A Cautionary Note on Amphibole Geobarometry ^†

The classical Al-in-hornblende barometer has been very successful in determining the depth of intrusion of the metaluminous cordilleran granitoid plutons that bear the buffering assemblage at near solidus conditions: hornblende-biotite-plagioclase-orthoclase-quartz-sphene-two Fe-Ti-oxides (or one Fe-Ti oxide + epidote)-melt-vapor (e.g., [1,2,3]).

Ridolfi et al. [4] and Ridolfi and Renzulli [5] derived empirical amphibole-only barometric expressions that could be potentially applied to a larger number of phenocrystic assemblages from volcanic rocks. However, Erdmann et al. [6] claimed that these barometers are inaccurate and can give untenable estimates.

A graphical barometer based on the partitioning of Al and Si between amphibole and plagioclase was derived by Fershtater [7] using amphibole–plagioclase compositional pairs of rocks from the Urals. More recently, Molina et al. [8] calibrated an empirical expression based on experimental data that can be applied to igneous and high-grade metamorphic rocks.

In order to compare the reliability of amphibole-only and amphibole–plagioclase barometry, in this work, we test the performance of the expressions of Ridolfi and Renzulli [5] and Molina et al. [8], using an experimental data set compiled from the literature that has been recently published by Molina et al. [9].

In accordance with Erdmann et al. [6], the test reveals unsustainable pressure estimates with the amphibole-only barometric expressions from Ridolfi and Renzulli [5]. In contrast, the amphibole–plagioclase barometer from Molina et al. [8] performs well and yields a precision better than 1.7 kbar for the Qz-Amp-Pl and Ol-free-Cpx-Amp-Pl assemblages, with the amphibole compositions having >1 apfu (23O; normalization to 13-CNK) Al, 0.05–0.27 apfu Ti, and <1.07 apfu Fe³⁺.