After I wrote my post ‘Central volcanoes of Vatnajökull’ a discussion ensued on the Lakí fissure eruption, both about the anatomy and its timeline. In the end I offered to write a post on it. I not only armed myself by reading more than 30 scientific articles on the subject, I also delved into the largest unpublished scientific material existing for the Lakí lavas. I did this on explicit permission by a mining company, who also graciously offered to pay me for writing the article and publish it on Volcano café. This gave me the opportunity to go far beyond in regards to the research I always do before writing a regular post.
The data covers 8 314 surface samples, 42 drill cores and five stratigraphic dig sites with every sample having been through geochemical analysis. There is also complete geomagnetic sling data for the area and satellite data with laser mapping and high resolution gravitational anomaly data. I here want to point out that the data is collected for a purpose, and I am not allowed to give out data regarding concentrations on any Rare Earth Mineral (REM), nor am I allowed to publish a map showing where some members of the lavas are situated.
Here comes the difference, in science we freely publish our data, in the corporate world it is business secrets. In reality even after taking the restrictions into account the remaining data is far beyond what a scientific project could ever afford.
I will put in a list of articles that I have used in the end of part 3 and 4.
I highly recommend anyone who has not read ‘Central volcanoes of Vatnajökull’ to do so carefully, otherwise you will most likely be lost fairly quickly, see it as the prequel to this 5 part article.
Previous Lakí research
Up until 1993 research into the Skaftár Fires was based upon a model using the Krafla Fires as a type case. This has proven to be a large mistake. Sadly the image of a series of fissures opening up and closing one at a time, in turn drawing magma from a shallow central magma chamber resulting in fire curtaining and mild strombolian eruption characteristics was highly false. It in turn has caused climatologists to gravely misjudge the factors behind the climatic effects of a large rifting fissure eruption.
I will return to the climatic driving factors in part 4 and in part 5 I will return to the magma behind the eruptions. In part 2 I will go through the timeline for the Skaftár Fires eruptions and in part 3 I will talk about the central volcano of Grimsvötn and a bit about the other central volcanoes on the Grimsvötn Fissure Line.
First I would though like to talk about the name of Lakí; it is after all very misleading since Lakí never erupted. Lakí is an old laccolite, a geolithic artifact that is a remnant of old quarternary volcanic activity. The latest eruption was before glaciation. During the eruption it was too hard for the opening fissures to penetrate, something that in the end had a bit of consequence for how the eruption unfolded. So, I prefer the Icelandic term of Skaftár Fires. The main reason is that the Icelandic name is nicely plural.
I will come back to this in detail in the next part, but the Skaftár Fires consisted of at least 10 fissure eruptions and at least 5 other eruptions.
Anatomy of an eruption
All of the fissure eruptions are sufficiently alike each other that I can start off with describing one and just note the few differences in the next chapter. I have for practical reasons chosen the first eruption as my type case.
For those of you who remember the Krafla Fires eruption, strike that image out of your brains, as we will see the Skaftár fires was a much more violent affair going through several more stages.
At the time of the onset of the Skaftár Fires the tectonic rift zone running between Myrdalsjökull (Katla) up to and including the western part of Vatnajökull had not suffered a fissure eruption since 1477 Veidivötnahraun eruption. In the zone there are 3 large fissure swarms that have had large fissure eruptions.
Basically one can say that the tectonic rift zone is being pulled apart with a speed averaging at 28mm per year, during non eruptive times the spread rate is lower, about 17mm annually as shown by GPS measurements. One can look about the area as a large rubber band that is pulled apart until it snaps, and as it recoils the rubber band moves much further than the pent up strain. At the sites of the fissures the movement was much larger than the pent up 8.6 meters. It is the phrase “on average” that bites us here. The actual fissure width was between 50 and 150 meters at the surface, and that is not even half the truth as we will see later.
In the beginning
What we do not know is what set off the eruptive phase, and we might never really know. It could just have been the rubber band snapping, but at the same time we know that the hotspot under Iceland is cyclic so it could have been a cyclic high (a pulse) that set off the rifting episode. Be that as it may, the rest we can infer from physical evidence and witness reports.
Early in the day on the 15th of May 1783 tremors was noticed in Skaftártunga. It was not anything unusual about that, the local priest though made a dutiful notation about it in his daily journal. The perceived tremor continued up until the early afternoon on the 29th of May. Not even the duration of the tremoring was really that noteworthy for a hardened Icelander.
Nowadays we know what was happening. It was a large scale earthquake swarm running thousands of small earthquakes per day ranging up to 3M. As an earthquake swarm it was fairly large, but nothing that does not happen every ten years or so in Iceland.
In the end it was not a normal earthquake swarm, or even a normal pre-eruptive earthquake swarm. It was the tectonic strain closing in on the point of a critical fail. And that critical fail started on the early afternoon on the 29th of May as the earthquakes suddenly got much bigger. Due to the distance of the witness reports we know that the earthquakes had a minimum of 5M, probably up towards 6M since they started at the bottom of the crust and we know that people had to live outside in tents due to damages to the houses. During the next ten days the earthquakes opened up a wedge shaped fissure from the bottom of the crust, as extremely gas rich magma started to travel upwards the wedge widened at the bottom and more magma poured up. By now the eruption was a self sustaining machine rapidly hammering its way up towards the surface.
As the magma rapidly rose upwards at an average speed of at least 3 km per day it started to degas at an ever increasing rate, and since the volume was very large and the magma gas content was so high a bow wake of gas preceded the magmatic ascent upwards. I will return to how we know about that gas front wave in part 4.
At 09.00 on the 8th of June the bedrock above the rising magma and gas become too thin to be able to withstand the extreme pressure of the rising magma and gas. To compound matters even more the top layer had a water table ranging from 150 to 300 meters depth, mainly consisting of bogs and small lakes.
As the compressed superheated sulphuric and fluorine gas punched through the bedrock the combined force of the hydrothermal explosion, the phreatic detonation and kinetic gas release pulverized the bedrock with tremendous force in an explosion that lasted for about two days. The larger debris and medium sized rubble rained down over the surrounding area, and the coarser sands fell over Skaftártunga and the stoic priest noted that another of Iceland’s eruptions had started. The force was though energetic enough to literally pulverize most of the overlaying bedrock into exceedingly fine dust. The length of the first fissure was 1.6km.
On the 10th of June magma was spotted for the first time as it came flowing out of Skaftárfljót. This first magma was not in an overly thick layer and came out of a fire curtain 1400 meters high. As the fissure continued to widen the height successively lowered as pressure dropped in the system, but the magma had barely started to arrive. The main bulk arrived during the night of the 11th of June, it quickly filled up the river gorge of the Skaftár river gorge and then moved onwards at a speed of up to 3 km per hour.
At that time a second fissure had already opened ejecting a fire curtain and a third was about to open up. For every opening fissure the pressure dropped at our initial fissure and after a while the main activity from it was quiet effusion of a steady stream of lava and strombolian activity at the cones that had formed. The activity continued like this with diminishing activity as every new fissure opened up, but the effusive activity at the fissure did not stop until the 7th of February 1784, and the lava flow did not halt until April of the same year.
As the eruption continued the lava changed that came out of the fissure and eruptive layer evidence give at hand that the temperature of the melt increases as time went by. The magma that came up after the fissure surge tended toward olivine, in the end almost pure foersterite, an extremely high melt point olivine member, was erupted. The bottom layer is almost exclusively blasted bedrock, but over that you have an intricate layering between ashes, lavas, finer grained bedrock as the fissures opened further and further away, and strombolian artifacts. One interesting feature is that some of the fissures produced large amounts of Pele’s hair and Pele’s tears.
At least two fissure openings had a minimum columnar height of 12 kilometers. We know that due to eyewitnesses stating that they could see the columns from a great distance. Problem is just that the previous calculations pretty much did not feel confident with stating higher columns. As we will see in part 4 this was wrong.
I previously mentioned that the magma was very gas rich. We know this from the amount of vesicularisation (fancy word for bubbles of gas in the rock) of the lava. The lava that ended up at the outer of the lava floods cooled rapidly as it contacted with cool rock and retained the high vesicularisation. The lava that cooled slower lost all of its vesicularisation (bubble free). This caused the lava to lose half of its volume, and then it had lost a lot more already during the ascension due to degassing.
One should note that the eruption created a series of Grabens along the entire stretch of fissures. A Graben is created as the area around a large fissure eruption settles due to the area below have emptied out all available magma. You could in a way say that it is a non round caldera formation through subsidence. Most of the Graben line was covered by the magma, but in many places the Graben structures are evident. We should remember that not only did Lakí erupt 15.1 cubic kilometers of lava (DRE), it also erupted tephra (0.8 to 1.9 cubic kilometers DRE) and blasted a far larger amount of rock out. To further compound problems we have all that area evacuated as the wedge formed from the mantle upwards to the surface. So, in reality the entire magmatic emplacement was more into the hundreds of cubic kilometers if we count all the magma going up into the crust. No wonder we got a Graben formation when all that magma cooled and shrunk.
No evidence has been found of more evolved magmas in either the ashes, tephras, or in the lavas. Instead all of the samples point to unevolved fresh magma arriving at high temperature directly from the mantle/crust boundary, also the high gas content points to such an origin. There is no sample pointing to magma having either encountered previous melted material, or rested in a magmatic chamber of any sort. This disqualifies any central volcano having taken any part in the Skaftár Fires. In the next episode we will discuss briefly the eruptions at Grimsvötn and the other central volcanoes on the fissure swarm during the time of the Skaftár Fires, and in part 3 we will do a deep dive into that particular subject.