Multi-elemental plagioclase-melt modeling unravels (dis)equilibrium crystal growth during dynamic magmatic processes

Assessing magma storage conditions relies critically on crystal-melt equilibrium models. However, natural rocks are replete with disequilibrium features that testify to open-system processes and episodic crystal growth. Interface kinetics at the crystal-melt boundary lag behind changes of physicochemical conditions determined by magma dynamics. This poses petrological modeling with a circularity problem due to the interdependence between models based on intensive and extensive parameters. Here we discuss results of a numerical approach that aims at circumventing these difficulties, applying it to an extensive new dataset of plagioclase crystals from Stromboli volcano, characterized by sieve-like An -rich crystal zones with An 69–92 and oscillatory-zoned An -poor crystal zones with An 59–74 , paired with a comprehensive database of matrix glass compositions. Our approach exploits major element partitioning between plagioclase and melt to identify equilibrium melt compositions and solve for the best-fit intensive parameters, and then tests the equilibrium pairs with trace element partitioning models. The well-constrained plumbing system of Stromboli, in which shallow, degassed magma frequently interacts with deeper, volatile-rich magma, represents the ideal test bed to evaluate iterative modeling approaches. Results show that the method used reproduces known equilibrium conditions and identifies disequilibrium crystal compositions formed during dynamic processes, including magma mixing, ascent and crystal resorption. By deconstructing the chemical record preserved in individual crystal zones, this study provides new quantitative constraints on crystallization processes within the Stromboli plumbing system. It is found that, while An -poor compositions invariably resemble equilibrium, the composition of the An -rich plagioclase zones is the result of disequilibrium partitioning triggered by concurrent magma interaction, cooling and degassing. Ultimately, our findings provide fundamental insights into crystal-melt partitioning during plagioclase near- and far-from-equilibrium growth, demonstrating also the broad utility of similar numerical approaches for interpreting magma evolution in open-conduit volcanic systems.