Igneous rock associations 20. Pearce element ratio diagrams: Linking geochemical data to magmatic processes

Abstract

It has been nearly fifty years since Tom Pearce devised a type of element ratio diagram that isolates the effects of crystal fractionation and accumulation (sorting) hidden in the chemistry of a suite of igneous rocks. Here we review the guiding principles and methods supporting the Pearce element ratio paradigm and provide worked examples with data from the Mauna Ulu lava flows (erupted 1970-1971, Kilauea Volcano, Hawaii). Construction of Pearce element ratio diagrams requires minimum data; a single rock analysis can suffice. The remaining data test the model. If the data fit the model, then the model is accepted as a plausible or likely explanation for the observed chemical variations. If the data do not fit, the model is rejected. Successful applications of Pearce element ratios require the presence and identification of conserved elements; elements that remain in the melt during the processes causing the chemical diversity. Conserved elements are identified through a priori knowledge of the physical-chemical behaviour of the elements in rock-forming processes, plots of weight percentages of pairs of oxides against each other, or by constant ratios of two elements. Three kinds of Pearce element ratio diagrams comprise a model: conserved element, assemblage test, and phase discrimination diagrams. The axial ratios for Pearce ratio diagrams are combinations of elements chosen on the basis of the chemical stoichiometry embedded in the model. Matrix algebra, operating on mineral formulae and analyses, is used to calculate the axis ratios. Models are verified by substituting element numbers from mineral formulae into the ratios. Different intercepts of trends on Pearce element ratio diagrams distinguish different magma batches and, by inference, different melting events. We show that the Mauna Ulu magmas derive from two distinct batches, modified by sorting of olivine, clinopyroxene, plagioclase and, possibly, orthopyroxene (unobserved).