Numerical Modelling of Trace Elements Behaviour in Deep Magmatic Processes

Trace element analysis of rocks and minerals can provide valuable insight in the petrogenesis of the crust and therefore may offer fundamental constraints into geodynamical processes. But crustal building involves a multitude of magmatic, metamorphic, or metasomatic processes, and the trace element behaviour during each of these processes is complex, which makes the interpretation of trace element composition of the final crustal product non unique.. Here, we present a novel numerical tool for the prediction of trace element behaviour during magmatic processes. To develop such tool, we used the following approach. First, we compiled the predictive models of partition coefficient based on the lattice strain model (LSM) available in the bibliography. Next, we built our own predictive models by applying statistical regressions on large dataset of trace element partition coefficients when the LSM is not suitable for their prediction. In our study, we focused on the phases that play an important role in trace elements fractionation in mafic systems such as garnet, olivine, pyroxenes, plagioclase and amphibole. We combine this tool with a state-of-the-art numerical model for melt generation, transfer, and crystallisation, allowing us to predict the trace element behaviour for a complex sequence of processes occurring at the mantle-crust interface within arc systems .