Partitioning pumice into rafts and submarine deposits during the 2012 eruption of Havre submarine volcano

Abstract:

The 2012 explosive eruption of Havre submarine caldera produced a >1.5 km³ pumice raft and a ~0.4 km³ submarine deposit that consists, in part, of pumice. Fauria et al. (in review) has shown that pumice can float because of trapping of non-condensable gas (air or magmatic gases) by water. With the goal to better understand how pumice is partitioned into floating and sinking regimes, we use laboratory experiments and X-ray microtomography to probe the floatation behavior of both Havre 2012 raft, and seafloor pumice (collected via ROV). In the first set of laboratory experiments we dry Havre raft and seafloor pumice, place the ambient temperature pumice on water, and measure the time it takes the pumice to sink. We find that the Havre raft pumice floats for much longer (>100 days and still floating) than pumice collected on the seafloor and of similar size (23 +/- 20 days). Furthermore, we find that the floatation time of the seafloor pumice scales with pumice volume to the two-thirds power as predicted by a gas-diffusion control on pumice floatation. These observations suggest that (1) ambient-temperature Havre raft pumice more effectively traps gas than ambient temperature seafloor pumice, and (2) seafloor pumice was hot and possibly filled with steam during emplacement (because additional mechanisms are required to make pumice sink quickly). We conduct a second set of experiments where we place ambient temperature Havre raft and seafloor pumice in water for several days and then encase the pumice in wax to preserve the internal distribution of fluids. We image the gas and liquid in these pumice with X-ray microtomography and find that both raft and seafloor pumice trap gas. By measuring the amount of trapped gas, we can test hypothesis (1) above. Furthermore, we use X-ray microtomography to image the internal pore size distribution of Havre raft and seafloor pumice. We find that seafloor pumice have elongated vesicles with a single size mode while raft pumice exhibit a bimodal pore size distribution. Future analyses will help link our understanding of eruption dynamics as preserved in pumice pore size distributions with an understanding of pumice floatation, gas trapping, and partitioning into rafts versus wholly submarine-transported deposits.