Research

Background

1993 Plinian eruption of Lascar volcano, Chile

Nine percent of the world's population lives within one hundred kilometres of an active volcano. For these people, the most important questions are "when will the volcano erupt?" and "what will happen when it does?". My research addresses this second question: the question of volcanic eruptive style.

Volcanoes show a broad spectrum of eruptive behaviour, from violent, explosive eruptions, (e.g. Vesuvius, 79AD; Krakatoa, 1883; Mt St Helens, 1980) to more gentle, effusive eruptions such as those characteristic of Hawaiian volcanoes and Etna in Italy.

A pyroclastic flow, head-on!

Explosive eruptions are amongst the most hazardous of natural phenomena. They throw large volumes of ash and pumice upwards in an eruption column, which may punch tens of kilometres into the atmosphere, spreading ash over a wide area.

Explosive eruptions pose a very significant hazard to the surrounding population. Pyroclastic flows - avalanches of hot rock, ash and pumice - may travel many kilometres from the volcano. These flows can scour sturdy buildings to their foundations and have been responsible for some of the worst natural disasters of the last century.

A lava fountain in Hawai`i

At the other end of the scale, effusive volcanic eruptions are much gentler, and tend to produce lava fountains and lava flows. Whilst these may be very spectacular, they pose much less of a risk to human life. The economic impact of an effusive eruption may still be significant, however, as lava flows can destroy buildings and infrastructure, and may leave agricultural land sterile for centuries.

A lava dome at Mt. St. Helens

Another type of low-energy eruption is the extrusive eruption. This type of eruption produces a lava dome - a mound of almost-solid lava that builds up over the top of the volcanic vent. This can fall away unpredictably, breaking up to form a pyroclastic flow similar to those created by explosive eruptions.

The distinction between "effusive / extrusive" and "explosive" eruption is not as straightforward as it might seem. Not only is there a full spectrum of types of eruption between these two cases, many volcanoes will produce both effusive and explosive eruptions during their lifetime, sometimes during a single eruptive episode. Some volcanoes, such as the Soufriere Hills volcano on the island of Montserrat in the West Indes, switch unpredictably between the extrusive formation of a lava dome, to explosive eruption, over very short timescales.

Figure 1: A schematic model of an erupting volcano

The degree and nature of the hazard presented by an eruption is initimately related to the style of the eruption. Effective hazard management must be built upon a good understanding of the physical origins of the hazard itself. I work towards explaining why some eruptions are explosive, some are effusive, and some switch between the two. In particular, I aim to show what physical properties of the magma, and what physical processes within the volcano, are responsible for determining the style of the eruption. In order to do this, we first need to build a general physical model for a volcanic eruption.

Volcanic eruptions are driven by buoyancy forces arising from the growth of bubbles of magmatic gas. Figure 1 shows a schematic cross section through a volcano during an explosive eruption. Deep in the earth, pressures are high enough to dissolve gases such as water (which is a gas at these high temperatures) and carbon dioxide into the molten magma. As magma rises, the pressure drops and the gas becomes less soluble. When the pressure drops low enough, some of the gas will come out of solution, forming bubbles in a process called nucleation.

The bubbles cause a dramatic decrease in the density of the magma, causing it to rise buoyantly. As the magma rises further, the pressure drops and more gas comes out solution, making the magma more bubbly. The pressure drop also causes the bubbles to grow decompressively. These effects combine to make the magma less dense, causing it to rise more rapidly. This begins a positive feedback of buoyant rise, decompression, bubble growth and decreasing density, which drives the magma to the surface, where it erupts.

The upward acceleration and expansion of the magma gives rise to immense stresses within the magma. If the stresses are large enough, they can overcome the internal strength of the magma, causing it to fragment into ash and pumice. This fragmentation of the magma is the defining characteristic of an explosive eruption.

Current research

The simple physical model of volcanic eruption outlined above is likely to be broadly applicable to most eruptions. The question then arises: what features of the model control the style of eruption? It is becoming clear that much of the eruptive behaviour of a volcano is controlled by properties and processes that operate at the meso-scale: the scale of bubbles and crystals within the melt. To-date, this area has been somewhat neglected, as volcanologists have focussed their attention on two extreme length scales: at the volcano scale (field volcanologists, conduit flow and eruption modellers, hazard managers) or at the micro-scale (petrologists, geochemists).

My volcanology research comprises several parallel and complementary investigations of fundamental magmatic properties and processes at this mesoscopic scale. These investigations combine laboratory experiments and analytical and computer modelling. My aim is to bridge the gap between measured properties of the magma and the eruptive behaviour at the volcano scale.

My current research encompasses the following themes - links take you to a more detailed (and technical!) description of each topic:

• Rheology: This describes the way a material flows in response to a stress. Magma and lava are usually a "multi-phase" mixture of liquid (molten rock), crystals and bubbles. This gives them a complex rheology which influences the way they erupt and flow. I use a combination of laboratory experiments and computer simulations to investigate this rheology and assess the implications for volcanic eruptions.
  
• Magma permeability: During an eruption, magma may become permeable, allowing gas to escape from the bubbles within the magma. The rate of gas escape is an important factor in determining how violent an eruption will be. I have developed numerical models to explore how permeability develops during an eruption.
  
• Stromboli and other persistently active basaltic volcanoes : Many basaltic volcanoes erupt persistently for decades or centuries. The behaviour of these volcanoes is dominated by the rise of gas bubbles through the relatively low-viscosity magma. I have used laboratory experiments to develop physical models that quantify aspects of this behaviour, including explosivity of strombolian eruptions, and the stability of lava lakes.
  
• Magma mingling: Eruptions are often triggered when a fresh batch of magma enters a magma chamber. This work uses a variety of experimental and analytical approaches to understand the timescales over which mingling of magmas causes eruptions.
  
• Fault slip: The numerical models I have developed to investigate magma permeability have application in other geoscience problems. Fault slip in thermally unstable rocks is an example.