Laki deconstructed: I. prequel

Central volcanoes of Vatnajökull

With all volcanic activity seemingly on hold during the shut-down, here is a re-post from the dungeons, originally published May 29, 2013, and written by Carl. It leads into a post on Laki which we hope to put on in the next few days. After that, who knows. Perhaps something on the developments at Anak Krakatau, some shatteringly beautiful volcano you’ve never heard of, the best piece of music to accompany an eruption. But first, back to Iceland.

You may notice that this post predates the Bardarbunga eruption. It was ahead of his time.

Grimsvötn 2011 eruption arial photo

Icelandic volcanism in brief

To understand our Icelandic volcanoes we have to go back to how they formed and what drives them. As most of you know there are two functions driving the volcanism in Iceland. The first one is the hotspot. It supplies Iceland with an almost never ending supply of nice juicy magma. The other force is plate tectonics.

Most of what we perceive as volcanoes on Iceland is situated on the zones where the Icelandic part of the Northern Mid Atlantic Ridge separates with half of Iceland being American, and the other half being Eurasian. As it spreads a fissure opens up, and magma pours up to fill the void. The last sentence is more important than you might think.

Wikimedia Commons. General tectonics, fissure swarms and some central volcanoes. Please observe, this is not a complete map.

The Icelandic volcanoes have all started as a fissure rupturing: think Krafla here. If the fissure eruption continues parts of the fissure will close up and cones will start to form. In the end just one cone will be erupting and a shield volcano is born. The 30 cubic kilometer Theistareykjarbunga volcano is a great example of this early stage for a volcano.

Due to the duration of the eruption there will be a lot of residual heat and left over magma under the newly formed shield volcano. Some of them never erupt again. But some of them get new magma infused into them and as time goes by a magmatic system forms under the mountain with magma chambers, feeder tubes, root feeding tubes and so on. Now and then these volcanoes will erupt, and as time goes by they will change into more evolved types. Also the magma will change into more evolved types as the magma will sit in the system for quite some time before being reheated by new magma. In this case a stratovolcano will be forming and we have a new central volcano.

During normal times these volcanoes will erupt magma that arrives in pulses into the system. This is how Grimsvötn functions when it has its regular eruptions, like the one we saw in 2011. The same fissure can through different fissure eruptions form several central volcanoes. The same fissure that formed Grimsvötn has also formed several other central volcanoes. The most well known of these are from north to south Háabunga, Thordharhyrna, Geirvörtur, Hágöngur and Eldgigur. At least four of those are proven to have erupted without any fissure eruption being present: these are Grimsvötn (obviously), Thordharhyrna, Hágöngur and Eldgigur. Háabunga would probably be a safe bet, but there is no geological or historical evidence for it. Eldgigur has erupted three different forms of evolved tephras, and a small bit of Lakí type magma on top. This volcano actually did the reverse order of how Icelandic volcanoes normally form.

But, during the Laki eruption all of them erupted at the same time. Here is a very important thing to remember: the Icelandic volcanoes are bimodal. Most of the time they function as singular entities, but when a rifting fissure episode occurs they function as a big fissure with many, or all, of the volcanoes on the fissure swarm erupting at the same time.

So when reading what I write below, remember that we are discussing the first “business as usual” style of eruptions for these volcanoes. We will therefore not see any evidence of them being fully interconnected. But, as we will see they are all separate central volcanoes during normal times, not sub-units of the two big central volcanoes (Bárdarbunga and Grimsvötn) even though they formed from the same basic two fissure swarms.

Not all earthquakes are equal

A normal misconception is that all earthquakes happening at a volcano is volcanic. Problem is that many of them are just in a secondary way related to volcanic activity. As we go about business we have to remove the purely tectonic earthquakes: remember that these volcanoes are on top of the spreading MAR. The next two sets of earthquakes to remove are the ones caused by cooling magma causing the mountain to shrink, and earthquakes caused by the mountain settling and pressuring downwards. These two types might be secondarily caused by the volcano, but they are not signs of how the volcano works before an eruption. The cooling earthquakes are though a good sign that the mountain is not going to erupt for the time being.

As new magma is pushed in it fractures the surrounding rock. These earthquakes have a slightly different signature to them. And they are normally easy to correlate since there will be signs of the uplift and inflation. The second group will be magmatic earthquakes. These have a very different signature, like no other earthquakes on the planet. To simplify it beyond belief, they look “gurgly and wet”: these are your basic long period events caused by increasing pressure in the volcanic system. There are of course more types of earthquakes than this.

Normally we go about our own little business plotting all of the earthquakes that occur in these volcanoes, and then we try to draw conclusions. Problem is that now and then the non-volcanic activity earthquakes lead us astray. Just look at the mess below. It is hard to see anything useful from when you look at all the earthquakes from 1992 to the 2011 eruption of Grimsvötn.

Image from Icelandic Met Office created by Krisín S. Vogfjörd. All earthquakes 1992 to the 2011 Grimsvötn eruption.

If we now compare with an image with only the earthquakes caused by volcanism showing we will have a much clearer image of what is going on.

Image from the Icelandic Met Office, created by Kristín S. Vogfjörd. Showing selected earthquakes.

I am not going to discuss Kverkfjöll, Esjufjöll and Öraefajökull here. I am going to concentrate on the volcanoes on the Bárdarbunga and Grimsvötn fissure swarms. But, remember that the only reason we call these the Grimsvötn and Bárdarbunga fissure swarms are due to those two central volcanoes having the largest number of eruptions along the line. One could as well have called the Grimsvötn fissure swarm the Thordharhyrna fissure swarm since it is the volcano in the middle. It is all a point of perspective really.

First of all, let me point out that there seems to be very little to no evidence for the volcanoes being interconnected during the normal eruption times. They really act as separate central volcanoes. I am though not going to go into depth of the Gjálp conundrum in this post. That will have to be saved for another time, here I will just point out that there is no evidence for any magmatic movement between Bárdarbunga that had a large earthquake that is believed to have caused the eruption of the Gjálp fissure from Grimsvötn central volcano.


We see that the earthquakes of Bárdarbunga are neatly centered around what most likely is the active magma chamber. A few earthquakes run down towards the depth, a sign of magma movement, but in a mature feeder that most likely is permanently open downwards, and this probably means that most of the earthquakes before an eruption would be around the chamber. Please note that there are two small tendrils of earthquakes moving towards Grimsvötn fissure swarm. The smallest of these is rather insignificant, but the other leading towards the fissure swarm north of Grimsvötn is a bit of a conundrum. I do not have a good explanation for that one. It might be related to the rifting fissures of Iceland, but that is just a guess. I will return to this in the conclusion.


Most of you will notice that there is a very strenuous separation between the Bárdarbunga earthquakes and the earthquakes of Kistufell. At the same time there is no clear evidence of a connection between them either. Probably Kistufell is a central volcano in its own right and erupts separately; I guess time will tell on that one.


Most people see Hámarinn as a subset volcano of Bárdarbunga, but as we see there is no evidence of that in the dataset. To the contrary, it steps out like a rather significant central volcano in its own right. At the same time we have the Loki-Fögrufjöll volcano that is seen as a subset of Grimsvötn.

In the data we though see a completely different picture grow forth. Hámarinn is very clearly separated from Bárdarbunga, and the same goes for Loki-Fögrufjöll lateral fissure system being separated from Grimsvötn. Instead we see that Loki-Fögrufjöll apparently is a fissure volcano emanating from Hámarinn volcano (or the other way around). On the other hand there is a clear trend pointing towards Grimsvötn. And, at the same time there is a shorter trend running down towards Thordharhyrna volcano. I will come back to this in the end.

It is also well to note how active the feeder tubes have been at depth, a sign of heightened activity in the system during the timeframe of the dataset. Interestingly enough there are 3 clear deep feeders being active leading up to all 3 eruptive vents. To me this points toward a less evolved volcanic system than either Bárdarbunga or Grimsvötn, and at the same time with the potential of developing into a very advanced interconnected volcanic system over time.


Grimsvötn has been very active during this timeframe. It is the only volcano that has erupted among those we are discussing here, during the timeframe of the dataset it erupted no less than 3 times. It is therefore surprising how little volcanic earthquake activity the volcano has shown. Almost all of the earthquakes have been around the chamber or chambers. Very little activity shows towards depth, a probable sign of a permanently open feeder system.


Laki fissure line

As I wrote above, this is how it looks during normal eruptive times. I will though have to put in a caveat here: it is an assumption, and we do not have data for how it looks when other volcanoes than Grimsvötn erupts. The most likely thing is that the picture will not change that much, we will only get more at the same locations.

Let us now look at the small enticing enigmas we found and how that might affect things during a rifting episode opening up a fissure swarm. First thing we have to acknowledge is that we know very little about this style of eruption. We have no data set for an eruption like that. In reality we have at best secondary data or written records for the Lakí eruption. And from the Lakí eruption we only have the larger strokes, not the more minute details that happened.

What I find intriguing are the trail of earthquakes running from Bárdarbunga towards the northern part of the Grimsvötn fissure swarm, the Hámarinn/Loki-Fögrufjöll pointing towards Grimsvötn and the line running from Hámarinn towards Thordharhyrna.

We know that Thordharhyrna is a part of the Grimsvötn fissure line due to them having co-erupted during rifting fissure eruptions. We also know that there is a known mechanical link between Thordharhyrna and Hámarinn. And Hámarinn we know is on the same fissure swarm as Bárdarbunga.

I think that the central volcanoes work as weaknesses between the fissure systems, and that during the larger rifting fissure eruptions they tend to open transverse fissures between the fissure systems and each other. I know that this is a bit of a shot in the dark, but there seems to be a bit of evidence for the possibility in the data set. If so it would explain a lot about where the huge amounts of magma comes from during a Lakí or Veidivötn eruption. Of course this in the end would have to be corroborated by hard data. But the idea is intriguing enough to mention as it is.


24 thoughts on “Laki deconstructed: I. prequel

    • Anyone have something even better? Deep Purple doesn’t quite fit the volcano colour scheme, perhaps.

      • How ’bout
        “There’s a Fire Down Below” (Bob Seger)
        Not particularly colorful, but you get the idea?

          Our moon still tectonically active??
          Still adjusting after 4,3 billion years.
          We use to discus and discard the moons tidal role for seismic and volcanic activity on earrth. What about the other way with earth squeezing the moon? Of cause, the moon is side locked to earth so not so much energy there, or?

  1. Come one guys, it’s obviously “Ring of Fire” by Johnny Cash.

    • Every time there’s an eruption in NZ/|Japan/Indonesia etc the press report states (it must be compulsory) that the eruption was on the Pacific ‘Ring Of Fire’ Can’t help thinking that the Estate of Johnny Cash ought to be getting royalties 🙂

      • That’s why I posted it. His song might predate plate tectonics theory.

  2. Nice article. Is it hard to distinguish between the different types of earthquakes? And why is this kind of information not shared with the public? A simple letter behind de magnitude would give a clue (M, T, C, O, U).

  3. Just checked the Thorbjörn GPS readings and it seems, since around April 20, the inflation has stopped and the area is instead subsiding. GPS trajectories have basically reversed.

    • That sucks… nothing happens

      Anyway: Have you checked the Grimsvötn Cumulative Seismicity Graph recently?
      Is the magma chamber growing towards the surface under the South caldera wall?

      The magma chamber is saied to be very shallow just like Kilaueas summit chambers.

      Grimsvötns upper chamber is shallow indeed. Will it break a path to surface soon?
      What will happen to Grimsvötn now?

        • I hopes the magma column or chamber blast a hole in the icefree south caldera wall.
          And we gets tall lava fountains blasting out.

          Or I wants an open conduit down to the lava.
          But Thats Impossible with this kind of basaltic volcano.

          It will be same kind of glacial surtseyan pheratomagmatic show as 1998 and 2004

        • Albert
          While this have been discussed extensively

          But since Grimsvötn is a large and very vigorous and open magma source.

          Is there any chance for Shield like Trölladyngja to grow example near Thordarhyrna?

          But I knows that law of physics probaly prevents constant open eruptions in the Grimsvøtn system.

          Iceland does mini flood basalts and other fast fissure events

          Rather than slow shield building

          Iceland is acually a small LIP

          • Shields tend to grow in new locations, on top of the rift. Put a big cone on top and the pressure probably prevents an effusive eruption there – it will come out in a less obstructed place. If you want to know what Grimsvotn can do, look at Laki. But building shields requires a relatively stiff rift. I think the dead zone is too flexible and opens up in too many places. You get a series of small cones rather than one big one. Shield volcanoes may form more easily north of Grimsvotn.

        • Maybe that will happen at Greip?
          A Surtseyan phase starts up at Greip?
          And builds a long lived Puu Oo like cone that grows out the icesheet. Maybe even forming a shield like Theistareykjarbunga there?

          In Grimsvötn its more faster short eruptions
          Thats the norm as Carl and Irspsit says.
          As carl also made an article about.
          Fast surtseyan – pheratoplinian eruptions are norm for Grimsvötn
          This is in this Article lack of Icelandic lava lakes

          But It woud not supprise me at all,
          If Grimsvötn blasts a Overlook like hole in its South Caldera wall.. Imagine driving up to Svianukur and looking down into the angry raging lava pit fountains

          Grimsvötn is open conduited

        • Most of the upper earthquakes are associated with expansion of the magma reservoir in under the southern caldera wall roughly where the increase in geothermal heat is.
          But I looked at Satelite of Grimsvøtn a few weeks ago and 2019 years open caldera wall lake where frozen again.

  4. Albert, at the beginning, you wrote: “After that, who knows.” May I most respectfully recommend Billy Mitchell Volcano on Bougainville Island, VEI 5 and 6 previous eruptions, 1030AD and 1580AD, respectively. I found no previous reference in the Volcano Cafe archive.

    I’ve learned very much from your essays and comments by others. Thanks and may your heath remain good. // jms

Comments are closed.