A submersible study of the products of a large submarine eruption demonstrates the influence of the ocean on eruption dynamics.

No abstract

25 26A long-standing conceptual model for deep submarine eruptions is that high hydrostatic pressure 27 hinders degassing and acceleration, and suppresses magma fragmentation. The 2012 submarine 28 rhyolite eruption of Havre volcano in the Kermadec arc provided constraints on critical 29 parameters to quantitatively test these concepts. This eruption produced a > 1 km 3 raft of floating 30 pumice and a 0.1 km 3 field of giant (>1 m) pumice clasts distributed down-current from the vent. 31We address the mechanism of creating these clasts using a model for magma ascent in a conduit. 32We use water ingestion experiments to address why some clasts float and others sink. We show 33 that at the eruption depth of 900 m, the melt retained enough dissolved water, and hence had a 34 low enough viscosity, that strain-rates were too low to cause brittle fragmentation in the conduit, 35 despite mass discharge rates similar to Plinian eruptions on land. There was still, however, 36 enough exsolved vapor at the vent depth to make the magma buoyant relative to seawater. 37Buoyant magma was thus extruded into the ocean where it rose, quenched, and fragmented to 38 produce clasts up to several meters in diameter. We show that these large clasts would have 39 floated to the sea surface within minutes, where air could enter pore space, and the fate of clasts 40 is then controlled by the ability to trap gas within their pore space. We show that clasts from the 41 raft retain enough gas to remain afloat whereas fragments from giant pumice collected from the 42 seafloor ingest more water and sink. The pumice raft and the giant pumice seafloor deposit were 43 thus produced during a clast-generating effusive submarine eruption, where fragmentation 44 occurred above the vent, and the subsequent fate of clasts was controlled by their ability to ingest 45 water. 46 3 47

Pumice rafts are floating mobile accumulations of low-density pumice clasts generated by silicic volcanic eruptions. Pumice in rafts can drift for years, become waterlogged and sink, or become stranded on shorelines. Here we show that the pumice raft formed by the impressive, deep submarine eruption of the Havre caldera volcano (Southwest Pacific) in July 2012 can be mapped by satellite imagery augmented by sailing crew observations. Far from coastal interference, the eruption produced a single >400 km2 raft in 1 day, thus initiating a gigantic, high-precision, natural experiment relevant to both modern and prehistoric oceanic surface dispersal dynamics. Observed raft dispersal can be accurately reproduced by simulating drift and dispersal patterns using currents from an eddy-resolving ocean model hindcast. For future eruptions that produce potentially hazardous pumice rafts, our technique allows real-time forecasts of dispersal routes, in addition to inference of ash/pumice deposit distribution in the deep ocean.

The 1875 rhyolitic eruption of Askja volcano in Iceland was a complex but well-documented silicic explosive eruption. Eyewitness chronologies, coupled with examination of very proximal exposures and historical records of distal deposit thickness, provide an unusual opportunity for study of Plinian and phreatoplinian eruption and plume dynamics. The ∼17 hour-long main eruption was characterized by abrupt and reversible shifts in eruption style, e.g., from 'wet' to 'dry' eruption conditions, and transitions from fall to flow activity. The main eruption began with a 'dry' subplinian phase (B), followed by a shift to a very powerful phreatoplinian 'wet' eruptive phase (C1). A shift from sustained 'wet' activity to the formation of 'wet' pyroclastic density currents followed with the C2 pyroclastic density currents, which became dryer with time. Severe ground shaking accompanied a migration in vent position and the onset of the intense 'dry' Plinian phase (D). Each of the fall units can be modeled using the segmented exponential thinning method (Bonadonna et al. 1998), and three to five segments have been recognized on a semilog plot of thickness vs. area 1/2 . The availability of very proximal and far-distal thickness data in addition to detailed observations taken during this eruption has enabled calculations of eruption parameters such as volumes, intensities and eruption column heights. This comprehensive dataset has been used here to assess the bias of volume calculations when proximal and distal data are missing, and to evaluate power-law and segmented exponential thinning methods using limited datasets.