Understanding Katla Volcano in Iceland

Recent seismic activity in the region of Katla has led researchers to believe it will erupt again. In order to understand the potential hazards of new eruptions, researchers need to understand how these eruptions historically behaved. In their new paper titled Large Explosive Basaltic Eruptions at Katla volcano, Iceland: Fragmentation, grain size and eruption dynamics, Schmith et al. (2018) investigated why this volcano is so explosive. In order to do this, they used size, shape, and composition of the ash fall deposits from the eruptions in the 17th and 18th century.

Katla, one of most hazardous volcanoes in Iceland, is located in the southeastern part of the country (Fig. 1 & 2). The volcano lies 1450 m above sea level, and the caldera covers an area of nearly 110 km2. Katla is topped by the 230 m thick Mýrdalsjökull glacier which covers an area of ~ 600 km2, which melts and causes lahars and flooding during eruptions. Katla has a history of major explosive eruptions that occur on average about every 50 years. These eruptions are sometimes large enough that ash falls on mainland Europe, including during the 1625 and 1755 CE eruptions.

The 2010 Eyjafjallajökull eruption in Iceland was small compared to past eruptions of Katla, yet airspace in 20 countries was closed for 5 days to commercial air traffic, ~ 10 million travelers were affected, and the world’s average temperature fell by 1 ℃. A future Katla eruption is predicted to be ten times larger than the 2010 Eyjafjallajökull eruption. Heat from Katla causes glacial outburst floods when parts of the Mýrdalsjökull glacier melt and breach the crater of the volcano creating large lahars. During the last major eruption of Katla in 1918, the lahar produced was so large that it extended Iceland’s coastline in the region by 5 km (Katla’s Hazards, 2019).

Figure 1: A: Iceland with the region of interest shown in the red box. B: Zoomed in view of the region of interest showing locations of the Eyjafjallajökull, Katla, and Surtsey volcanoes in southeastern Iceland. Satellite imagery courtesy of Google Earth.

In order to understand the mechanisms of past eruptions and the extent of magma/water interactions, the researchers first turned to ash deposits. They evaluated the amounts of tachylite (heterogeneous glass), sideromelane (homogeneous glass), lithics (rock pieces), or crystals. By understanding the amounts of each of these are present in the ashfall, scientists can infer whether the explosivity of an eruption is more fragmentation/gas dissolution driven or driven by interaction of water and magma to cause explosive steam and lava eruptions.

Figure 2: Volcanic ash from previous eruptions found within the Mýrdalsjökull ice. Figure from British Geological Survey (Katla volcano, 2019).

Next, Schmith et al. (2018) determined the extent of fragmentation at the time of the eruption. As magma rises to the surface, gasses dissolved in the magma begin to come out of dissolution and form bubbles. As the magma rises further the bubbles expand more and more until there are more bubbles than magma, resulting in fragmentation. When the bubbles all finally pop, an explosive eruption occurs. This process is similar to opening a shaken bottle of soda. When you open the bottle the pressure is decreased and bubbles begin to form. The bubbles degassing rapidly cause the soda to explode. By understanding the fragmentation process of the Katla volcano, geologists can infer past and future eruption sizes.

Finally they determined the water/magma interaction of glacial meltwater of the 1625  and 1755 eruptive events. To this end they examined the size and sorting of the ash grains. A decrease in grain size and increase in the sorting of ash indicates a decrease in the ash column height.  A decrease in grain size with a decrease in sorting of ash grains indicates increases in fragmentation intensity (Fig. 3).

Figure 3: Plots showing grain size base on median size (φ) in each unit vs. sorting. Figure adapted from Schmith et al., 2018.

Overall Schmith et al. (2018) found that both the 1625 and 1755 ash deposits were composed mainly of the two glasses, tachylite and sideromelane, followed by crystals and lithics (Fig. 4), which indicates that the eruptions were not water driven.

Schmith et al. (2018) found that 1625 CE eruption was a combination of magmatic degassing and water/magma interaction (phreatomagmatic) and  that the water/magma interaction generally decreased over the course of the eruption. However, toward the end of the eruption, the water/magma interaction became the larger driver of the eruption over fragmentation.

Figure 4: Components of ash fall deposits for the 1625 and 1755 Katla eruptions. Figure from Schmith et al. (2018).

The 1755 CE eruption was similar to the 1625 eruption in that it was both a result of fragmentation and of water/magma interaction. However, in the case of this eruption the water/magma interaction was only dominant at the beginning of the eruption, and fragmentation dominated most of this eruption.

By understanding how water from the glacier interacts with lava and superheated material, scientists can use the data from this study in order to help predict potential mechanisms for future Katla eruptions. This will allow insight into how far the ash from an eruption might travel, and the extent of lahar flows and flooding. Understanding the dynamics of Katla and the Mýrdalsjökull glacier allows governments to plan for hazards resulting from a Katla eruption and put protections in place to help keep the residents of Iceland and mainland Europe safe.

References

Katla’s Hazards Katla Volcano, http://katlavolcano.weebly.com/katlas-hazards.html (accessed March 2019).

Katla volcano, Iceland | Volcanology | Icelandic volcanism | Our research | British Geological Survey (BGS), https://www.bgs.ac.uk/research/volcanoes/katla.html (accessed March 2019).

Schmith, J., Höskuldsson, Á., Holm, P.M., and Larsen, G., 2018, Large explosive basaltic eruptions at Katla volcano, Iceland: Fragmentation, grain size and eruption dynamics: Journal of Volcanology and Geothermal Research, v. 354, p. 140–152, doi:10.1016/j.jvolgeores.2018.01.024.

GPS Used To Predict Height of Volcanic Plumes

 Photo of eruption Grímsvötn, May 2011. (SPR AFP/Getty Images) 

Scientists in Iceland have developed an innovative method to predict volcanic plume height using changes in magma pressure.  Using a series of anchored GPS stations, they were able to detect inflation of the Grímsvötn volcano before its eruption in May 2011 and the sudden co-eruptive subsidence.

These GPS stations, originally installed to monitor the movement of tectonic plates over time, continuously recorded position and elevation during the pre-eruptive and eruptive phases of the explosion.  They found that the volcano expanded radially approximately 20 inches in the hour preceding the eruption and sank 10 inches during the initial eruptive phase, resulting in tilting of one hundredth of a degree.  By tracking the change in volume of the volcano, the scientists were able to track the change in pressure of the magma chamber during the eruption.  This process is similar to a tire deflating – as air (or in the case of a volcano, magma) leaks out the tire pressure decreases. 

Based off this change in pressure, the total change in volume of the magma chamber was .027 km3, about 10x smaller than the total erupted volume. Pumice is extremely porous and can contain over 50% void space; this is a result of bubbles being trapped in the solidifying melt as they expand due to decreased pressure at the volcanic vent relative to the magma chamber. The scientists were then able to determine the rate of the eruption using the overall duration and amount of ejected material.  The mass eruption rate is directly correlated with the height of the plume; the faster the eruption rate, the taller the plume. Using this novel method, they estimated that the plume height peaked at 15 km, which compares favorably to satellite imagery.

Column height is an important parameter to predict where an ash cloud will travel.  The heat produced by jet engines cause the ash particles to melt, gunking up the engine and causing the plane to stall.  Currently, the protocol for avoiding this is to completely close airspace.   The eruption of Grímsvötn, which lasted for seven days, caused airspace closures in northern Europe, grounding approximately 900 flights.  Better predictions of column height would result in more accurate hazard maps and fewer plane delays. This new method to estimate plume height uses real time data that does not rely on satellites.  Aviation experts can then better seed their models to create smaller no fly zones due to more accurate predictions of ash movement.

Original Article: Hreinsdóttir, Sigrún, Freysteinn Sigmundsson, Matthew J. Roberts, Halldór Björnsson, Ronni Grapenthin, Pórdur Arason, Thóra Árnadóttir et al. "Volcanic plume height correlated with magma-pressure change at Grimsvotn Volcano, Iceland." Nature Geoscience (2014).

Pouring Lava on Ice: for Science!

Eruption plume above Eyjafjallajokull volcano, Iceland, April 17, 2010.
 In the foreground is the meltwater floodpath from eruption-induced ice melting.
Photo credit: Eyjoful Magnusson, University of Iceland. Source

In 1973, a new volcano erupted on the island of Heimaey, part of the Westman Islands off the southern coast of Iceland, and threatened the very existence of the inhabitants and their way of life as they knew it. The main town of the island, Vestmannaeyjar, became partially inundated with lava but the inhabitants knew they would be able to rebuild their town. However, the lava was also advancing toward the only harbor on the island, threatening to seal off the harbor and destroy this fishing community’s livelihood as well as one of Iceland’s main fishing port. The Icelanders came up with a novel plan to desperately attempt to save the town. The idea? Throw enough cold water on the lava in hopes of creating a solidified lava wall to divert the rest of the lava away from the harbor. It took two months of water pumping and the entire arsenal of pumping ships from the area (high-capacity pumps were even brought in from the US) but the harbor was eventually saved. This event is one rare case of a “successful” fight against an advancing lava flow; however, half of the town was still destroyed and a similar event could occur at any time.

Lava flows are not the only volcanic threat to life and infrastructure. Lava flowing over ice or snow can induce rapid melting, leading to an unexpected debris filled flood event. Volcanic eruptions are dangerous and need to be quantitatively measured in order to better understand the hazards associated with them. Specifically, understanding the interactions between lava and ice is critical in developing hazard assessments. We care about the interactions between lava and ice because of the rapid melting involved and the increased risk for flash flooding and debris flow events, such as lahars and jokulhlaups. However, because volcanoes are often in difficult to reach places and an eruption involving lava and ice is inherently dangerous, it is difficult to observe the lava/ice interactions closely. Moreover, a lava flow traveling over ice may melt sufficiently through it and begin flowing underneath and/or through the ice, where it is no longer observable from the surface.

In an attempt to address these observational and measurement difficulties, a group of scientists out of Dickinson College, Pennsylvania, in collaboration with Syracuse University, New York, have created a way to observe and measure the interactions of lava with ice and snow (figure 1). In short, their experiment is to pour basaltic lava onto layers of ice or snow in a controlled setting to observe and measure the interaction. Multiple experiments were performed using blocks of ice, shaved ice, and sand (to simulate an ash covered ice surface).

Images of the experimental lava flows advancing over ice and snow. The top image is lava flowing over an ice bed 12 cm thick (the width of the flow is 15 cm). The second image is lava flowing over snow 12 cm thick (the width of the flow is 20 cm). Images are from the original article. 

Experiments reproduced two end-member melting behaviors consistent with field observations: (1) flows that melted quickly through ice and (2) flows with slow initial melting. In experiments without a sand layer, lava advance was inhibited due to it quickly melting through and sinking into the ice. Eventually the lava melted through the ice to the container bottom and again flowed downslope along the container-ice boundary. These experiments show the ability of lava to exploit these crevasses to flow downward and through ice. By controlling the starting geometries for ice-lava boundaries we can test different hypotheses for ice confinement of lava flows.

One of the more interesting observations is that during some experiments the lava would appear to skate across the ice; this is thought to occur from the trapping of a vapor phase between the ice and lava (the ice transitions to steam, is trapped, and acts as a lubricant between the ice and lava). The existence of this trapped gas may also act as a buffer for heat flow between the lava and ice, which would melt the ice slower and allow the lava to advance further before sinking into the ice.

Formation of bubbly Limu o Pele surface in experimental lava flow. Liquid water and stream from the melting of the ice blocks below the lava is incorporated into the lava. The steam expands within the lava and forms large bubbles that then solidify, known as Limu o Pele. Image is from the main article.

Rapid and extensive formation of large bubbles (Limu o Pele, see figure 2) occurred within the lava during the no-sand experiments. This may be the first published demonstration of large-scale Limu production with water originating underneath lava. The abundance of bubbles that formed as lava flowed across the ice indicates that there may be a strong relationship between external water incorporation and the formation of Limu.

Additional reading: 

http://www.earthmagazine.org/article/do-it-yourself-lava-flow

http://lava-dev.syr.edu/research.html

            http://www.nsf.gov/news/news_summ.jsp?cntn_id=116833