USE OF A GEOCHEMICALLY DIVERSE PCA TRAINING SET TO ASSESS THE TECTONIC SETTING OF THE WICHITA GRANITE GROUP, SOUTHWESTERN OKLAHOMA
Principal component analysis (PCA) compared the WGG to a training set of well-defined and recent analyses of granites taken from GEOROC Compilation Files (2023-12-2JETOA_GRANITE_parts 1&2.csv: DIGIS Team). To construct a significant-but-manageable set, this study used 306 complete analyses of selected Phanerozoic granitoids from 15 “location” designations for the 10 common major-element oxides and 4 trace elements (Rb, Sr, Y, and Zr). Selections represented a range of Fe/Mg ratios and alkalinities across a-variety of tectonic settings: collisional arcs and orogenies, post-collisional and within-plate extension, as well as a potentially oblique setting (Coastal Maine Magmatic Province). To this, the study added 89 complete analyses of the WGG. All individual primary components were z-scored based on the set average and deviation.
The resulting first 3 principal components covered 43, 22, and 12 % of the variance, and PC4 describes less than 10%. Dominant loadings (eigenvectors) included Fe2O3T, MgO, Al2O3, SiO2, the alkali oxides,P2O5 and Y. The WGG PCA most greatly differed from arc granites, followed by orogenic examples. The Coastal Maine complexes produced wide, individually-distinct scatter that partially overlapped with the WGG in all three principal components. WGG showed the strongest PC-space affinity with extensional granites from the Younger Ring Complexes, Nigeria, the Poyi Intrusion, Beishan, and post-collisional members of the Tasmanian Lachlan Fold Belt and the Hainan Island Granites of China.