An Aqua image of the August 2006 eruption of Home Reef in Tonga. Pumice rafts like this could be an example of a Tangaroan eruption, where lava is erupted as a buoyant foam, neither explosively nor as a lava flow. Image: NASA Earth Observatory Submarine volcanism is one of the frontiers of volcanology of which we are only scratching the surface. We have so rarely seen a wholly submerged eruption in action, and typically when we do know one is occurring, getting down there to see what is happening is very difficult. Instead, we have to look at the volcanic debris (tephra) that makes it to the surface during the eruption (like the "coconuts" at El Hierro or the Havre pumice raft) or wait until after the eruption to collect pieces of volcanic debris from the seafloor. This means we need to work backwards -- what are the processes that formed this volcanic deposit? Clues include the textures of bubbles or minerals in the tephra, shapes and sizes of bubbles, the density of the material, the types of fractures present in the chunks and much more. It is all the forensics of volcanology that allows for building a model of how eruptions that we never saw happening might have actually occurred. A great example of this is a recent paper in Nature Geosciences by Melissa Rotella and others. In it, they suggest that some submarine eruptions are neither explosive nor effusive, the two typical endmembers of volcanic activity. Instead, the eruptions produce a buoyant volcanic foam that rises through the water column to the surface and float until the chunks get waterlogged or wash ashore. The eruption isn't explosive as there isn't the violent fragmentation you need for an explosive eruption. It isn't effusive either as no lava flow or lava dome is produced. Instead, you get a slow release of blobs of volcanic pumice foam that cool as they rise. Now, that is cool! This Tangaroan activity (as they suggest to call it) is not anything that has been identified before, so what is the evidence? Rotella and others examined tephra deposits from the submarine (mostly) Macauley volcano in the Kermadec Islands north of New Zealand (home to volcanoes such as Havre and Raoul). They looked at material erupted during a significant eruption that occurred ~6,100 years ago from Macauley that produced somewhere between 1 and 5 cubic kilometers of material (it is tricky to get accurate volumes from submarine eruptions). Most of Macauley, including a 8 by 11 kilometer caldera, is underwater except for Macauley Island, so they looked at both pumice deposited on Macauley Island and dredged from the seafloor at Macauley. Then they set about to measure the density and bubble shape/size for these clasts. What they found is that the dredged pumice from the seafloor had a very different character than the pumice that was deposited on land. The dredged pumices had a wider distribution of density, from 0.20-0.5 g/cm^3 versus the narrow range of densities found in pumice deposited on land (which is centered around 0.4 g/cm^3). The dredged clasts also have a different texture, with lots of evenly-shaped bubbles in the low density chunks but as the density increases, the bubbles get more elongate. These textures can be found all in the same chunk, which means it can't reflect different eruption styles, but instead changes in the chunk itself as it cools. These differences in bubble shape and density in the same clast is not seen in pumices collected from typical explosive eruptions (like the pumices collected above the ocean surface at Macauley Island). So, it seems that the Macauley pumices can't be produced by a standard explosive eruption. What about a lava flow or dome? In those cases, you tend to find tiny crystals (microlites) that form as the slowly erupting lava cools and distorted bubbles, caused by stretching of the viscous lava as it erupts - both of which are present in the dredged Macauley pumices either. That means it can't be effusive either.
Figure 4 from Rotella and others (2012) showing schematic the production of buoyant lava blobs from a Tangaroan-style eruption. The chunks of frothy magma break off at the vent and rise upwards, slowly continuing to cool. Eventually, the chunk will crack and fragment, allowing it to become waterlogged and sink.
Image: Rotella and others (2012), Nature Geoscience
. Instead, the authors suggest that the erupting volcanic material drips out as floating blebs of lava (see above). They don't fragment in the conduit like explosive eruptions, but they are too buoyant to form lava flows that stick to the seafloor. Instead, blobs of lava form and float upwards, still cooling and forming new bubbles and fractures. The authors point out some intriguing possibilities of Tangaroan eruptions that have occurred recently (but not identified as such) including meter-scale rhyolite blobs from West Mata in the Mariana Islands, large pumices from Ilopango in El Salvador (a lake eruption) -- even possibly more basaltic eruptions like Socorro Island in Mexico. I know that when I read this, Havre immediately jumped into mind as the plume from the Havre eruption seemed really small to produce such a large pumice raft, so many Havre was really another Kermadec Tangaroan eruption. As the authors suggest as well, maybe a lot of these large pumice rafts that have been spotted are not evidence of large explosive submarine eruptions but rather these "inbetweener" foam eruptions -- and it even makes me begin to wonder how seafloor coring might be overestimating the frequency of explosive eruptions if these Tangaroan-style eruptions are more common that we have previously appreciated.
