It came from under the earth

C&I Issue 3, 2014

For many years, the direct danger from lava flows has taken a bit of a back burner in the lexicon of hazards that keep vulcanologists awake at night. This is for two reasons: first, lava flows are – for the most part – slow-moving and so are easily avoided, unlike other volcanic hazards such as explosions, toxic gases and pyroclastic flows. Second, lava is not an easy material to study.

But lava is important not only from its real-world hazard but also in interpreting the movement of volcanic material in the geological past. Since lava covers much of our Moon, the terrestrial planets, and the volcanically-active moons of Jupiter as well as our own planet, lava research has implications for planetary geology as well.

At the Lamont Doherty Environmental Observatory at Columbia University, US, Einat Lev and colleagues have been observing and modelling the flow of lava in the lab. ‘Lava flows hold key information about fundamental processes such as planetary evolution and the creation of new crust,’ Lev points out, ‘but the dynamics of lava flows are complicated [because] lava is not a simple liquid, but rather a multi-phase fluid containing liquid melts, gas bubbles and solid crystals.’

Lava forms when igneous rocks are heated and erupt. There are four chemical types; felsic, intermediate, mafic and super-heated ultramafic. Felsic lavas such as rhyolite are extremely viscous, and typically fragment as they extrude, producing blocky debris known as autobreccias. The high viscosity and strength is the result of high silica, aluminium, potassium, sodium and calcium, forming a polymerised liquid rich in feldspar and quartz, with a higher viscosity than other magma types. Felsic magmas can erupt at temperatures as low as 650 to 750°C. 

Intermediate or andesitic lavas are lower in aluminium and silica, and usually somewhat richer in magnesium and iron. They are hotter than felsic lava –in the range of 750 to 950°C – and less viscous. The higher temperatures tend to destroy polymerised bonds making the magma more fluid.

Mafic or basaltic lavas are typified by their high iron and magnesium content, erupting at temperatures in excess of 950°C. Basaltic magma is lower in aluminium and silica, which further reduced polymerisation within the melt. Viscosities can be relatively low, although still thousands of times higher than water. The low degree of polymerisation and high temperature favours chemical diffusion, so it is common to see large, well-formed ‘phenocrysts’ within mafic lavas. Basalt lavas tend to produce low-profile shield volcanoes or ‘flood basalt fields’, because the fluidal lava flows for long distances from the vent. The thickness of a basalt lava, particularly on a low slope, may be much greater than the thickness of the moving lava flow at any one time, because basalt lavas may ‘inflate’ by supply of lava beneath a solidified crust.

Ultramafic lavas such as komatiite and highly magnesium magmas erupt at temperatures of 1600°C, creating a highly mobile liquid with viscosity as low as that of water. Ultramafic volcanism does not occur today with most of this type of volcanism occurring in rocks greater than 1Ga years old (Proterozoic eon)

Lev points out that the rheology – or flow dynamics – of lava depends on four things: its initial composition, temperature, crystallinity and ‘and its tendency to form internal bubbles’ – its vesicularity. Over the course of eruption and flow emplacement, lava cools, de-gases and deforms, changing its composition and vesicularity. Its flow characteristics alter as a consequence.

Lev’s approach is to melt lava in a giant furnace at 1350°C, then pour it down a stainless steel slope while imaging it with advanced photographic and thermal techniques. There are two methods: Particle Image Velocimitry whereby particles in the rock are identified and tracked by computer to see how they move; and Structure from Motion studies whereby the movements and changing shapes of flows can be reconstructed in three dimensions from cameras mounted above the apparatus. Lev can change the substrate over which the lava flows in the chute and thus observe how lava flow changes while flowing over different surfaces. Her thermal imaging cameras accurately show the changing distribution of temperature in the flow as it moves and cools. ‘Never before has so much basaltic rock – up to 300kg – [been] poured out to create lava flows in a controlled environment,’ she says. ‘This scale enabled me and the rest of the Lava Project collaborative team to study processes that have not been addressed in lab-based experiments... for example, the interaction of lava with snow and ice, [the effect of] channels of different shapes [and] flow beds of different roughness levels.’

Lev’s work has already had important implications for the study of lava flows from volcanoes in Polar regions. One of the group’s most intriguing finds has been that lava flow speeds are orders of magnitude faster when lava travels over ice than over other materials. This is because a melt water lens forms beneath and ahead of the lava flow, so it hydroplanes across the ice surface. However, in experiments where melt advanced on shaved ice or ice covered with a layer of sand, the higher porosity and insulation prevented the formation of an obvious water or steam layer.

Lev’s ice and lava work also has more rarefied implications. Using the team’s ability to control the starting conditions of lava/ice interactions, the researchers believe they will be able to better interpret ancient landscapes where lava and ice have co-existed, thereby adding a new geomorphological tool to the understanding of ancient landscapes. Lev’s work highlights the ability of lava to exploit discontinuities, such as crevasses, to burrow down and through large volumes of ice. The resulting trellis pattern is distinctive and explains what must have caused the unusual features at the Gigjokull lava field during the 2010 Eyjafjallajokull eruption in Iceland. Mars is also thought to exhibit glacio-volcanic features and modelling the structures here on Earth may throw insights in the history of the red planet.

Hugh Tuffen’s research at the UK’s University of Lancaster looks at the properties of lava on the microscopic scale. Like Lev, he is interested in using lab-based observations to be able to model both modern and ancient lava flows. Of the four factors that control lava rheology, the one that interests Tuffen the most is crystallinity.

‘Silica crystals tend to grow in low-temperature basaltic lavas as they cool. Basaltic lavas are more prone to polymerisation than rhyolitic lavas with a lower silica content and this makes the lava far more viscous, so strongly retarding the advancing flows. Therefore any prediction of how fast and far lava’s flow needs us to understand how fast crystals grow inside lava. [At present] predictions are made using flow models that account for the viscosity of the lava and how it is affected by cooling, but crystal growth is mostly ignored and we want to change that.’

Using a microscope linked to a heatable stage, Tuffen is able to take solid lava, remelt it and then observe what happens in the laboratory when it is cooled. ‘By directly measuring the rate at which bubbles and crystals grow... we are able to observe changes in the nature of the magma that affect the way in which the lava is moving.’ In particular, he is interested in the thresholds at which the lava flow starts to slow down and when it begins to form a containing crust. It turns out that this is highly variable and depends on the concentration of silica in the lava.

So not all lavas are the same. The paradox is that basaltic lavas – full of flame and violence – move quickly and then, as a result of their high silica content, cool and crystallise quickly. On the other hand, slow moving rhyolitic lavas refuse to countenance crystallisation and remain molten for much longer.

Rhyolitic lavas also accompany a more deadly form of volcanic apocalypse – the famed and feared pyroclastic flow. Because rhyolitic lavas flow and cool so slowly they can often seal the vents from which they originated. If this happens pressure build up can occur and when the pressure reaches explosive point huge amounts of debris – boulders, dust and pebbles can be hurled high into the atmosphere. Quite apart from the damage to life and property caused as the larger ejecta come back to Earth, the dust shroud can cover all in its path, suffocating life and crushing property under its weight – as happened to the citizens of Pompeii in AD79.

In addition, rhyolitic lava fields are prone to explosions as gas (derived from ground water and trapped air as well as decomposing volatiles trapped by the lava) vesicles become trapped in cooling lava and then are heated again underneath by the still-flowing lava that refuses to crystallise. These explosive aftershocks can go on for years after the eruption has ceased. The menacing quiescence of rhyolitic lavas is one of the main findings of Tuffen’s work. ‘Quite simply, they may not be safe to go near for up to five years after the eruption,’ he says.

John Stephenson, a volcanologist in the Department of Earth Sciences at Edinburgh University, UK, points out that this new work on lava flow would not have been possible 20 years ago. ‘A lot of papers came out in the early 1990s dealing with the theory of how lava flows. But they couldn’t get any real data, certainly not at the speed with which we can now. By imaging lava as it flows, as well as heating it to watch the speed of its crystallisation, we now know that the old models were broadly correct but in need of refinement.’

These hidden complexities explain the importance of work on lava flow. ‘The important thing controlling lava hazards is how fast the lava cools – and this is a function of its crystallinity,’ says Tuffen. ‘We are getting to the stage whereby we will be able to predict the behaviour of lava flows in different volcanic regions from an understanding of its original composition. That kind of ability could be very helpful in saving lives.’

The silica factor

Lavas with a silica concentration of less than 50% are basaltic. ‘This is  the red runny stuff you get in Hawaii and Etna … the sort of typical lava that one watches on the telly that has a viscosity that is not much more than chocolate syrup,’ says Tuffen. Its low silica content also means that as it cools it grows crystals and the faster it cools the larger the crystals are. This has important implications for lava flow.

Lavas with a silica concentration of greater than 70% are rhyolitic; the most typical example is the volcanic glass obsidian. Such lavas are typical of the volcanoes in Tuffen’s field area in Chile. Obsidian type lavas do not crystallise readily and stay molten for up to a million times longer than basaltic lavas. Tuffen points out that these can be more destructive than basalt type lava flows. Although they move more slowly than basaltic lavas they stay molten for much longer, forming a deceptively inert crust beneath which the underlying lava will continue to move for months and even years.

Richard Corfield is a science writer based in West Oxfordshire, UK

Become an SCI Member to receive benefits and discounts

Join SCI