Basements of these terranes are of accretion, island arc and turbidite origin, intruded by magmatic rocks and overlain by volcanic-sedimentary cover. Conducted researches consider area of island arc Kema terrane, which occupies large territories of the Strait of Tartary continental coast. Au-Ag deposits here are related to the origin and development of the active volcanic boundary. Genesis of the ore objects here is epithermal and related to intrusive and effusive complexes of Upper Cretaceous and Paleogene ages. Detailed field lithogeochemical surveys were held on a prospective Kolumbe site where Upper Jurassic–Lower Cretaceous siliceous and terrigenous rocks are intruded by Cretaceous granites of Tatibinski and Olginsky complexes. Research of the geochemical features was done using Principal Component Analysis, the effective method of dimensional reduction and data filtration. Implementation of that method in geochemical data analysis facilitates outlining of the associations of the elements by unifying them into principal components and discover their spacial distribution related to geological complexes of the research area. Determining of the number of principal components used explained variance covered by these components. Derived maps of the spatial distribution of principal components that include elements associations allowed to determine the locations of the local geochemical extrema. Grouping of elements into principal components made it possible to assume the staging in ore deposition in a prospective Kolumbe site. Conducted studies could invigorate further more detailed researches of the minerageny of the considered area.

volcanic belts, machine learning, geological mapping, principal component analysis, exploration geochemistry