An evaluation of the Rb vs. (Y + Nb) discrimination diagram to infer tectonic setting of silicic igneous rocks

Abstract

The most commonly used tectonic discrimination diagrams for granites were introduced by Pearce et al. [Pearce, J.A., Harris, N.B.W., Tindle, A.G., 1984. Trace element discrimination diagrams for the tectonic interpretation of granitic rocks. J. Petrol. 25, 956–983.]. Since then, many studies have shown that some granites defy classification or their geochemical assignment does not fit with the geodynamic environment in which they are thought to have formed. In this paper we evaluate the performance of the Pearce et al. tectonic discrimination method, specifically, the most widely-used Rb-(Y + Nb) diagram, using a new data base of over 250 occurrences worldwide, the tectonic settings of which are fairly well known. We conclude that a correlation of geochemistry and tectonic position exists, but that ambiguities and misclassifications arise from one or both of the following factors. First, complex or polyphase orogeny can mix source rocks of different tectonic provenance. This is common in continental arcs and collisional settings, which can be closely associated in space and time with extensional regimes. Second, differentiation can produce compositional trends which cross field boundaries, especially the VAG to WPG boundary. One can minimize this problem by using less felsic, noncumulate members of cogenetic series. We demonstrate the inherent weaknesses of trace element tectonic discrimination diagrams. Such diagrams are of little use if applied alone, but they can be valuable in combination with other methods such as dating and geologic assessment.