NON-PARAMETRIC STATISTICS AS A TOOL FOR PROVENANCE ANALYSIS IN GRAVEL DEPOSITS: VINCHINA FORMATION (MIOCENE, ARGENTINE) AS A STUDY CASE

The Neogene Vinchina Basin developed between 27°and 33°S as a foreland basin as a response to the Andean deformation in the southern part of the Central Andes. The Miocene Vinchina Formation (Turner, 1964) was deposited mainly in fluvial, fluvial-aeolian, aeolian, and lacustrine environments, reaching up to 6400 meters in thickness representing the main depositional unit in the basin. This unit is remarkably exposed along the Sierra de Los Colorados (La Rioja Province, Argentina), where this study was carried out. The results from statistical analyses from 33 gravel beds, their modal compositions, compositional trends, and distribution patterns, based on in-situ lithological clast-counting are presented. The statistical approach was carried out by using hierarchical clustering and principal components analysis (PCA), which permitted differentiate three compositional clusters thought to represent different petrofacies. The first cluster involves samples with compositions widely dominated by neovolcanic andesitic clasts, where intrabasinal volcanic effusions were the main detrital source. The second one comprises samples with mixed compositions dominated by palaeovolcanic clasts. Detrital sources of this cluster were primarily the cordilleran/precordilleran area, with subordinated contributions from the Western Sierras Pampeanas crystalline basement, and intrabasinal volcanic deposits. The third cluster involves mixed samples with a dominance of crystalline-basement supply. The main source area for these samples was the Western Sierras Pampeanas, although the cordilleran/precordilleran supply is present in significant proportions. This study illustrates the merit of using non-parametric statistics in provenance studies and especially, to detect internal compositional variations when multiple source areas are active.