Most explosive eruptions involve silicic magma where the treacly consistency of the magma traps bubbles inside the melt until the pressure builds to the point of explosive rupturing. In the case of explosive basaltic eruptions, the magma is much runnier and so the bubbles can rise up through the conduit and escape like bubbles in a glass of coke. This leads to the question: how can basaltic magma erupt explosively?
Many of Iceland’s volcanoes are situated beneath glaciers and as a result, it is commonly assumed that it is the interaction of hot rising magma with cold external water which enables the basaltic magma to erupt explosively. This interaction is called phreatomagmatism and can involve processes such as molten fuel–coolant interactions (think throwing water on boiling oil). However, there are also examples of explosive subaerial eruptions involving negligible external water in Iceland so phreatomagmatism is not the only answer to the question.
In order to examine the shallow conduit processes involved in both subaerial and subglacial explosive eruptions we turned to the 10th century Eldgjá eruption in southern Iceland. Eldgjá is a 70 km-long fissure associated with the Katla central volcano with 12 km beneath the Mýrdalsjökull glacier (figure 1a). The eruption produced 19.7 km3 of lava and 1.3 km3 of tephra in at least 16 explosive episodes.
The tephra deposit is a composite of multiple units (figure 1c) which can be categorised into two rough categories: shiny-black and well sorted tephra, and duller brown and poorly sorted tephra. We were able to identify the origin of each unit by mapping the pattern of dispersal. We found that the black tephra came from subaerial segments of the fissure whilst the brown tephra came from subglacial segments. The grain-size distributions of each tephra type are distinct with the black tephra being coarser and unimodal and the brown being finer and bimodal (figure 2).
Now that we had evidence of the end result of explosive activity in the subglacial and subaerial environments we turned our attention to processes which were occurring prior to fragmentation. The engine which drives a volcanic eruption is the exsolution of gas as the magma ascends. We used vesicle-size analysis (VSA) to study the formation and evolution of gas bubbles in the melt. VSA involves measuring the size and number of bubbles preserved within tephra samples (figure 3). The VSA results show no difference between subaerially- and subglacially-derived tephra (figure 4). This was confirmed using a pairwise two-sample Kolmogorov–Smirnov test.
The grain-size and vesicle-size results indicate that the Eldgjá magma was capable of fragmenting regardless of external water. The role of external water was relatively late-stage and limited to fragmenting an already disrupted magma which resulted in a more poorly-sorted and finer-grained deposit compared to the subaerially-derived material. As to why Eldgjá featured explosive episodes at all given its basaltic composition – this is likely due to a rapid ascent rate (similar to opening a bottle of coke too quickly) but more work is needed to quantify this. For more details and additional observations of the Eldgjá eruption please download our article on the subject – now available for free, for everyone through the journal Volcanica.
William Moreland is a postdoctoral research fellow at Istituto Nazionale di Geofisica e Vulcanologia – Osservatorio Etneo where where his research focuses on explosive basaltic volcanism, using the 122 BC basaltic Plinian eruption of Etna as a case study. This follows the theme of his PhD project which was carried out at the University of Iceland, studying the explosive episodes of the 10th century basaltic fissure eruption of Eldgjá.
Will’s research is based on a combination of fieldwork observations and measurements, and laboratory measurements of the physical properties of tephra, with a sprinkling of dispersal modelling. He is a Free and Open Source Software enthusiast with aspirations of contributing some day. Follow Will on twitter at @volcano_will.