Seismic unrest and the future of volcanism at Herðubreið

The Epic edifice of Herdubreid with the small spatter cone on top visible.

I had intended to write either an angry op-ed, or a piece about Guatemalan volcanism. But, as usual Iceland had to butt in and grab the attention, so here we go (again).

Seismic unrest at Herðubreið

Before I write about what has happened in the last few days and about what is going to happen in the near geological future we need a geological background.

The Herðubreið volcano formed during the last glacial-period when Iceland was covered by a 2km thick layer of ice. When the eruptions took place that formed Herðubreið they happened under the ice, melting a circular cavity in the ice. During the eruptions, a lake of water formed above the eruption and a circular cauldron formed with walls of ice.

This lead to lead to layers of pillow lava forming, pretty much as they would during a sub-aquatic eruption, but with a small twist. Due to the ice-walls of the cauldron the sides of the volcano became very steep, almost at a 90 degree angle and the top became flat.

This type of volcano is fairly rare and is called a Tuya, or a table-top mountain. And no volcano on earth is a better example of this process than Herðubreið. At the end the mountain became too tall for the diminishing ice age glacier and the final eruption occurred above the ice and a small spatter cone formed at the top.

This makes us fairly sure that the volcano has not erupted during the last 12 000 years since we have a glacial timeline for it.

Herðubreið is a bit of an oddity for Iceland since it is not situated on a known fissure swarm and seems to not have a fissure swarm of its own. It is therefore not a part of the regional east-west extension of Iceland and should therefore be relatively seismically inert.

We do though know that it is a central volcano of its own and that it used to have a developed magmatic system that was among the more active during the glaciated period prior to Holocene. If left to its own devices that magma system would after 12 000 years have become solidified and Herðubreið should have become a former volcano by now.

Nice and distinct rock breaking earthquakes detected at the Mokollar station between Saturday evening and Sunday evening. Image from Icelandic Met Office.

In 2008 a large seismic unrest started at the volcano of Upptyppingar to the south of Herðubreið as about 1km3 of magma intruded into the crust at 14km depth and there was a scare that the hilariously named Upptyppingar would erupt. If that had happened we can only hope that CNN would not have translated the name themselves, in the end the name Viagra Hill was suggested to the international press. And yes, the actual name is far worse…

That being that, in the end the bedrock above the intrusion was to solid for the magma and a northward trending dyke formed. After that we have seen between 1 and 4 vigorous earthquake swarms per year as the dyke propagated slowly northwards at an upwards angle. As the dyke arrived at Herðubreið it was at a depth of 7km.

What happened as the magma entered the old system of Herðubreið was so fantastic that it made me do splits while playing lip-banjo out of joy. I have several reasons for my joy. The first one is that something almost magical happened, as the magma entered the old system it started to form earthquake stacks towards the surface indicating reformation of the old volcanic system. This in turn caused the second reason for my joy, science have never before been able to follow in detail how it looks as a long dormant volcanic system comes out of dormancy. And the third reason is that the system must have stayed reasonably active, otherwise the volcanic system would have solidified beyond the ability to form new magma tubes towards the volcanic edifice.

During the last 3 years we have seen earthquake stack after earthquake stack form as the magma has tried to find pathways towards the surface, and every time the magma has come a little bit further up. The image that is starting to build of the volcanic system from these earthquake stacks are intriguing and gives us a very good image of how a reforming volcanic system functions.

The current seismic swarm

Map of quakes over the past 96 hours

Earthquakes for the past four days (96 hours), 18-21/03/17. Earthquake data IMO, map data Beardy Gaz, Google, FutureVolc and VC.

The current seismic unrest at Herdubreid captured on sunday evening. Image from Icelandic Met Office.

It is easy to get waylaid by small and badly located earthquakes. There has been quite a bit of talk about earthquakes at 2km. The reason that people bring up earthquakes at that depth is that it is thought of as the point of no return for buoyancy driven eruptions.

Above that depth the magma will start to rise upwards due to the magma being lighter than the surrounding bedrock due to the heat energy having expanded the magma. This is in most instances quite true, and at depths above that volatiles will start to nucleate out of the magma causing rapidly increasing pressure levels.

Only problem is that there has been few earthquakes at, or above, 2 kilometers.. If you are looking at earthquakes it is good to remember two basic rules. In all Icelandic volcanoes except Katla and Hekla any earthquake of less magnitude than M1.5 will be too small to have been located accurately. So, look at earthquakes above that size and you will get decent depth estimates. Second rule is to wait for the manual check of the earthquakes. The automatic system is giving unreliable depth locations.

If I look at manually corrected earthquakes above M1 there has been a few at 3.1km depth and one M2.1 at 1.2 kilometers depth. This is pretty much to be expected since each earthquake swarm has gotten about 0.5km closer to the surface. At the moment the swarm seems to be dropping in intensity, so it is unlikely that something will happen right now.

I have seen a couple of comments around the internet that this earthquake swarm is not caused by magma, all of them based on the IMO specialist remark about the swarm stating that there is no fluid signature detected. The IMO comment is of course correct, but slightly taken out of context.

The earthquakes are caused by the general pressure that the magma intrusion has built up under the volcano. This causes strain that fractures small rock faults and in time magma will move into them causing renewed stress. But this process does not give the telltale seismic signal of freely moving fluids.

To be able to see small quantities of magma moving distances of a few meters at a time you would need equipment that is both close by and that are applied in large quantities. And that is not the case right now.

Instead you look if the earthquakes form a distinctive earthquake stack that might be a magma conduit, and in this case, we do see one of those forming. In fact, it is a rather handsome little stack.

A peek at future possibilities

Earthquake stack made with 3D-Bulge by Jonet Greene.

The Upptyppingar to Herðubreið intrusion is one of the most well document in the history of mankind. It has given loads of scientific data about reformation of volcanic systems that are long dormant. This wealth of data also gives us the possibility to coax out metadata such as rate of upwards progression per earthquake swarm.

We know for instance that the rate of earthquake swarms is between 1 and 4 swarms per year, and that each swarm averages at 0.5km in upwards motion and that the initial northward motion has stopped in the Herðubreið area.

We also have ample evidence that the volcanic system of Herðubreið is rapidly reforming in the form of sills and dykes with a possible magma reservoir at 7km depth.

We also know that the bulk of intrusions in Iceland putters out before reaching the surface in Iceland. But, in this case the rate and intensity of the earthquake swarms does not indicate that it is about to putter out at Herðubreið.

All of this gives us a logical conclusion based on the evidence at hand. It is currently unlikely that Herðubreið will erupt from this earthquake swarm unless it continues for several more days, or even weeks, with the same amount of vigour. Instead it is likely that the earthquake swarm will slowly loose vigour and putter out. For now.

If we instead project what we know into the near geological future, we see that it would take anything from 1 to 5 earthquake swarms before we reach a point of no return depending on duration and size of the earthquakes in the swarms. And as that point of no return is reached an eruption will occur. So, let us say that an eruption at Herðubreið is likely to occur within 6 months to 4 years at a significant percentage rate of probability.

An upcoming eruption would occur within a few kilometres of the original volcano, or perhaps even up through the original volcano. Unless there is a pocket of evolved magma somewhere down there the eruption would be from a distinct vent and create a spatter cone and freeflowing lava. The eruption could be cyclical like at Krafla or be in one long go like at Holuhráun and the size would probably be somewhere in between those two eruptions.

An upcoming eruption would probably be more scientifically interesting than it would be interesting for the general population, but then on the other hand. People love Icelandic eruptions a lot so I may be wrong about that.

I must though admit that it will be a bit sad watching one of the most beautiful volcanic edifices be destroyed, but mother nature is after all not about aesthetics.


98 thoughts on “Seismic unrest and the future of volcanism at Herðubreið

  1. Sweet! and i’m still watching April the Giraffe because i still have lots of snow….. Best!motsfo

  2. Thanks for an interesting read! Also, thanks for the help with my user account. This is the first test, so let’s find out if this also ends up in the dungeon.

    I have to agree that it would be a shame if Herðubreið lost its iconic shape. One of the strongest memories I have from Iceland is driving along the part of highway 1 between Egilsstaðir and Mývatn and seeing Herðubreið in the distance.

    Back in the dungeon again, its cold indifference is made warmer with your constant visits! – Admin

  3. Iceland simply amazes me. For a country roughly the size of Ohio or South Carolina, there seems to be an insane amount of geologic beauty (both aesthetic and academic) there. Definitely on my bucket list. When the time gets closer, I will have to inquire about any “must do/must see” – just as long as one of those places doesn’t include doing a “*quick jaunt* to the top of Hekla”…. 🙂

  4. Great article.

    For a decent parallel to this situation, I strongly recommend everyone here read Erik Klemetti’s article on Lassen Peak rejuvenation. In this, he discusses how long dormant volcanoes can be brought back to life with new magma injections – basically what we are potentially seeing at Herdubreid.

    Personally, I wouldn’t think a new eruption here would be so effusive, at least at the start. There is presumably a decent amount of evolved magma in Herdubreid. 10,000 years is enough time for old magma to fractionate into more evoled magmas. The question would be how much of that magma is activated by the new injection of magma coming into the system.

    I don’t think an eruption here would be enormous in size, but I would personally guess that a new eruption would start out explosive before turning into an effusive eruption (which is common in Iceland). In a lot of ways, I would expect this to be somewhat similar to a Hekla style eruption if it were to go off, although I have no clue how large or small it would be.

    • I guess I have been influenced by Jesper Sandberg, but I can’t get the image of lava-flows plumeting 800 meters down the sides of Herdubreid. It would be gorgeous 🙂

      • Agreed 🙂

        I’m curious if magma does reach the surface, if it will go through the top, or if it will create a new vent to the side of the main edifice. I would imagine that it would create a new vent to the side since there would be less resistance that way, but who knows. If this were the case, it likely wouldn’t destroy the pristine Tuya.

        • Depends on if it occured Close to the Tuya or not really. Kistufell is a Tuya merged with both a shield and a strato-cone. The result is fairly ridiculous.

  5. Regarding the depth of the earthquakes.. is 0km sea level, the surrounding plain or the top of the volcano? Given the elevation of Herdubreid is about 1.7km, that makes a big difference when judging how close the swarm is to the top. Also, from the map it looks like the swarm is in the plain next to the volcano, so could this be a fissure type of event?

    • I’m not sure about IMO specifically, but USGS has it as follows:

      “This depth may be relative to mean sea-level or the average elevation of the seismic stations which provided arrival-time data for the earthquake location.”

    • It is the average of the seismometers height above the ocean. So, around 400ish meters up.

      There is no known fissure at Herdubreid, so it is more likely that this is a local crustal weakness.

      • I do not think it would be surprising that the August 2014 rifting created a stressed area in the area of the current swarm.

        The underlying pressure may be taking advantage of this stress and creating a new “crack”?

        • Doubtful since this series of earthquake swarms predate 2014 with 6 years.
          No, this is most likely an old process of volcanism. It is just that we have never seen it Before since both Upptyppingar and Herdubreid are not the most common of eruptors.

          • What caused me to think this is a mechanism was looking at this map


            If you grab the blue slider on the chart at the bottom of the map you can control the animation

            I zoom in so you can see BB and Askja and move the slider to just before the 2014 action and as the dike runs away from BB you can see the line of quakes that suggests the dike will run past Askja right onto the current area of the swarm

            Link to full size


          • Ian, can you do the same map plot and time scale, but only include quakes greater than or equal to 1.5 please? I’m curious as to whether it gives a clearer picture of the activity. Can the quakes be colour coded in relation to depth like Andrew’s plots?

          • The problem is the temporality of the swarm. FIrst you had Upptyppingar in 2008 that slowly and in spurts meandered over to Herdubreid where it arrived in late 2012.
            In 2014 you had the large rifting eruption, and at the same time the Askja leg happened, but that was also not related to Upptyppingar.

            But I give you that it has been a busy corner in the last ten years.

  6. For some reason I found this screamingly funny, especially the suggestion that CNN would translate the name of Volcano #2 as Mt. Viagra. I shared it with a literary discussion group who appreciate good satire though I think I am on shaky ground (pun intended) reading it as such.

  7. Here is my new video, in light of the Herdubreid activity, a closer look at overall seismicity and how the latest swarm fits in. Data used is from the IMO earthquake catalogue, and only manually checked earthquakes were used.
    I recommend watching in full screen and HD.

    I am working on a portable version of the software with the packed data, to share it individually, and so it works as a install package together with the data and everything, so the plan is to kinda make it like “plug and play” or “install and spin”. But its not ready yet.

  8. fascinating. Mt Viagra put Bard under stress
    21.03.2017 17:48:55 64.666 -17.399 3.1 km 3.1 99.0 6.7 km ENE of Bárðarbunga

  9. Ármann said eruptions could be in store: 500 earthquakes in the swarm
    Editorial DV
    13:52> March 21, 2017

    About 500 earthquakes have been Herðabreiðartöglum near caldera. Ruptcies could be available for the eruption which could even take place in the coming weeks. Another possibility is that earthquakes are to release the pressure after the eruption of the volcano. This was stated in a conversation with Gissur Ármann Höskuldsson eldfjallafræðing in hádegisfréttum Channel 2. …

    Released from the pit of despair – Admin

    • Giggle translate of the rest:

      “The largest earthquake in the swarm was almost three points but significantly reduced the cycle yesterday and last night. Asked what could be read from these earthquakes Ármann says that two possibilities are in the situation. One possibility is that after the big eruption in the volcano area that is adapting and slowly releases a voltage. “Then it ends with earthquakes and maybe open some cracks,” says Ármann and there is nothing to worry about. But there is another possibility and it could mean that an eruption is coming.

      “On the one hand there is this dynamic that penetrated Álftadalsdyngju the river near many years ago and has been expanding west along,” says Ármann and adds that this behavior is somewhat strange.

      “When last heard her, it was in this area where these earthquakes is now. It could then mean that someone present in the ash system we had a volcanic eruption in the next few weeks or months or years. ”

      Ármann says that magma is moving closer to the surface and it could be shorter in the eruption area. Ármann concludes:

      “Not necessarily close to Herðubreið but in a box system itself because it is moving into the caldera system when it goes like this the west.”

      So it seems he’s saying (trying to understand the translation!) that this could be tectonic adjustment to the Holuhraun eruption dyke. OR it could be reactivation of an earlier intrusion which may be moving magma into the Askja system (that’s the only “caldera system” in the vicinity!) leading to an eruption in the relatively near future.

      When we were filming the Holuhraun eruption we stayed with Ármann and his team at the Dreki hut for a few days. He was surprised to find a cameraman who could talk about non double couple quakes – and I found him impressive; he should be listened to.

      • My Icelandic is extremely rudimentary but the words translated “box system” and “caldera system” in the last paragraph of the article are, are, respectively (and exactly as they appear in the article), “öskjukerfinu” and “Öskjukerfið”, the first part of which are inflections of Askja, which, to complicate matters, itself simply means caldera or box. So I agree, he does seem to be naming Askja as a possible eruption site.

  10. Peanut Gallery comment. At about 7 km, I calculate the lithostatic stress as about 185.35 MPa. Generally you assume 2700 kg/m³ (granite), but doing the same calc for 3100 kg/m³ (basalt) you get about 212.80 MPa. Since the whole complex is under tension, the ambient stresses can be lower as the region is stretched/thinned.

  11. Once I asked a Holuhraun scientist in the area, during the 2014 eruption (about the quakes often happening at Herdubreid) and he told me that area (at Herdubreid) behaves like a fracture region, and actually one can see stacking of N-S faults progressing towards the northwest. Problem is, fracture regions like SISZ also have eruptions, because magma can escape upwards in the narrow faults. This activity at Herdubreid is so recurrent and trending slowly upwards that I predict it will finally end up as an eruption.

    The curious is how an eruption would take shape.

    There are many shields in the nearby area westwards. So I think things can get interesting there.
    Herdubreid tuya might have been formed by a similar large-sized efusive eruption under the ice cap (the first eruption), which would have been basically a shield eruption under the ice, resulting in the table mountain shape. So, eruptions in the region could be large-volume.

        • I should though Point out that it is a very rough estimate since we could not find exact measurments of the width of the volcano at both the top and the bottom of the edifice.

          • If I take some rough averages to make it a basic cylinder and discount the splatter cone on top for height, then I get r=3.2km, h=0.75km which equals V≈24.13km3

          • Bear in mind that the edifice is only what remains now after the retreat of the ice cap after the last ice-age. Some will have been eroded during the retreat and afterwards.

            I do note that KarenZ has retrieved back the Mantle of Galumphing Dungeonism from Tomas Andersson. /Admin

          • That produces a good estimate, it is probably slightly more than that since it is not a perfect cylinder.

          • You can use google earth to do this. If you open it, zoom to Herdbubreid, there is a tool in there that can provide a measurement between two points (in a line). If you just click the edge of the top or the edge of the bottom, you can get a pretty good estimate of the diameter.

          • Yes, that’s what I used to estimate the dimensions, very roughly…

          • This puts Herdubreid in the range of 20km3, very similar to the other shield volcanoes in the area between Bardarbunga and Tjornes. And there has been several since the ice age, at least 10 (making an average of one such eruption every 1000 years).

            This was my rationale to why I thought that Holuhraun would have become a shield volcano.

            (In Iceland, there has been only one shield eruption in historical times, and that was around year 1000, in the west side of Langjokull. It started as 3 vent fissure and then progressed towards a single or double vent shield volcano, but records from that time are very very scarse. Volume was around 10km3).

  12. I remember back in 2014, a set of weird stories occurred in the Herdubreid and Askja region. It was just before the Holuhraun eruption started, some two weeks before. Tourists reported a large boom sound coming from within the Herdubreid mountain. And two weeks before that, a large landslide and another rumbling occurred in Askja.

    Another thing I remember too, was as the Holuhraun dike progressed northwards, it seemed to jump across Askja and continue towards the Herdubreid area. Either this or the weaknessness in the crust were propagated in that region.

    • The Holuhraun dike was pointed straight at the Askja caldera; there was much discussion about what would happen if it reached there!

      Then it stopped dead and erupted vertically.

      What exactly do you suggest “seemed to jump across Askja and continue towards the Herdubreid area”? It certainly wasn’t the dike or the associated seismicity.

  13. Swarms have now occurred around 3km deep under Herdubreid, and at times up to around 2km deep.

    If each swarm moves magma 0.5km upwards, then we should need about 4 swarms, assuming things do not develop faster as magma nears the surface. This is about 1 to 4 years until the eruption.

    Also I disagree that this area shouldn´t be expected to erupt. The previous eruption in Askja in 1961 occurred just outside the eastern edge of the caldera. This swarm is about 16km northeast (the area is only slightly east, about 10km, of the general Askja swarm region, where some shields are located, north of the caldera)

    • Perhaps this could influence/cause a potential eruption in Askja, with it being close, maybe in the same area as 1961? I guess it depends if Herdubreid is truly an independent volcano or if there’s some link with Askja.

      • I don’t think there’s any link between the two, it’s not part of Askja’s fissure swarm and the intrusion is unlikely to jump across into Askja now it’s so close to the surface. It’s possible stress changes could influence Askja down the line (decades), but this intrusion seems to be fizzling out, for now…

  14. 37 Years Ago: Mount St. Helens blinks awake.

    On another note, Cerro Azul (Galapagos Isl.) is getting restless, eruption anticipated.

  15. Some of you, are much more mathematically inclined than I am. So, a gift link for you.

    Magma chamber behavior beneath a volcanic edifice” Pinel, Jaupart (2003)

    In that paper, you will find a description of “Hoop Strength” and how it applies to a pressurizing magma chamber. In essence, the tensile strength of the rock in the chamber walls is augmented by the Lithostatic stress. As the chamber pressurizes, when it overcomes the hoop strength of the “chamber”, a dike forms and begins propagating. “Chamber” as I use it, refers to the the area of many thread like tendrils of magma that make up the partially molten rock. Many months ago on here (well, in VC) someone likened a magma chamber to like a ball of cotton. Not every part of a ball of cotton has plant material fiber, much of it is open space. In this mental model, each fiber equates to a tendril of magma, and the open spaces to the host rock. Near the middle, the tendrils are quite dense and approach being a contiguous molten area, on the periphery is the less molten “mushy” area. If you concider this “chamber” as one volumetric space, that is where the idea of breaching the chamber walls comes in. The idea is that once a dike begins, it will continue to propegate as long as it has enough pressure to overcome the “Hoop Strength” of the surrounding rock. But, once the pressure of the chamber drops below that point, the dike will slam shut.

    The Lithostatic stress field that declines with shallowness, also tells you why dykes almost always trend towards the surface.

    As Carl has noted, at some level you begin to see gasses exsolving from solution. This acts to ramp up the pressure of a dike as it approaches the surface.

    …in other VC mentioned ideas, one that I see merit in, the rapidly exsolving gases at the tip of the actively propagating dike can act as a sort of cutting torch to help fracture the rock ahead of the magma finger.

    • ” at some level you begin to see gasses exsolving from solution. This acts to ramp up the pressure of a dike as it approaches the surface.”
      This should not really be possible. The simple start of exsolving gasses can’t increase the pressure at the level where the magma protruded, simply because the exsolvation is pressure driven. As the first gas comes out and the pressure rises a hair the exsolvastion should stop.
      What i guess of what could be happening is that magma has a hard time filling small cracks because it will solidify very quickly and seal it of. Exsolving gasses can reach in to cracks easily and transfer the deap pressure into higher regions of the lithosphere. Higher up, the lithospheric pressure is lower and the gas can cause the cracks to widen and provide a path for the magma, wich then can exsolve more gasses…

  16. Question.
    It may be a dumb one… I don’t know… I’m just letting intuition be my guide here.
    After leaving school I had an abortive career as a welder. In my apprenticeship I was taught that a sound weld is usually stronger than the surrounding metal.
    Now when I look at that beautiful summit plateau structure of Herdubreid, and when I read Carl’s words about how the vertical “crown” was formed, intuitively I’m thinking “weld”….
    So my question is, due to the manner in which it was formed, is it likely that the summit is going to be tougher for a dike to penetrate than the surrounding rock, thus making a fissure or some new vent on the flanks more likely than a summit eruption… assuming of course that it erupts at all?
    I am also asking based on an assumption that any dike is headed for the summit…So I suppose my question is really, would the cap be strong enough to divert the course of the dike?

    • It is not a stupid question at all.
      In fact it is so good that I really should do an article about it.
      Let us just say for the moment that it has turned out that nature is a really crappy welder… 🙂

      • Actually, having the top partially pop off like an overpressured screw-top can and having a ring of lava vents pouring down the sides would be a novel eruption type. It’ll probably just burn through like a 40 megawatt laser through a balloon, but it would be interesting if it was pretty well welded.

        • ….Which was also part of my musing… Sort of , like a ring fault , but on the outside of the edifice?
          I’m not reading any reports of any seismic activity of note anywhere on Iceland since the start of the week, which seems out of character with the last few months.

          Nothing at Herdubreid seems to have been expected, but nothing around Vatnajokull either?

          • There’s been some stormy weather over Iceland and will be for the next day, that always dampens down the sensitivity of the seismometers, especially on the Vatnajökull icecap and exposed volcanic areas like Herðubreið.

    • Definitely not a dumb question – I actually mentioned a similar thing in a comment higher up that I would partially assume that an eruption would not erupt through the top of the summit for this reason.

      The assumption of course is that the old conduit would have to be hardened for this to be the case. After 10,000 years of no eruptions, this is likely the case, especially the closer you get towards the surface.

      • Apologies for not noticing that !
        Yes. My thinking was that under such circumstances, the rock in its molten form would have had an opportunity, given that it was constrained at the time of the last eruption, to let off gas, have some “impurities”… a term I use here with some qualification, and so on, to either burn off, or bubble out, or settle, creating a more condensed structure, and they were then presumably pressurised from above by reforming ice.

    • More important than the “weld” most probably is the weight of the plateau structure itself – gravity is an important factor in volcanic eruptions like we learnt at Bardarbunga. However, a well developed feeder pipe below the tuya cone for example may be the most easy path for the magma to reach the surface. I suppose it is different from case to case, without a generally applicable rule.

  17. Do we have GPS data for the area? Is there any indication of inflation? If not, an eruption should still be a long way off.

    All earthquakes seem to avoid the edifice itself: there are outside of the area of the volcano. That is also true for earlier swarms, isn’t is? The activity isn’ t strong enough to move underneath the weight of the mountain (dykes tend to move in the direction of least resistance, so not normally towards higher ground. That is also why the Bardarbunga dyke didn’t extend into Askja: it stopped at the lowest point)., and there is no indication that the old magma underneath is reactivated. I read this as that any eruption would likely occur on the open ground beside the volcano, and not on the summit.

    Are you sure Askja has nothing to do with it? The dyke points roughly away from Askja.

    • Any indication of inflation? Don’t know how much use this will be but the CGPS that I’m looking at were last updated on 20/12/16 and show data over only a short period, but GIGO, RIFC were showing inflation, and THOC was not. DYNG seems to be deflating. I haven’t been able to find any CGPS stations nearer to Herdubreid.

    • I will try to answer your question as well as is possible Albert.
      First of all, the original intrusion was in 2008 and was visible on GPSes and gave the same location centroid as the earthquake swarm centerpoint at 14km below Upptyppingar.
      The later movements of magma from Upptyppingar towards Herdubreidartögl and Herdubreid has not given any useful measurments since the magma was moving away from relevant GPS stations. Askja GPS station was at the time to intermittent and to far away to yield useful data.

      Yes I am quite sure that Askja has nothing to do with it. There is a clear temporal progression of the earthquake swarms from Upptyppingar via Herdubreidartögl to Herdubreid prior to the dyke that led away from Askja.
      The Askja dyke is well known radial fissure that ends at the rim of Askjas fissure swarm and the dyke ends 12km from Herdubreid and Herdubreid in turn is well outside of that fissure swarm.

      • One should not underestimate the link between different volcanic systems in Iceland.

        To put things quite to the extreme: Reykjanes and Tjornes are in fact connected to one same source, the Icelandic plume (that actually bifurcates into two parts before reaching the crust, one spreading southwest, another northwards). Or even more extreme: other research has shown different hotspots can be even connected some 1000km deep in the mantle. So there is always a “link”.

        But away from this “big picture”, when Krafla erupted in the 1970s, Askja immediately responded and deflated. That also seemed to show a “link”. Magma may be playing around in much more exciting ways, more deeper beneath the crust, in the upper mantle, in ways that connect Icelandic volcanic systems located quite apart!

        Obviously we don´t see that as earthquakes, because that´s just in the mantle. But the connection and the movements are still occuring.

        Research has shown that as the hotspot nears the crust, under Grimsvotn/Bardarbunga area, it divides into a branch travelling northwards. If such a pulse occurred in previous years (2010, 2011, 2014…), such large pool of magma has also travelled, deep beneath the crust, from Bardarbunga into Askja and further northwards, so it could have affected also the present intrusion (as well as further southwestwards towards the dead zone and Hekla/Katla region). We cannot confirm nor deny this, so at the moment its only speculation.

        • Any movement of magma from the plume into the crust, is often noisy at least to some degree (read as -earthquakes-).
          As I showed in my 3D visualisation yesterday, Herdubreid area does seem to have its own deep feeder to the SW, and the Uppty 2008 magma emplacement also had its own.

          No obvious cinnection to the Askja systems is to be seen from the seismicity.
          Its even hard to see one between the 2008 Uppty emplacement and Herdubreid.

      • Carl,

        Have a look at the map I posted up near the top of the comments – seismicity 1997-2014.

        And here’s another one for the 2008 intrusion specifically:

        I note a couple of things…

        First the much-mentioned ‘Upptyppingar’ intrusion was actually in a slightly different location to the northeast – in Álftadalsdyngja. Significance? Are they considered separate systems or all part of one system?

        Second in the total seismicity 1997-2014 map I posted – heck I’ll post it again here so you don’t have to scroll up 🙂 – we see the seismic cloud of 2008 to the lower right. And we see seismicity in roughly linear features in Herdubreidartögl and Herdubreid:

        But I see no *seismic connection* between them. Nothing resembling the Holuhraun dike seismicity for instance. They’re separated by seismic ‘dead zones’. So what exactly is the evidence and basis for the widely-mentioned belief that the 2008 intrusion magma is ‘migrating’ to the west and into Herdubreid? Looking at the map there’s no obvious connection.

        • The general trend aligns with the rifts to the east. The activity studiously avoids the volcanoes themselves and this suggests it is not (just) faulting: whatever is happening is inhibited by the cold and dense volcanic plugs – it can’t break them.

          We may be looking at stretching due to the plate movements. The plates move away and cause a general weakening of the area but the tear hasn’t yet happened. So you get a series of almost-tearing events. I don’t see evidence for direct connection between the various swarms but they will have a common cause. It is a prelude to a rifting event.

          It is not inconceivable that the actual eruption will be somewhere else, for instance Askja’s magma could move into the new gap.

          • Rifting in Iceland occurs along the Fissure Swarms. Herdubreid is between the fissure swarms of Kverkfjöll and Askja. So there should be no, or very Little extensioning going on in the area.

          • Carl, what you said is true, but we have seen numerous examples of magmas jumping fissure swarms. Obviously, Holuhraun is our obvious example, but just going off seismicity, it’s easy to see that there is a connection to Kverkfjoll and other nearby volcanic systems.

            I guess my point is, I don’t know if we can write off a relationship just because a volcano isn’t directly located on the primary fissure swarm.

        • Hello Mike!

          I will try to answer your question.
          The initial swarm and a couple of the following swarms was from the Upptyppingar volcano, but as you mention slightly offset on a fissure going from Upptyppingar. Upptyppingar and the nearby area in turn is a part of the Kverkfjöll Fissure Swarm, that is why you have the NNE alignments of the swarms as the occur along the fissures in that swarm.
          Here is where it becomes slightly problematic. As you notice it became seismically inert and “jumped” out of the Kverkfjöll Fissure Swarm and into an area that is not part of any known fissure swarm, but that has a number of non-fissure volcanic features.
          There is though a temporal trend from Upptyppingar towards Herdubreid, and there is also an upwards trend over time. I Think there may be a “dead zone” area here, but I can’t show any proof of it.

          • How can we define a fissure swarm? Obviously the quakes at Herdubreid are following the same NNE alignment like Kverkfjöll, Askja or Holuhraun do. What is the difference? The depth of the fracture?

            What is more, it seems like they are connected to each other – so can we say that there may even be a connection between the Aska fissure swarm and the one of Kverkfjöll?

            And I have struggles finding a site to translate Upptyppingar – giggle translate does not work.. I took a grasp of the direction it points (pun intended), however I would like to see the actual translation 🙂

          • The actual name might not be that funny. In Icelandic it is common in some cases to replaces O by Y, and in this case it simply means High-Toppings.

          • Well, there is an incongruency with the name and it’s actual appearance.

      • Carl, you are presenting a rift as something fixed, like a mid-ocean rift where the plates cleanly separate. And at the volcano with funny name there is a very nice rift, as seen at the right in the topographical map. But rifts can move. The rocks between the old and new rift would be jumping plate, and that is common. Mid-ocean rifts can jump by hundreds of kilometers, with the fragments connected by transform faults. That has happened around Iceland too. If 100’s of kilometers is possible, 10’s of kilometers cannot be excluded.

        In Lurking’s ‘dead zone’, there are a number of rift scars, formed at different times. Eldgja erupted into what is the rift valley at this time, but there are older features parallel to it. It is like pulling paper apart: it is unpredictable exactly where it will tear. Once magma has solidified, the old rift can be welded shut and a new tear will happen somewhere else – probably near, but not in exactly the same place.

        Another way that I look at it is to imagine a dyke trying to find the path of least resistance. It may follow the existing rift exactly, but in fact the crust is equally weak nearby so it may just run parallel but offset to the rift. This heats that area and weakens the crust. The rift may jump to the dyke.

        Of course this is rather speculative regarding this swarm, but rifting may happen away from the rift.


    1) If one follows the alignment of the 1961 fissure eruption at Askja, then 15km later, one arrives at the present Herdubreid dike. This shows the possibility that the present intrusion is still Askja magma.

    2) Interestingly, if one extends the alignment of the 2014 Holuhraun dike into Askja, it comes at the 1961 Askja eruption site.

    3) However we should not discount the possibility that the intrusion at Upptyppingar in 2008 crossed jumped into Herdubreid too, as the tectonic alignment in the region seems similar to a fracture region: faults parallel to each other towards the northwest. Upptyppingar itself is in a region with many fissural volcanic features that seem to originate from Kverfjoll fissure swarm. Many of these are Pleistocene, and run all the way to the north coast of Iceland, located to the east of Tjornes (not shown in the map).

    4) In this region, in the north coast of Iceland (not shown in the map), one finds also a small Holocene lava field, that supposedly belongs to Askja, as it follows the alignment northwards of the 1875 Askja dike event. This shows how far can dikes travel. The 1875 fissure eruption site was located more than 60km north of Askja caldera.

  19. For TJElJohnny:
    “And I have struggles finding a site to translate Upptyppingar – giggle translate does not work.. I took a grasp of the direction it points (pun intended), however I would like to see the actual translation 🙂”

    Giggle does work, sort of: Don’t enter the name as one word, but as two words, separating the first syllable (Upp) from the rest That gets a result 🙂

    • Sort of related to the idea of some words originating as run-on sentences or phrases?

      island mountain glacier → Eyja fjall jökull

      Months ago, someone with first hand experience noted how the mountain looks like an island poking up above the plain. So, the glacier on top gets that name. The glacier on the mountain that looks like an island. Linguistically it’s a nightmare for those of us who are non-icelandic, but the construction of the word makes perfect sense.

    • “A penis shape”! But interestingly, if you separate out “pingar” into it’s own word you get “dripping” for that. What does this say about Icelandic?

      • It says that like most of our languages, words and phrases can be very dynamic in meaning. “Some settling may occur” typically means “mostly empty” in the grocery store.

      • Oh bother! Sorry! I get NASA updates and it includes ESA, so I didn’t stop to check which source. Not enough coffee in me yet – too early for me! Thanks for spotting my mistake!

      • It still would be an interesting cage match…

        They could even do it old school… with slide rules.

        But I doubt ESA would run into the problem of mixing units of measurements in the calculations…

          • This sort of problem is far wider than you might think in software in general. Data structures are swapped between processes (eg subroutines) that are increasingly complex. That, and undefined conditionals (if..then) which absolutely cannot happen (but do) probably account for almost all the intransigent software bugs which are hard to locate and are often serious and apparently random. If the data structures are not checked to be identical problems will occur even if ‘that’s impossible’.

            Some languages are better than others but tend to have a steep learning curve. The best known is Pascal (and its son Delphi) although others can be set to have rigid type control – or not! When competent in Pascal (also very good for hardware control systems) it becomes normal for programs to pretty well run perfectly first time.

            I was always amazed when this happened….

            These are now pretty well replaced by easy-to-start but poorly structured languages like C and python.

  20. Askja must have known that it was mentioned. A 2.1 yesterday.

    23.03.2017 07:11:36 65.040 -16.771 2.1 km 2.1 99.0 7.8 km W of Dreki

    From IMO.

  21. Hi,
    Now we get depth 1.0km
    25.03.2017 07:45:13 65.172 -16.361 1.8 km 1.1 99.0 0.6 km WSW of Herðubreið
    25.03.2017 07:41:26 65.159 -16.374 1.0 km 1.1 99.0 2.0 km SW of Herðubreið
    Should i book a flight to Iceland in the following weeks?

Leave a Reply