Reconstructing eruptive processes from volatile distribution in volcanic glass

There are two broad categories of volcanic eruption: explosive and effusive. Explosive eruptions produce tall clouds of ash and pumice, which may fall over a wide area, and hot, fast-moving pyroclastic flows, which pose an acute hazard to local communities. Effusive eruptions produce slow-moving flows and domes of lava and are usually much less hazardous. Many volcanoes, such as Mt St Helens (USA), Merapi (Indonesia) and Soufrière Hills (Montserrat, a UK territory in the West Indies) may erupt either effusively or explosively, and may switch between these styles during a single eruption. This is particularly hazardous because these switches may happen with no warning, making effective hazard management very difficult. Consequently, understanding the causes of transitions in eruption style is one of the most important goals in volcanology.

This study will develop a new tool that will help volcanologists to understand why volcanoes switch between eruption styles. The tool can be used at so-called "silicic" volcanoes that have a particular type of magma that is very sticky (viscous). Because the magma is so viscous, these volcanoes are capable of producing the most violent types of eruption. One such volcano is Novarupta in Alaska - its eruption in 1912 was the most powerful of the 20th century and the fourth most powerful in the last 1000 years. We will focus our study on this eruption because it switched eruption style several times, and because it has exceptionally well-preserved deposits of the magma that it erupted.

Volcanic eruptions are driven by buoyancy forces that arise when water, which is initially dissolved in the magma while it is stored underground, comes out of solution and forms bubbles of steam -much like the formation of bubbles in champagne, which cause it to spray out when the cork is popped. Our new tool works by measuring how much water is left in the magma when it cools, and how it is distributed around the bubbles that are 'frozen' in when the magma cools. A PhD study conducted by members of this research team has shown that it is possible to reconstruct the changes in pressure and temperature that a sample of magma experienced during eruption by measuring the way that the water is distributed around the bubbles. We do this using a spectroscopic technique that can make maps of the water distribution that are accurate to a few thousandths of a millimetre. By understanding the differences in pressure and temperature history of magma samples that were erupted explosively and effusively, we can determine what physical differences caused the change in eruption style.

We will use artificial pumice samples, which are produced in the laboratory and therefore have very well constrained 'eruption' conditions, to test and perfect our tool. We will then apply it to samples from the eruption of Novarupta in 1912. There are a few specific questions that we want to answer, the most important of which are: Are the switches in eruption style caused by changes to the way the gas comes out of the magma? If so, what causes the change, and where in the volcano's plumbing system does this happen?

Finding the answers to these questions won't just help us to understand the Novarupta eruption, but to understand why eruptions that switch between effusive and explosive are so common at silicic volcanoes. This will have impact well beyond our study. It will help volcanologists to work out whether a new eruption is likely to switch in style; it may even allow us to work out what signs to look for that a change in style is imminent. Ultimately, this will help to protect at-risk communities from one of the most serious natural hazards.

The primary non-academic beneficiaries of this research programme are volcano observatories, hazard management and civil defence teams, school and university students, and the PDRA associated with the project.

Explosive silicic eruptions pose a global hazard and volcano observatories and hazard management teams around the world are tasked with planning for, monitoring, managing and mitigating such eruptions. Whilst the risk is greatest around the Pacific Rim, the UK is also susceptible to hazards from explosive silicic eruptions. Montserrat, in the West Indies, is a British Overseas Territory that has been severely impacted by the ongoing eruption of the Soufriere Hills volcano. This eruption is an exemplar of the hazards of effusive/explosive switching. The British Isles are also at risk from explosive silicic eruptions of Iceland's volcanoes; for example, Askja (1875) produced a very large, and well-documented explosive rhyolitic eruption which sent ash clouds over mainland Europe.

The proposed study will produce tools and knowledge that will be of value to hazard managers planning for, or in the event of, an explosive silicic eruption affecting UK interests. Our research programme is primarily 'blue skies', so the impact will mainly be realized following further studies by us and other researchers. Those studies will apply our tools beyond our case study eruption (Novarupta 1912) to improve understanding of the eruptive history of other silicic volcanoes, allowing better assessment of the risk of future events. The key to achieving this impact is engagement with academic beneficiaries, which is discussed elsewhere, and with observatories. Our approach will be shaped through collaboration with staff at the Alaska Volcano Observatory following fieldwork on Novarupta (see AVO Letter of Support). The timescale for impact is 3-10 years.

We will exploit the phased nature of our numerical model development programme to produce educational resources for undergraduate and postgraduate numerical modelling courses. This is discussed in detail in the Pathways to Impact document. The resources that we produce will help to produce graduates with strong quantitative modelling skills, in line with NERC's 'most wanted' skills for doctoral graduates. The timescale for impact is 2-5 years.

The wider public (including schools) will benefit from close interaction with active research during outreach events and schools visits. We will also develop teaching resources and learning activity packages alongside other traditional outreach activities (see Pathways to Impact). We aim to inspire interest in physical science in general, and volcano science in particular, amongst these users.

Our PDRA will benefit from specific training and development of analytical and/or numerical modelling skills, as well as highly transferable multimedia, presentation and written and oral communication skills.