CLASSIFYING A-TYPE PLUTONS USING PRINCIPAL COMPONENT ANALYSIS, WICHITA GRANITE GROUP, SOUTHERN OKLAHOMA

The alkali-feldspar granites of the Wichita Mountains, southwestern Oklahoma, are the exposed expression of magmatism during the Eocambrian rifting of the Laurentian margin. These exhibit typical A-type characteristics: ferroan, alkali-calcic MALI, and elevated Y, Nb, Zr, and Rb. They are the products of dry magmas with an average zircon saturation T= 880 °C. Rapakivi and granophyric texture are prominent features in some of these shallow intrusives. Collectively known as the Wichita Granite Group (WGG), mapping and petrology placed these into eleven lithodemic units; each is presumed to be a pluton, representative of a single intrusive event.

Establishing relationships among the plutons can illuminate the distribution and longevity of felsic sources during this rifting event. Given similar ranges expressed on Harker plots, Myers and Gilbert (Ok. Geo. Notes, 1981) collected the lithodemes into three classes: Mount Scott, Reformatory, and Mountain Park, based on the available data at the time. In the intervening years, additional whole-rock geochemistry data has obfuscated these proposed classes. A few bivariate trends produce clusters relevant to the class designations, such as Sr and Ba vs. Rb; most are poorly resolved.

This investigation utilized principal component analysis (PCA) to evaluate potential relationships. PCA is a multivariate statistical tool that uses correlations in a database to construct a series of orthogonal basis vectors to maximize variance. This study ran the Z-scores for the current compiled list of WGG analyses (n=71) on major-element oxides, selected LILE, and Zr. The eigenvalues for PC1, PC2, and PC3 are those greater than 1.0, and represent 54.9%, 19.7%, and 11.8% of the variance. Divalent major elements produced clear positive loadings, and SiO₂ clear negative loadings, in PC1. The PC2 and PC3 loadings were dominated by the monovalent major elements, along with Sr and Zr. The resulting scores suggest that the original classes hold merit. The Mount Scott and Reformatory bodies occupy distinct fields; the former shared by the related Saddle Mountain Granite. A third field is comprised of the Quanah, Cooperton, and Headquarters Granites. Other units show considerable scatter, warranting further investigation into the nature and processes of lithodeme heterogeneity. Encouragingly, the plot is useful for discriminating smaller, anomalous bodies mapped within pluton boundaries, including two recently characterized members associated with the Quanah Granite.