Crossed contrails in the sky above Edinburgh this evening, looking south toward the Pentland Hills. Click to enlarge.

In 2014, Scotland will have a referendum to ask the people if they want to become an independent country.  The Scottish National Party will be campaigning hard for a ‘Yes’ vote.  Does that extend to some kind of deal with air traffic control?

Or perhaps it is a good omen ahead of Scotland against England in the 6 Nations rugby this weekend?

The image is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License. You can use it if you link back to this page.

This post describes my highlights from the 30^th Nordic Geological Winter Meeting that took place in Iceland last week. The most interesting results include that Iceland’s glaciers may be gone in 200 years, that the May 2011 Grímsvötn eruption was supplied by a fresh input of magma, and that melt flows upwards into the columns that form when lavas are cooling.

The meeting took place in Harpa, Reykjavík's newly-opened opera house and conference centre. It was attended by over 250 participants, mainly from Nordic countries. Sessions covered topics such as tephrochronology, glaciology, and igneous, metamorphic and resource geology.

Iceland’s shrinking glaciers

The alarming conclusion of Helgi Björnsson’s talk was that Iceland’s glaciers may have less than two centuries left. This seems incredible, given that they are so huge, but Björnsson demonstrated how their rapid demise results from a positive feedback loop or vicious circle: the smaller these ice caps get, the faster they will shrink. Understanding this depends on a concept known as the Equilibrium Line Altitude (ELA).

Above the ELA, more snow and ice accumulate on the glacier in the winter than are lost in the summer. The opposite is true beneath the ELA, which is known as the ablation zone. As long as the area above the ELA is large enough to collect enough snow to replace the ice lost beneath it, the glacier will be stable. With rising global temperatures, the ELA is rises, the glaciers lose more ice than can be replaced by annual snowfall, and they shrink. But this alone doesn’t explain the rapid shrinking.

Björnsson and colleagues have mapped the bedrock surface beneath Iceland’s glaciers using ground-penetrating radar. This is how we know, for example, that Katla’s glaciers are ~500 m thick. But this means that the surfaces of the glaciers are only above the ELA because of the ice itself. For example, 65% of the area of Vatnajökull is above the ELA, but only 20% of the bedrock is. As glacier shrinks, the surface altitude of the ice decreases, which reduces the area above the ELA, which makes the glacier shrink faster and so it continues.

There are equations that show how snow is compacted into ice. There are equations that show how glaciers flow under their own weight. There are equations that show how ice melts at lower, warmer, elevations. Combining these with the surface shapes of the glaciers, the radar maps of the bedrock and estimates of future climate (2°C increase by 2100), Björnsson and colleagues modelled the glaciers to see how they would change shape in the future. They concluded that they would be mostly gone in 150-200 years.

The disappearing ice was a theme in other glaciology talks and even in other sessions. Another glaciologist, Thorsteinn Thorsteinsson showed the results of 25 years of monitoring the Höfsjökull glacier, during which period it has lost 6.5% of its mass. The summers of 1991 and 2010 were particularly bad for the glacier, when a coating of fresh volcanic ash caused extra melting. Over in the geodynamics session, Thóra Árnadóttir presented results from GPS stations that are looking for tiny movements of the land surface caused by plate tectonics or volcanic activity. Every station that they put on the mountains poking through the glaciers (nunataks) shows a strong upward motion. This is caused by the land rising as the weight of the ice disappears, and is something to bear in mind when looking for signs of inflation at ice-covered volcanoes.

The Reykjanes peninsula basking in the midday January sun. This is where the mid-Atlantic ridge comes ashore. If you swam due south from here, the next land that you would hit would be Antarctica.

Fresh magma drove the Grímsvötn 2011 eruption

Olgeir Sigmarsson’s talk looked at the geochemistry of the magma from the Grímsvötn eruption in May, and concluded that the reason that the eruption was so powerful is that it was driven by a fresh batch of hot, gas-rich magma from deep beneath the volcano (see my post from August to see the deposits close up).

Geochemistry is a very important tool in understanding volcanoes. It relies on two facts, which will be familiar to any students of a Volcanology 101 class:

1. Magma is a cocktail of many different chemical elements and their oxides e.g. SiO₂, Al₂O₃, MgO, CaO, F, Cl.
2. Magma doesn’t all melt or freeze at once or at a single temperature.

As magma solidifies, crystals grow, and the concentration of the ingredients of those crystals is reduced in the remaining liquid. For example, MgO is removed from the liquid when crystals of the mineral olivine ([Mg,Fe]₂SiO₄) form. This process is called fractional crystallisation.

A key piece of evidence in Sigmarsson’s talk was the concentration of the element thorium (Th) in samples of ash and pumice from each Grímsvötn eruption of the 20^th century.  These show an increase in Th concentration over time.  This suggests that they all came from the same source beneath the volcano (the magma chamber), where growth of crystals of low-Th minerals results in an increase in concentration in the remaining melt. This melt was erupted in each of the events that produced the samples e.g. in 1998 or 2004, and was probably left over from the huge Laki eruption in 1783. The Grímsvötn 2011 samples contain much less Th than previous eruptions. This implies that it is a different magma: hotter, richer in gas, and closer to the composition of the original melt that formed in the mantle.

Melt migration into lava columns

Hannes Mattsson gave a presentation suggesting a previously-undescribed process within cooling lavas that explains the patterns in polished lava tiles (see image) as part of a study into the formation of columnar joints in basalt.  This work has already been published in the journal Nature Communications.  Joints are a system of cracks in a rock, and columnar joints break the rock into columns. Famous examples of columnar jointed lavas are the Giant’s Causeway in the UK, the Devil’s Postpile in the USA and Svartifoss in Iceland.  Mattsson and colleagues found that melt flows into the columns as they form.

Basalt tiles on the conference centre floor. These were quarried from columnar-jointed lava, and each tile is a slice through a column like a plank of wood. The pattern, which resembles the grain in wood, results from melt movement within the columns as they cooled.

Columnar joints form because lava contracts as it cools. Usually, the top of the lava, exposed to the air, cools most quickly and the cracks grow downwards into the flow. Experiments show that basalt lava shrinks by 10-15% when it solidifies, but the volume of the gaps created between the columns is much less than this.  Mattsson’s group thought that this is because more material flows into the columns as they form.  Sometimes you can see where finger-like parcels of lava have done this, but they are not very common, so another mechanism is required.

This mechanism is also based on the fact that magma doesn’t freeze or melt all at once.  As the lava cools, the first crystals to form are tiny, white, matchbox-shaped grains of plagioclase feldspar.  The dark bands in the tiles are rich in a black-coloured mineral called titanomagnetite, which is the last mineral to start crystallising.  If lots of plagioclase crystallises, then the crystals can lock into each other and form a network with some strength.  By this point, the lava is effectively rigid, but the remaining melt is still free to move through the gaps between the crystals.  It is equivalent to sucking coffee through a sugar cube.

Titanomagnetite, as the name would suggest, is magnetic. To test their hypothesis, Mattsson and colleagues measured the magnetic field across the column. It showed that the titanomagnetite grains were all lined up parallel to the column’s edge, instead of begin randomly orientated, or lined up with surface of the ground.  This demonstrates that melt flows into the columns from the molten interior of the flow, and seems to confirm their idea.  The different colours of the individual bands could be explained by different pulses of melt moving in.

If you visit Iceland, you will surely see these patterns, as the same tiles are used in the floor of Keflavík airport. Now you know what they are.

Other highlights

The conference was opened by the president of Iceland. He knew how to play to a room full of geoscientists. In his address, he said that seven-day creation myths are clearly false because Iceland was still being created. I imagine that he is equally unimpressed by stories about walking on water; in Reykjavík, people play football on the city pond.

• Thor Thordarson explained that subglacial basalt eruptions may release less sulphur dioxide than their subaerial counterparts because quenching of the magma by meltwater can freeze the gas inside the pumice and ash.
• Anja Schmidt looked at how tiny sulphate particles formed from volcanic sulphur dioxide would be dispersed across Europe by another Laki-sized eruption. Using equations linking air pollution and cardiopulmonary fatalities, she predicted that around 140,000 extra deaths could be caused in Europe (depending on weather conditions).  The full study was published in PNAS last year.
• Guðrún Gísladóttir looked at sediment accumulation in western Iceland in lake and soil cores.  You can easily spot the time when Iceland was settled around 900 A.D. because there is a distinctive ash layer from an eruption at the time.   Before settlement, Iceland’s largest herbivore was the goose. The arrival of tree-chopping men and their vegetation-munching livestock coincides with a big increase in wind-blown sandy material in the cores as the plants that previously held it in place were stripped away.
• Rikke Pedersen concluded that the best way to explain a donut-like ring of uplift around Hekla between 20 and 40 km from the crater is that material is being added to a magma chamber deep below the volcano, while the upper part is subsiding under its own weight.
• Halldor Björnsson used webcam footage from Eyjafjallajökull and point-tracking computer software to study how the way that the plume rose above the volcano. These results will feed into improved understanding of how ash clouds form. This study is another great example of quality science being done using public data and a bit of programming skill.
• Sean Pyne-O’Donnell reported on the identification of ash grains from Alaska and the Cascades in peat bogs in Newfoundland, showing that ash from very large eruptions can travel great distances. He suggested that some ash grains found in Europe, and which have no chemical match with known Icelandic eruptions, may have come all the way from the Americas.

Ash from Iceland fell in the UK

My own talk showed that the recent eruptions of Eyjafjallajökull and Grímsvötn both deposited ash in the UK. It included a map of where Grímsvötn ash was found in the UK based on the tape samples sent to the British Geological Survey after our call last May. Thanks again to everyone that sent them in. The talk made the front page of the local paper, which was exciting, even if all the quotes about other European locations that were attributed to me actually came from someone else.

Acknowledgements and caveats

I am grateful to Marie Curie Actions, who provided funding to allow me to attend this meeting. All of the information presented above is based on my understanding of the conference and is therefore only as accurate as my note-taking on the day.

Note: 02 December 2011.
The current media interest in Katla does not stem from a recent change in activity at the volcano, but from an article published on the BBC website today.  The same thing followed a Guardian article earlier in the year.  Activity at Katla is still elevated, as it has been for six months already.  There is no new evidence today that an eruption is likely very soon.  In geological terms, imminent could mean weeks, months, or maybe years.  This post was written in early November.

Katla rumbles on.

The unrest began in the summer, when a small flood broke out from beneath the glaciers and washed away the bridge over the Múlakvísl river.  Since then, there have been a number of small earthquakes every day.  This is more activity than usual, but less than, for example, occurred at Eyjafjallajökull in the weeks before the 2010 eruption.  It is not clear where this is heading, but Iceland continues to prepare for an eruption and local towns have been running evacuation drills.

The remains of the bridge at Múlakvísl that was washed away in the flood of July 2011, with the replacement bridge in the background. Note the yellow earth-moving equipment high on the hillside. They aren't parked up there for the view.

While the international media focus on the potential ash cloud and ‘travel misery’, the real destruction in Iceland will be caused by meltwater floods.  Katla is covered by glaciers up to 500 m thick.  These represent a huge reservoir of water, waiting to be unlocked by the heat of an eruption.  A look at floods from the last time round, in 1918, gives an indication of what could be expected.

The jökulhlaup from the 1918 eruption

There is a word in Icelandic for such floods from beneath the ice: jökulhlaup.  Translated directly into English, it means glacier leap.  This seems appropriate, as the water bursting under, over and through the ice tears off huge chunks and carries them suddenly forward down the valley. When the water subsided after the flood of 1918, the plain was strewn with giant icebergs, up to 60-80 m high.

Eyewitnesses said the speed of the Katla 1918 jökulhlaup was “so great that a fit man could not have avoided it”.  Escaping from the eastern side of the glacier, the flood wave reached the ocean in 45 minutes.  These reports correspond to a speed of 10 metres per second (36 km per hour).  The high-water mark of the jökulhlaup is recorded on the slopes of small hills along the way that were scoured clean by the swirling torrent.  Near the glacier, the waters peaked at 25 metres deep.

Haukur Tómasson combined velocity and the depth of the flood with the shape of the channel and came up with an estimate for the peak discharge of 300,000 cubic metres per second.  The total volume of water was estimated to be around 8 cubic kilometres.  The flooded area was as much as 700 square kilometres.  There was sufficient sediment in the flow to extend the coastline by 5 kilometres.

To put these figures in context, the average discharge of the Mississippi is a relative trickle at 17,000 cubic metres per second!  The flood is equivalent to pouring out Loch Ness onto an area half the size of Greater London in less than 8 hours.

The jökulhlaup from the next eruption

To understand the hazard from future flooding, researchers in Iceland used computer models to simulate what would happen if another Katla 1918-sized flood was to occur at the volcano.  They found that floods would reach the roads to the east of the volcano in 1-1.5 hours after the eruption began; a very short window to get people to safety.  Few people live in this region, but destruction of the road would be a major blow for the tourism-based economy of the area.

Iceland is a country about the size of Ireland.  For 9 months of the year, you cannot cross the middle as the roads are blocked by snow.  Highway 1 is the tarmac ring-road that runs around the country near the coast.  If it is destroyed, it means that you can no longer drive directly from Vík in the south, to Höfn in the south east (271 km, about 3 hours).  Instead, you have to go all the way round the other way (1068 km, about 13 hours).  A severe jökulhlaup could close the road for months.

The models showed that a flood travelling westward would sweep across the farmlands of the Landeyjar district, which is home to around 600 people, within 3-10 hours.  Homes, farms and livestock would be destroyed.  Fortunately, investigation of deposits from past jökulhlaups in the region suggest that this area is flooded rarely, perhaps only once every 500-800 years.

Predicted times for Katla flooding from 1918-sized jökulhaup. Click to enlarge. Source: Gudmundsson, M. T., G. Larsen, Á. Höskuldsson, and Á. G. Gylfason (2008), Volcanic hazards in Iceland, Jökull, 58(58), 251-268.

Summary in a single picture

Sources: Amazon map – Natural Earth data; Iceland map – atlas.lmi.is; Iceberg image – Larsen, G. (2010), 3 Katla: Tephrochronology and Eruption History.

Around 150 million years ago, plankton floated in warm seas.  Using energy from nuclear reactions in the Sun, they built their bodies from protein and fat and carbohydrate.  Then they died and their bodies sank to the muddy sea floor.  Over time, the mud was buried, compressed and cooked, and the plankton bodies became an oily liquid.  Much later, the liquid was pumped, refined and sold.

Now the liquid is exploding, releasing heat and gas.  The gas expands and turns a turbine.  The turbine sucks air through the engine and pushes it out into the atmosphere behind.  The atmosphere pushes back with an equal force, driving forward the engine and the wing that it is attached to.

The wing has an asymmetric shape that Deflection of air by the wing forces the air passing over it travel faster than the air that passes below.  The air below presses on the wing with more force than the faster-moving air above, lifting up both the wing and the metal tube to which it is attached.  The metal tube, and the seats inside, travel at 900 kilometres per hour across the ocean.

And the ocean, which is over 3000 km wide and averages ~4000 m deep, didn’t even exist back when the plankton were catching the rays.

Map of the age of the ocean floor. Most of the floor of the Atlantic Ocean is less than 100 million years old. Much of the source rock for North Sea oil was formed in the Jurassic (145-200 million years ago). Image from National Oceanic and Atmospheric Administration (http://www.ngdc.noaa.gov/mgg/ocean_age/). Click for for large version.

EDIT: 10 N0v 2011 – altered wing lift mechanism.  See comment below.

Geomorphology is the study of the formation of landscapes, and may not seem immediately relevant to the farmyard.  But heaps of wet grain and mountains of fractured rock are shaped by the same processes.  It’s all just a question of scale.  This post illustrates some of the similarities.

Wet wheat behaves like fractured rocks

Freshly cut wheat, piled up into heaps as it waits to be dried, is a common sight across Scottish farmyards this September.  The grain will be fed through the grain drier, which blows hot air through it (like a huge, diesel-gulping, hair drier) and reduces the moisture content to less than 15%.  The drying process prevents the grain from germinating or going mouldy during transport and storage.

In a good year, minimal drying is required as the sun does most of the hard work while the crop is still standing in the field.  This summer has been rubbish, however.  There have been only short dry windows between bouts of rain, meaning that lots of wheat had to be harvested before it could dry out naturally.  It has left farmers cursing and big heaps of wheat with up to 30% moisture standing on the farms.

As grain is taken, bit by bit, from the heap to the drier, the geomorphological connection becomes clear.  Wet wheat is slightly sticky, so the grains do not flow past each other very easily and the pile can behave as a solid whole.  This effect is especially apparent if the grain has been sitting for a day or two.  If vertical walls remain as grain is removed from the edge of the pile, farmers know that the grain will need a lot of drying.  The moisture content is a strong control on the behaviour of the pile, and the size of walls that it can support.  This is demonstrated in the picture below.

Grain heap ~3 m tall. The grain was cut from a number of different fields on different days and the moisture content varies from 24-28% on the left of the image, to 20-24% at the right. The ability to support vertical walls decreases with decreasing moisture content.

The wheat walls are not stable for long and are soon sculpted into cliffs and valleys and ‘scree slopes’ (also known as talus ramps in countries where they eat tomaytoes) of loose grains by a series of small landslides (or wheatslides?).  The results are strikingly similar to real-life mountain landscapes.  Given sufficient time, all the cliffs will collapse to form a pile with uniform, smooth sides.  The angle at which the loose grains are stable is called the ‘angle of repose’ and is higher in the wet grain than the dry.

Beans means bedrock

The collapse of the grain heap is interesting because it demonstrates processes that take thousands of years in real mountains, but in the space of a few minutes or hours.  The farmyard is like a large-scale version of an experiment published by Alex Densmore and colleagues in 1997 in the journal Science, and which is now carried out in undergraduate labs each year in order to demonstrate erosional processes.  They built a perspex box with a wall that could slide up and down,  and filled it with elongate, red beans, which simulated the bedrock in mountainous regions.  As they lowered the sliding wall, the beans spilled out over the edge.  At each stage, they weighed the beans and traced the profile of the bean surface on the transparent side of the box.

They found that the beans were not lost continuously, but fell in bursts in a series of unpredictable landslides of differing sizes.  Steep slopes could develop and last for a short time, but would be swept away in bigger events.  In fact, the vast majority of material (70%) was removed by just a small number (10% of the total count) of large events.  Densmore and colleagues concluded that the variety in landslide size was down to the existence of regions in the pile where the beans were aligned and which behaved as coherent lumps.  When they repeated the experiment using white beans (which have a more spherical shape), they found that the landslides were more regular and evenly-sized.

A closer view of the grain pile. Internal structure, formed when the pile was pushed up, is visible as aligned grains form light and dark bands. Collapse of the vertical walls leads to formation of 'scree slopes' of loose grains.

In real-life mountains, the sliding wall of the perspex box represents the lowering of the baselevel of a valley by erosion, perhaps by a stream or river.  The red beans represent bedrock that contains a network of fractures that are much weaker than the rock itself.  Such studies are important because landslides are a real danger in many parts of the world (there were a number in Nepal and India following an earthquake in the area on Sunday), and if we can understand how they work then the areas at greatest risk can be identified and avoided.

Beans versus wheat, wheat versus mountains

In the grain heap, the lowering of the baselevel is controlled by a forklift taking wheat from the foot of the slope.  Instead of a gradual lowering, this happens quickly, then the grain pile has to catch up.  While the grains can align as the beans did, grain moisture is a key factor in slope stability and dry grain behaves more like white beans.  A few large landslides make much bigger changes than many small ones.  The erosion of the walls is helped by drying of the surface by the sun, or by the wind, or by grains from above sliding down scree chutes.

(a) Steep buttresses and scree slopes, grain pile, Fife, Scotland. (b) Steep buttresses and scree slopes, Tre Cime de Lavaredo, Dolomites, Italy.

The landscape in the Dolomite mountains of Italy bears an uncanny resemblance to the grain heap.  There, the mountains are adjusting to lowering of the baselevel by glacial erosion during the last Ice Age.  The rock contains different layers 20-200 cm thick, that make up a network of planes of weakness and frost, sun, wind and rain erode it away.  Scree chutes run between the peaks and build up at the base.

Though these mountains are literally ‘as old as the hills’ the lessons learned from the small size, high-speed example of the grain pile gives a glimpse into their future state.  The area has been designated at UNESCO World Heritage Site for its spectacular scenery.  But if you want to enjoy it, you’d better hurry up.  In a million years, it will be nothing but a pile of scree.

Here for a limited time only: Cima Piccola, of the Tre Cime, Dolomites, Italy. The orange line is the classic rock climb, Spigolo Giallo. As with the 'cliffs' in the grain heap, it is just a matter of time until all this is just scree.

Meanwhile, back on the farm

Fascinating though all of this is, I’m sure that most farmers would much rather see two weeks of wall-to-wall blue skies, and sun-dried grain flowing into the grain store as smoothly as water, than ponder the fate of distant mountains.  Hopefully the weather will pick up in the next few days.

Last month, the Institute of Earth Sciences of the University of Iceland and the Iceland Glaciological Society organised an expedition to Grímsvötn to study the deposits of the eruption that took place there in May.  This post describes the journey to the volcano and the effect of volcanic ash and other debris from the eruption on ice and snow.

Driving on ice

Grímsvötn lies beneath the Vatnajökull ice cap, which is the largest glacier in Europe.  At 120 km by 90 km by 900 m thick, it is huge.  If, for some reason, you needed to hide the English Lake District, you could just slip it underneath.  There would be room in there for the Yorkshire Dales, too.  To get to the volcano it is necessary to drive across this expanse of ice, in our case approaching from the east.

Fortunately, driving on glaciers seems to be a relatively routine practice in Iceland.  Two vehicles, equipped with winches, planks, shovels, compressors and a host of spares and tools can cross snow-covered ice with each providing a backup for the other.  The other vital component is a set of massive, waist-high tyres.

The final stage before driving onto the ice is reducing the air pressure in the tyres, which spreads the weight of the vehicle and greatly improves their grip.  It may look like the truck is loaded with everything but the kitchen sink.  That isn’t exactly true, but the large white object in the trailer is a household chest-freezer.  This was brought to store samples containing ash-filled hailstones, which were produced during the eruption.  At 08:00hrs, we were ready to get on the ice.

Follow the black ashy road

Ash from the Grímsvötn eruption is scattered thinly over the eastern part of the glacier.  The dark coloured ash absorbs sunlight, heating the snow beneath and causing extra melting.  Thus the surface of the snow, normally flat, is currently very uneven with curving ridges and smooth, wide valleys around 50 cm deep.  In places they resemble frozen waves on a choppy sea.  Crossing them is equally rough, and the maximum speed was around 10 km/hr.

The eruption took place in the late spring, so there has been little new snow to bury the ash.  Where it was blown into the tracks of vehicles that crossed the glacier back in June, it defines a vague road of long, dark, parallel trenches caused by the extra melting.

Snowfall accumulation and melting

At a seemingly anonymous point, defined only by GPS coordinates, is a snow monitoring station.  It uses echo-sounding the measure the distance from the sensor to the snow surface.  In an average year, 6 m of snow fall onto the glacier in this area.  Of those, 1.5 m melt and evaporate away; the remainder compacts and joins the rest of the ice flowing slowly to the lowlands.  We stopped to adjust the height of the sensor and to download the data.

Thicker tephra and dirt cones

Tephra is the technical term for all the ash and pumice and rock fragments that are thrown out by an exploding volcano (strictly speaking, ash only refers to material less than 2 mm in diameter).  Closer to Grímsvötn, the tephra got gradually thicker and coarser.

Once the tephra thickness gets more than a few centimetres, the effect that it has on the underlying snow changes.  While the black tephra still absorbs extra sunlight, if the layer is thick enough then the heat isn’t transmitted through to the snow.  Instead of enhancing melting, the underlying snow is protected.  The result is a rough landscape of dirt cones.  While the cones may be a metre high, the tephra only accounts for a black, sandy skin over their snowy cores.

Approaching the huts – Plan A

We were aiming for three wooden huts built on a ridge of rock called Grímsfjall, that pokes out above the ice.  The normal approach is to drive up the slopes from the south, as the direct route from the east is blocked by a series of large crevasses.  It appeared, however, that this route was also impassable.  Where the tephra was even thicker (10-15 cm), it gave more uniform protection to the snow beneath, resulting the a series of large plateaux about 2 m taller than their melted-out surroundings.

After driving back and forth along the foot of Grímsfjall and scanning the slope with binoculars, we could not see a way through.  Potential leads were explored on foot, but they they were dead ends, finishing in steep dirty walls.

Approaching the huts – Plan B

It was decided that the best approach may be to try a bit further east, where the tephra was thinner.  Driving up the slopes, things seemed promising as the black piles gave way to more open snow fields.  We progressed steadily higher until CRUNCH!  The front of the truck fell into a crevasse.  These deep cracks form in the glacier where the ice accelerates down the slope and they are covered in snow-bridges that hide them from view.  As the truck drove onto it, the bridge collapsed and it dropped down to its axle.

Removing the truck was suprisingly straightforward.  Jump out, clear the snow from the front, set up a jack on the far side of the crack (which was only about a metre wide), jack up the truck, reverse out.  It took less than 20 minutes.  We looked for an area to the side where the crevasse was narrower then continued upwards.

CRUNCH!  Another crevasse.  We were going to need another plan.

Approaching the huts – Plan C

Where two high ash-covered platforms met, there was sometimes a notch in the wall.  The notches were too narrow to drive through, but with a bit of shovelling, they could be enlarged sufficiently to allow the trucks and trailer to pass.  A route was staked out on foot, then the gateways were dug through the snow.  Four breaks were enough to give us access to the huts.

The Grímsfjall huts

Once through the worst of the tephra platforms, it was easy going to the crest of Grímsfjall and the huts.  We finally arrived there at 20.30 hrs.  With good snow conditions, the journey across the glacier can take just two hours.  It had taken us over twelve.  The difference was due to the tephra on the ice.

Some observations from the fieldwork are described in the following post (Grímsvötn 2 – What was in the plume?) and there are some more vehicle-themed photos in another (Grímsvötn 3 – Bonus truck pictures).

The May 2011 Grímsvötn eruption blasted ash and pumice and rock fragments (collectively known as tephra) through the Vatnajökull glacier, forming a massive plume up to 20 km tall.  It was the biggest eruption in Iceland since Hekla 1947.  Locally, ash rained down across southern Iceland turning day into night, while the finest grains were swept across the Atlantic to be deposited at least as far afield as in the UK.  There are nice images of the plume and its consequences at The Boston Globe website.

Last month, the Institute of Earth Sciences of the University of Iceland and the Iceland Glaciological Society organised an expedition to Grímsvötn to study the deposits of the eruption.  Following an adventurous journey across the glacier to the volcano (see Grímsvötn 1 – Crossing the glacier and Grímsvötn 3 – Bonus truck pictures), it was possible to see the deposits of the eruption, and thus what was in the plume, at first hand.  This post shows what was erupted and explains some of what the deposits can tell us.

All this should be snow

The dispersal of volcanic ash is controlled by the wind.  During the May 2011 Grímsvötn eruption, southerly winds blew the top of the plume northwards, while northerly winds blew the lower part to the south.  It was the lower part that contained most of the tephra that was deposited near the volcano.  Tephra is the technical term for all the ash and pumice and rock fragments that are thrown out in an explosive eruption (strictly speaking, ash only refers to material less than 2 mm in diameter).

This view, looking south from across Vatnajökull, should be dazzling white snow and ice all the way to the coast.  Instead, it is black: a vast dark plain of pumice and ash extending tens of kilometres from the crater.

Calculating the erupted volume

Digging through the black surface reveals the bright snow beneath.  This makes it easy to measure the thickness of the tephra layer in different locations.  A simple, but very important, question can be answered from the resulting map of tephra thickness: how much material was erupted?

At this location, if you keep digging, you will pass nothing but snow, ice and ancient tephra layers for over 700 metres before you finally reach the bedrock.

Closer to the crater, the tephra gets thicker and contains coarser, gravel-sized, pumice grains.  At this site 8 km downwind of the crater, the deposit is nearly 2 m thick, and the hole took eight people over an hour to dig.  It is clear why lone murderers favour shallow graves.  A layer of ashy-hailstones that fell during the eruption has refrozen into an icy layer near the base.  These were collected and transported back to Reykjavík in a freezer.

Other things that tephra can tell us

The deposits mainly contain ash and pumice: broken-up, bubbly rocks.  The ash is fragmented pumice and looks like black sand or grit.  All of this rock was hot enough to be liquid magma just moments before it erupted from the ground.  A detailed look at some of the grains can tell us more about the eruption.

The first photo shows tephra from a layer that was full of smooth brown spheres called accretionary lapilli.  If you cut one open, you find concentric rings (like in a gobstopper or an onion) of very fine ash grains.  These form as the ash is swirled around in the turbulent plume.  Helped by moisture, fine grains stick to the outside of the growing lapillus, building it up layer by layer.

It is important to understand accretionary lapilli because if these fine ash grains are sticking together and falling out onto the glacier then they aren’t drifting off downwind to bother European airports.

The second photo shows a piece of golden-coloured pumice.  It is very lightweight and contains millions of tiny bubbles.  This is unusual for basaltic tephra, which commonly has only a few, large bubbles.  Bubbles in volcanic rocks form when gases dissolved in the magma are released (exsolved), usually in response to decreasing pressure as the magma rises up from depth.  Thermodynamically, it is much easier for exsolving gas to join an existing bubble than to form a new one, so pumice with lots of tiny bubbles tells us that the gas was all trying to get out in a hurry.

The golden pumice therefore means that the magma rose very quickly from deep beneath the volcano.  This is consistent with the very intense eruption.  Geochemists can measure how much gas is still dissolved in the rock, and how much is dissolved in material trapped inside crystals that formed at depth.  From this, they can estimate the depth at which the magma was stored before the eruption, and how much gas (such as SO₂) was released.

The southern crevasse field

Great volumes of ice near the crater was melted during the eruption.  Since then, the glacier has flowed back toward the crater, producing a network of crevasses on the surface.  Unlike crevasses in a more alpine setting, tephra has fallen into and filled these ones, allowing the area to be explored in relative safety.  Unlike crevasses in a more alpine setting, these ones have steam coming out of them.  Just a few centimetres down, the tephra pile here is still warm.

The long, tall walls produced by the crevasses are a volcanologists dream, as they expose all the individual layers produced by different stages of the eruption.  These can be traced and measured over a wide area without the need for any digging whatsoever.  Here, the deposit is rich in pale-grey, angular, dense rocky chunks, 10s of centimetres in diameter.  These were cold pieces of old lava flows or other parts of the volcano that were ripped up and spat out during the eruption.  They are heavy, and rained out from the plume close to the vent.

A view of the crater area

Looking across the crater area. Click for bigger version

The photo shows the crater area, looking from the west.  The crevassed area in the foreground, with all the exposed tephra layers is clear.  The ridge on the skyline is Mt Grímsfjall, and the huts are located at the far end.  The cliff is about 200 m high.  The eruption began along a 1.5 km fissure running parallel to the cliff, before focussing on a few craters.

The lower, flat, area of ice sits mainly on top of the permanent subglacial lake, Grímsvötn, which is kept from freezing by geothermal heat at the base.  The water drains periodically from Grímsvötn in floods called jökulhlaups.  The lake in the foreground has formed since the May 2011 eruption by surface water flowing in, and has flooded the site of the craters.  The black area at the far end of the lake is not a beach, but a raft of floating pumice stones.  The vertical ice cliffs are capped with tens of metres of tephra, and sometimes come crashing down into the lake.

It is a spectacular place.

The journey to Grímsvötn is described in the previous post (Grímsvötn 1 – Crossing the glacier) as well as the effect of the tephra on the glacier.  The following post (Grímsvötn 3 – Bonus truck pictures) describes the difficulties of working on the tephra-covered glacier.

Last month, the Institute of Earth Sciences of the University of Iceland and the Iceland Glaciological Society organised an expedition to Grímsvötn to study the deposits of the eruption that took place there in May.  This post describes some of the difficulties in working on the debris-covered glacier.  There are other posts describing the adventurous journey up to the volcano (Grímsvötn 1 – Crossing the glacier), and the huge volume of material produced during the eruption (Grímsvötn 2 – What was in the plume?).

Hiking in the clouds

The journey across the glacier had been slow, as patchy cover of volcanic-debris and uneven melting had created a maze-like landscape of ridges and platforms on the surface of the glacier.  It was decided that the best way to get to the interesting geology would be to hike the 5 km from the huts.  The cloud was down, it was 2 degrees C, damp and windy.  The loose pumice and ash were soft underfoot.  Visibility was 20-100 m and navigation on the featureless terrain was by GPS.  The hike took 3 fairly-miserable hours in each direction.

It was decided that the best way to get to the interesting geology would be to somehow get the trucks through to where the surface was more continuous.

Building a road

The following afternoon, a ‘road’ was constructed.  Narrow ridges were dug through, holes were filled with snow, sharp edges on walls were softened to form ramps.  It was hard work.  Fortunately, the definition of ‘road’ for these trucks is quite broad.

The road meant the we could get to the interesting region in 1.5 hours, and arrive dry and warm, ready to work.  We could also move quickly between sites where the surface was smooth.  We used it for the following days, and scenes like the one below became almost routine.

Welding in a blizzard

During the days that we worked on the glacier, the snow was melting.  By the time we were ready to leave, it had lost about 10 cm, weakening the snow bridges that we had used to cross crevasses on our way up.  On the homeward journey, a bridge collapsed beneath the front left tire of the big Ford F350 and it dropped to its belly.  We couldn’t get it out using the jack alone (as we had done previously) so the Hilux came round and winched it out.  A close look showed that the axle had been broken.  We had a spare, so it was changed there and then.

Further inspection showed that part of the suspension had been damaged, too.  It turns out that we also had a generator and welding equipment, so that was fixed there and then as well.  On a glacier, in a blizzard, in less than two hours.  It takes a week to get an appointment for the annual MOT test at my local garage!  It seems that a mechanic is top of the list of things to bring on this type of expedition.

Bridging the crevasse

To finally cross the crevasse, it was ‘bridged’ with wooden planks, which spread the weight of the truck across a wide area of snow.  The same principle allows a skier to cross snow bridges that a hiker would fall through.  Once out of the crevassed area, the return journey was straightforward, but slow.  We reached the edge of the glacier 12 hours after setting off.

The outward journey to Grímsvötn is described in the post Grímsvötn 1 – Crossing the glacier, and there are descriptions of the results of the eruption itself in Grímsvötn 2 – What was in the plume?