The 2012 Havre eruption evacuated a crystal-poor rhyolite (∼3–7% crystals) producing a volumetrically dominant (∼1.4 km3) pumice raft, as well as seafloor giant pumice (5–8%) and lavas (12–14%) at the vent (∼0.1 km3), both of which have subtly higher phenocryst contents. For crystal-poor rhyolites like the Havre pumice, it can often remain ambiguous as to whether the few phenocrysts present, in this case, plagioclase, orthopyroxene, clinopyroxene, Fe-Ti oxides ± quartz, are: (a) autocrysts crystallizing from the surrounding melt, (b) antecrysts being sourced from mush and the magma plumbing system, or (c) xenocrysts derived from source materials or chamber walls, or (d) possibly a combination of all of the above.

In crystal-poor magmas, the few crystals present are strongly relied upon to constrain pre-eruptive conditions such as magmatic temperatures, pressures, water content and fO2. A detailed textural and compositional analysis combined with a range of equilibrium tests and rhyolite-MELTS modeling provide the basis for distinguishing autocrystic vs inherited crystal populations in the Havre eruption.

This dataset consists of the following files, which are also listed in the 'Supplementary Material' section of the paper, 'Defining pre-eruptive conditions of the Havre 2012 submarine rhyolite eruption using crystal archives':

1. Supplementary Materials 1: Sea water contamination removal method

2. Supplementary Materials 2: Havre 2012 pumice raft diversity

3. Supplementary Materials 3: Fe-Ti oxide Mg/Mn equilibrium test

4. Sample raw data

Further information:

• X-ray Fluorescence spectroscopy (EPMA): Used for whole-rock geochemistry (Table 1).
• Electron Probe Microanalysis (EPMA): Used for mineral compostion and glass composition (presented in figures and raw data in the paper's Supplementary Material).