A bit about the importance of data-plots
It is now many years since I did any heavy scientific lifting. But, the fun part of writing for Volcanocafé is that I get to do quite a bit of work in the field of volcanology in a relaxed popular science setting.
It is a rather well-known secret that I lean on the amazing plotters that we have had through the years, all the way from the grandmaster of Geolurking, to the latest plotter in the form of Andrej Flis. Sometimes it all starts with me having an idea that I want tested against data, and other times it is the plotter that comes with a plot that shows something odd.
Through the years these marvellous plots have shown things hitherto unknown about Icelandic volcanoes. So far, we have found unknown magma reservoirs under volcanoes thought to be either extinct or highly dormant. We have found that at spread centre volcanoes there are often wedge-shaped dykes working as deep magma reservoirs.
My favourite is though that I got to test a more profound theory against real time data. Back in 2011 I had an idea that we would be able to track the Icelandic mantle plume magma pulse that comes around roughly every 200 years. That pulse heralds both an increase in the frequency of eruptions, and is also a sign that there might be a large rifting fissure eruption around the corner.
My idea was very simple, any pulse from below would be visible in the form of a marked increase in deep earthquakes in the crustal layer next to the boundary between the crust and the upper mantle. My mental image looked something like a soft pillow hitting the underside of a mattress in slow motion, with an epicentre close to the location of the plumehead.
In 2013 we started to see that increase in deep earthquakes, and from then on it was a case of chasing the earthquakes to try to see where the first eruption would strike. A few weeks prior to the start of the Holuhraun eruption I knew that the eruption would come from Bárdarbunga. It was the first time that an eruption had been accurately forecasted from a hypothesis that had not been tested, and it was predicted a year prior to onset (obviously the forecast was updated as new data came in to pinpoint the volcano).
My point is though that none of these things would have been possible without all the marvellous plots that are produced by the people behind Volcanocafé. It is through all the data-wrestling that we can reliably say things. Any hypothesis is just another wild idea until it is tested against the raw data given to us by nature.
In the case of Holuhraun we had a wild idea by me tested against data, then the idea was refined into a hypothesis, and later it was proven correct by a rather stupendous amount of magma. It still gives me a yummy feeling of science in my stomach.
Andrej strikes again
A couple of weeks ago I started to rewrite my woolly mammoth series about Icelandic volcanism, and I partially rewrote the part about the Reykjanes Ridge volcanism. In particular I wrote about the Eldey-Geirfuglasker Volcanic Belt.
The next day I was sitting down to plan for the next part of the series when Andrej sent me new plots that he had made off the Eldey-Geirfuglasker area. I must admit that I groaned a bit as I looked at those plots, since I was in complete shock. I went from being rather pleased with myself to realising that I had just published an article that was to be thrown unto the garbage-pile of extinct science.
But, since I am me, a little while later that yummy feeling of science started to spread throughout my body, and I got around to planning an entirely new article in the series about Icelandic volcanism. Now it is time to get to the candy itself.
The Geirfuglasker Central Volcano
For the regular readers of Volcanocafé it is well known that we use earthquake data positioning to track everything from magma reservoirs (shown as voids in the plots), to intrusive dykes, sills, and central volcano fissure swarms.
According to the classical model the Reykjanes Ridge is not containing any central volcanoes. Instead volcanism is caused by episodic intrusions as the Mid Atlantic Ridge is pulled apart. Whereas a central volcano has a complete magmatic system with a feed from the mantle leading up to either a deep reservoir and then onwards up to a shallow magma chamber, or going directly up to the magma chamber.
This magma chamber should be a permanent fixture, and not an ephemeral construct consisting of a onetime dyke or sill. In Iceland it is also more or less a given that a central volcano should have a fissure swarm of its own where known fissure eruptions has occurred.
None of these things was known to exist for the area as I wrote my previous article a little bit more than a week ago. Still, they did exist in glorious detail in Andrej’s earthquake positioning plots. In fact, it is one of the clearest instances we have seen, so it is a bit of mystery that it has been missed.
If we start at the bottom, we find an elongated SW/NE-trending deep sheet of earthquakes that continuous upwards. The bottom of this deep feed system is at 27km depth and it continues up to 15km where we find a kilometre-wide sheet-dyke that goes all the way up to 7km depth.
The sheet dyke shape of the magma reservoir comes naturally from the spreading of the MAR, what differentiates this dyke from others is the width, at a kilometre across it is sufficiently large to be a permanent fixture containing quite a bit of magma. I roughly estimate the magma content to be 35 cubic kilometres. Not that large for being Iceland, but still enough to be considered as a magma reservoir.
From that reservoir there are cracks leading upwards to Geirfuglasker and Skerja (see map, the volcanic feature next to Skerjadjúp). But, there is surprisingly little activity going on at Eldey. This means that an eruption at either Geirfuglasker, or Skerja could occur quite rapidly. And that an eruption at Eldey would only come after quite a prolonged seismic unrest.
It is though good to remember that this area of Iceland is intensely seismic, and that any eruption would most likely start after a rather hefty earthquake swarm. In other words, if it is not memorable it is not happening.
Now that we have found that the fissure swarm is extending to the south-west a few kilometres to Skerja, where it abruptly ends, it is time to see if we can find a north-east continuation of it. And it is here that we find something so dramatic that it is accompanied with the angelic choirs normally only associated with baby Jesus on Christmas.
What we find is a continuous feeder tube extending from the bottom of the magma reservoir under Geirfuglasker running at an upwards angle all the way up to the tip of mainland Iceland. This connect well with petrochemical data that tells us that the magma found at Eldvorp is of the same type found at Eldey (and hence Geirfuglasker).
Or, in other words, the tip of Iceland is a flank vent of the oceanic central volcano of Geirfuglasker. Surprise is a very mild word in this context.
Possible sources of error
In science one should always point out any flaws in a theory. And here there is a glaring one. By nature, the data we are using are earthquakes, and since the entire volcanic feature is situated on a major tectonic fault, the bulk of the earthquakes we are mapping are tectonic in nature, and not volcanic.
In an ideal world we would filter out all earthquakes not containing a magmatic or volcanic signature from our maps. This is though not possible from the data that we are using. Instead we are using shapes, figures, voids and discontinuations in the data to glean what is hiding inside the crust.
Here we interpret the data as being connected to overlying volcanic features. The data is though so impressive that we strongly believe that we are correct. But, we would be happy if any young strapping Ph.D. student would do the legwork and filter out the volcanic type earthquakes from the tectonic to produce a better validated model than we could.
I am certain that the same basic model would come out of such a dissertation. But if not, we would in the end have an even better scientific understanding of a truly intriguing part of Iceland. Being wrong is also a yummy part of the scientific process.
CARL REHNBERG, text & ANDREJ FLIS, Geoplotting