Chemical and textural zoning in crystals from volcanic rocks can be used to fingerprint the processes that lead to magma storage and eruption, but decoding the crystal record is not straightforward. In the last three years, we have developed new statistical tools and software based on plagioclase chemical zoning of hundreds of crystals that allows to objectively identify the crystal populations. Here we add a multifractal analysis of texture, and aim to link them to the results of crystal-rich simulations. We apply the approaches to a variety of volcanoes and eruptions and use cluster analysis to identify families and analogue systems.
- Earth Observatory of Singapore