The State of Affairs in Vatnajökull

Grimsvötn 2011 eruption. Photographed by Jóhann Ingi Jónsson.

Grimsvötn 2011 eruption. Photographed by Jóhann Ingi Jónsson.

Lately there have been a few earthquakes with a magnitude larger than M3 at the ring fault around the Bárdarbunga caldera. This has led to a lot of question about that particular volcano, and that is to be expected after a medium sized Icelandic eruption. They do tend to attract attention.

This has been coupled with an erroneous factoid that Bárdarbunga is inflating at a rate of 10 centimeters per year. In reality this factoid was caused by selective data-picking, GFUM-station (Grimsvötn) was omitted from the dataset. If you omit that station the data indicates 10 millimeters inflation per year at Bárdarbunga.

After reviewing earthquake and GPS data I think it is more furtive to take a broader look at what is going on under the Vatnajökull, there are after all more volcanoes there.


Image from University of Iceland and Dr Sigrún Hreinsdottir. The data does not support inflation at Bárdarbunga.

Image from University of Iceland and Dr Sigrún Hreinsdottir. The data does not support inflation at Bárdarbunga.

This volcano is a steady supply of Jökulhlaups caused by the geothermal energy release into its twin caldera. Sometimes it tends to have brief eruptions or suffers from phreatic detonations. It has been relatively quiescent after the medium sized eruption from Bárdarbunga. One interpretation would be that the eruption lowered the magmatic pressure in the volcano. If that is true we can expect a quiet period from this volcano.

On the other hand, there were a few deep earthquakes indicating fresh magma entering the system prior to the Bárdarbunga eruption, so there could be some small scale activity upcoming in the future, but nothing seems to be imminent.


Image from University of Iceland and Dr Sigrún Hreinsdottir. The dataset shows current rapid inflation at Grimsvötn.

Image from University of Iceland and Dr Sigrún Hreinsdottir. The dataset shows current rapid inflation at Grimsvötn.

If we look at the proven data at hand, the earthquakes, we get a different picture of what is going on. But, let us start with looking back at the events prior to the onset of the eruption. First we saw an intrusion coming up from depth north of the Bárdarbunga volcano, that intrusion in turn changed direction and headed almost dead south and entered the magmatic reservoir of Bárdarbunga. That reservoir was already teetering on the brink of an eruption and the fresh magma caused the reservoir to rupture in the ENE side and a dyke started to propagate.

The ensuing eruption caused a massive drop in pressure and the caldera floor fell like a gigantic plug as magma moved out. As the eruption dwindled the subsidence dropped accordingly until it stopped.

Now it would be easy to believe that the volcano would start to inflate immediately. But that is not how it works. Before any inflation can occur the pressure must first build, and that takes a long time since the magmatic inflow is not consistent, instead it happens in fits and bursts, almost always shown as earthquakes at depth or around the magma chamber. So far, there is very little evidence of this taking place.

So what then are those M3+ earthquakes? Well, it is just a bout of resettlement earthquakes after the medium sized eruption. The very shallow depths seem to indicate that it is the edges of the caldera plug that is settling downwards a little. This might be caused by seasonal ice pressure, it is after all winter in Iceland.


Image by the Icelandic Met Office. This is the best representation of data when looking for a probability for an eruption. As the cumulative seismic release nears the threshold the likelihood of an eruption increases exponentially.

Image by the Icelandic Met Office. This is the best representation of data when looking for a probability for an eruption. As the cumulative seismic release nears the threshold the likelihood of an eruption increases exponentially.

Well then, where did then all the inflation hoopla come from? Well, it is absolutely no GPS-data to support any 10 centimeter per year uplift at Bárdarbunga. But if we look at the usual suspect Grimsvötn we find that uplift. And that is not unexpected, after all Grimsvötn has the highest and most even rate of magmatic intrusion on Iceland.

When Bárdarbunga put in an appearance Grimsvötn was already busy inflating after the VEI-4 eruption of 2011. As the Bárdarbunga eruption happened it momentarily released tension in the magmatic reservoir of Grimsvötn, but as soon as the eruption ended the pressure continued the build up towards the next eruption.

Grimsvötn is normally fairly quiescent at the beginning after a big eruption, but the inflation is almost linear. This seems to imply that the inflation itself is almost aseismic, but as the pressure increases the magma reservoirs start to “creek and groan” as the reservoirs surrounding bedrock start to fracture with the mounting pressure.

The inflation-rate and a slight uptick in seismic release seem to imply that we are closing in on the next Grimsvötn eruption. GPS data points to it being close, 6 months to 2 years away. On the other hand cumulative seismic release data points to an eruption occurring 3 to 5 years from now. It is though a good idea to remember that the cumulative seismic release plot can increase very rapidly, and I think that it soon will do so.

If you are waiting for an Icelandic eruption happening, the best bet is Grimsvötn. As we get closer to the eruption I will be able to give a better prediction about when it will go off.


P.S. During the weekend the Whizzards of all Things Technical will do some changes and move the place to it’s brand spanking new permanent home. Hopefully you will not notice any difference at all, except a faster better Volcanocafé. D.S.

120 thoughts on “The State of Affairs in Vatnajökull

  1. Great article. Noticed there’s been a continuing swarm of deep quakes in the Tjornes fracture zone.

  2. Really interesting and informative. Thanks, Carl. I seem to have lost my CGPS links, so I’ll have to have a proper search; meanwhile, it’s nice to have Dync, etc. displayed.

    To the Technical Whizzards: have “New Post” emails been abandoned? Although the website seems to recognize me automatically, if emails are sent out, I’ve not been getting them.

      • I’d like to register a complaint to advertising standards: your post was niether verbose nor laconic.

        Pah! 😉

      • I’m probably being dense, or missing something obvious, but I’ve had a dig around and can’t see how to unsubscribe.

        However, not being able to escape Volcano Cafe isn’t the worst thing I could imagine. I mean, freshly grilled lamb, shark steak, lava cocktails and water straight from Carlavatn – who could wish for more? And apologies to verbosely laconic if what passes for my sense of humour was ineptly deployed.

        • I’m not able to dig around at the moment for the link but if you look at the bottom of any VC mail you’ve recieved it should have a link to subscription options

          • Thanks for that suggestion. I *think* I’ve made the right changes, so we’ll see how it goes.

  3. I see where Barda temporarily halted Grims seismic increase, but what halted it pre-2004 (at roughly the same time [+1500 days] as the post-2011)? Was there an eruption under Vatnajokull at that time as well? If so, maybe the correlation runs opposite as well (pressure from Grims forces an eruption elsewhere)?

    I know, I know, “correlation is not causation” all the time – but sometimes it is. Sometimes the paranoid have reasons to be so.

    • Good question actually. I will have to mull it over for a while. Think weeks here.

  4. Carl

    Great summary on the current status!

    Interested on your thoughts regarding the KISA GPS placed on the northern rim of the caldera. I see the vertical position is all over the place per usual in Iceland, but there is a significant amount of horizontal movement.

    I still think BB and Grims are connected, the evidence points that way

    • The dataset is not detrended so the swings are due to isostatic shift and other phenomenon. Also, it is a very short timeframe, but on average it gives no clear uplift.

      In regards of the connection. There is a fault running between them. It caused the Gjálp eruption. But, the latest eruption indicates that there is no direct magmatic connection.

    • Double whammy for the Aleutians, because flank collapses of Aleutian volcanoes have probably caused some pretty hefty tsunami as well (although mostly on the Bering Sea side of the chain)

    • I was poking around looking at the video linked in one of my old posts. It was the numerical recreation of the tsunami from the 1700 Cascadia originated tsunami. What I noticed, was that along with catching destruction from the tsunami as it passed, the Aleutians also get hammered a few house later by a massive reflection off of the Hawaiian–Emperor seamount chain.

  5. I found the cumulative seismic moment plot from Grimsvötn very interesting, so I did a similar one for Bardarbunga for comparison. Quakes were selected from the region 64.54-64.8N, 17.2-17.8W, starting from March 2015 when the eruption had ended. I also included a magnitude plot for all quakes and finally a magnitude plot only including quakes deeper than 15km to be able to see possible magma intrusions.

    Data set was extracted from IMO’s week by week “skjálftalisti”.

    • Nice plotting there.
      One thing to remember is that there might be differences between the two volcanoes, for one thing Grimsvötn is a high frequency eruptor that seems to go really quiet the first couple of years after an eruption compared to what we are seeing from Bárdarbunga.
      On the other hand we have no data from Grimsvötn after a large effusive rift eruption, except Skaftár Fires that pretty much put Grimsvötn to sleep for a while.

  6. “Hopefully you will not notice any difference at all, except a faster better Volcanocafé.”

    Would that include replacing that goofy temporary banner with the real Volcanocafé banner? Also a lot of content is still missing, including the entire Dragon’s Hoard section.

  7. One word folks…. Thorium. I believe Carl knows what I’m talking about. The game may well have changed in 2011 and the last couple of hundred years may not be a good guide to the next twenty years…

    • Mind explaining or teaching? I’d understand that there has likely been a change in magma composition, but would be nice to get what it actually means without having to sift through tons of research papers.

        • I just don’t get the point of intentionally teasing information and not actually explaining whats going on.

          • Sorry I’m on vacation with kids; limited time and typing on phone. Not teasing.

            I was referring to the paper a couple of years ago on historical thorium numbers for Grimsvotn magma. Summary: Laki magma was low in thorium. Each succeeding eruption has shown increasing thorium. This is interpreted as due to fractionation in crustal magma reservoirs. This trend continued until the penultimate Grimsvotn eruption in… 2005? In a geochemical sense, that eruption marked the final end of the Laki eruption.

            Grimsvotn 2011 was back to low thorium; this is interpreted as the system starting to erupt a new batch of very juvenile magma ‘straight from the mantle’ for the first time since Laki. That’s obviously an observation of some significance!

      • It is one of those occasions when the subject is a bit trickier than what can easily be explained in a brief comment.

        Since the concept is generally misunderstood for Icelandic purposes I will at some point return to the discussion again. I did though briefly delve into it as I used the Th in conjunction with negative Pb and positive Nb and Ta anomalies to predict Bárdarbunga and wrote about the upcoming eruption.

        But, the subject is so complex that one of these days I will have to get back to it.

        Hm… Nah, it is all a constipation to hide that it is the WHO that causes eruptions 😉

    • On a slightly unrelated note – does anybody have a link to the paper that details the mechanics of caldera collapses / edifice failure with regards to the size of an overlying cone and the depth of a magma chamber?

    • Hello Mike, yes I am very well aware of the difference in thorium and its possible implications for the future eruptions in Iceland and its deeper meaning.

      • Hi Carl (or Mike, Henrik or Albert), Can someone simplify that into short “laymans” terms what that actually means whether for Grimsvotn, Vatnajökull’s volcanoes or Iceland in general. Falling presence of thorium in lava @ Grimsvotn = potentially ? *Struggling with this one for the linked article, keep thinking it is implying Grimsvotn is about to go into the autumn years of its cycle as it winds down as a volcanic system. I’m fairly sure im reading this wrong but hey any help understanding this would be appreciated 🙂

        • Trying to explain it in a comment is just not feasible. It is a rather complex concept with widespread implications for Iceland geology.
          So, I reluctantly find myself on the road to writing an article about it. Reluctance stemming from it being a very difficult subject to explain in a simple manner and in a proper context.

        • Here would be my take (be warned, I’m a total numpty on such matters). Thorium is one of those elements which does not get taken up to any great extent in the crystal lattices of rock-forming minerals, so fractional crystallization will gradually increase the proportion of Th in the remaining melt – which is what has been happening at Grimsvotn since 1783. An injection of fresh magma, however, would have a Th content of the ‘primitive’ source material, much lower than in the part-crystallized older melt.

          Of course, that would seem to suggest that Grimsvotn, far from being in the autumn of its days, would be seeing a spring-like rejuvenation, which could be uncomfortable from a human perspective

  8. Well, Carl, the ‘hoopla’ can be placed very well.
    The meeting of the Monitoring group in which possible inflation under B was discussed, was half november…
    Looking at the available GPS data of stations in the vicinity of B., there could be seen a slight uptick, after a longer period of deflation. So, not strange, the monitoring group was alert to the broken trend.
    Now, half januari, GPS data show no prominent inflation any more.

    The onset of the Holuhraun eruption began with deep quakes, just north of B. Actually there have been (I think two?) serie(s) of deeper quakes north of B lately. One of them 15/11/15 at 26.3 km depth, most of them about 17 km. I learned, not every unrest at depth triggers an eruption, But it has been said many times also: this is a very well monitored rifting event with a huge pile of data and no one knows exactly to state what will happen next.

    Tomasandersson 2’s plots above are interesting, not only shallow quakes!

    It is fun to speculate a bit about the next move. My unexperienced amateur feeling is attracted more to the ‘failed” dyke intrusion in the Herdubreid-area north of B.

    VON GPS data (source
    Until half november (about the time the IMO monitoring group did discuss) a bit of an uplift can be recognized, after that the numbers do not confirm such trend.

    • I was preocupied back then, but even if I go on the Mid november data there was no real uplift.
      I would though write it off as a very very rare bad day at work on IMOs part. Come think about it, I think this might be their only mistake I am aware of. And that says a lot in their favor 🙂
      I have tremendous respect for them and their work is A-grade.

        • Thank you.

          Just to reiterate. I have the utmost respect for IMO. They do excellent work, but this time they jumped the gun a bit (first time I think).

          • I don’t think they have “jumped the gun” at all,if anything they are being cautious.But the evidence seems to suggesting a shallow magma intrusion ,something that I was theorizing over a year ago.There is no shield volcano,Holuhraun was just a symptom of a very large event occurring in the caldera.My guess a crypto dome of evolved magma.

          • Bardarbunga is by definition, a shield volcano, so I’m not sure why you are saying there is no shield volcano. That being said, defining a volcano as a shield, strato, caldera, etc doesn’t really accomplish much in describing the characteristics of a given volcano.

            Also, I strongly doubt a cryptodome of evolved magma is what intruded into Bardarbunga. Icelandic magma is always strongly basaltic when it comes up from depth. So any magma intruding into the caldera’s shallow chamber was likely similar to the magma we saw at Holuhraun. Rhyolitic magma in Iceland forms mostly via fractionation, so any evolved magma that may be in Bardarbunga would have already been there, not a new intrusion. In a general sense, lava-doming isn’t something you see in Iceland too much.

            Also, there is likely a shallow intrusion into Bardarbunga, but this isn’t anything to get all that excited about. After eruptions (such as holuhraun), it’s 100% normal for any volcano, and especially a large caldera to being refilling and rebounding. After all, there is both an active magma source below Bardarbunga with buoyant basalt, and a big accumulation region that recently evacuated a decent amount of magma. There is a decent chance we won’t see another eruption until the caldera reaches a fill level similar to the volume it had prior to the Holuhraun eruption. With that said, the pathway to Holuhraun is still hot and somewhat “open”, so we may see a series of basalt eruptions over the next few years here, but the subsequent eruptions would likely be on a smaller scale than the initial eruption.

            On the whole, I’m not really expecting anything too exciting to come from Bardarbunga in the future years. But there are a lot of other volcanoes around the world (and in Iceland) that promise more excitement, so it’s all good.

          • cbus05,is not all magma from depth,basalt.So Rhyolite anywhere in the world has formed from basalt at shallow depth?

          • @cbus: No, it is a compound bimodal magma caldera with ring fault volcanism. Far from a simple shield volcano.

            @geysesoze: There is an entire field of science called petrology. Start reading Carmichael and evolve from that.

          • The point I was making is the rhyolite is already in place and is being “energized by the massive influx of more primitive magma ,causing sub surface dome formation.Some of this influx of basalt erupted at Holuhraun and the rest interacted with a possible shallow rhyolite formation,basically kick stating the Caldera into life after period of dormancy(thousands of years perhaps?), in the Caldera part of the Bardarbunga system.Imagine an animal that was once asleep,now it has one eye open.

          • Much better answer geysersoze. Mostly I agree. But currently the pressure is to low. In a few years it is entirely possible that we will have an eruption at Bardarbunga as per your argument. Now, make the leap and remember that Grimsvötn is basically a more frequent erupter than Bardarbunga and you will understand that it is exactly the same argument I made about Grimsvötn.
            One other thing, not all Icelandic volcanoes start with unevolved magma. Hekla is one example.

          • @Carl: My fault then – I had thought I read in the GVP that Bardarbunga was commonly referred to as a shield volcano.

            Regardless, there are a lot of shield volcanoes out there that do not behave like shield volcanoes in the traditional sense. The only real defining characteristic of a shield volcano is its shape.

            @GeyserSoze: Yes, most magma starts out as basalt from depth, but Iceland is different because it does not sit on top of a continental margin (although there are some people who believe a fragment exists somewhere down there). Magma that rises in a traditional arc-based setting, or in a rift such as Africa will melt, and incorporate a sizable amount of continental rock, which leads to more evolved and gassy magma before the magma ever even reaches a magma chamber. As an example, Taupo and New Zealand’s caldera systems produce their prolific volumes of Rhyolite largely from the basalt melting the surrounding continental rock.

            In Iceland, there is no continental rock mass to melt. It’s all old basalt for the most part (I believe), so you don’t get the same type of assimilation. So in other words, on most normal continental margins, evolved magma can be produced either through fractionation, or through assimilation. In Iceland, you only really have the option of fractionation to occur (not counting some oddities like Hekla), which is a long slow process.

  9. Glad to see you back, Carl, and I hope you are recovered! Thank you for an interesting post.

    • And here is FED down below.
      I can’t find anything unusual on the borehole strainmeters. Nothings seems to point to earthquake activity.
      But, there are micro-seisms on FED. Could be rocks falling, frost-quakes, pretty much anything.
      If we get a couple of quakes and a bit of ruccus on the borehole strainmeters I will become a believer, but I am not one just yet 🙂


      And this is how it would look on a SIL-plot.

      • For those who are not familiar with Hekla and her antics. She is a downright nasty piece of work.
        Before an eruption you get a few small earthquakes, then either the volcano will do nothing, or 30 to 120 minutes later all hell literally breaks lose.
        Hekla does not do prolonged runups to an eruption, this is why it is bad mojo to climb that particular volcano. If anything happens you will not have time to get to safety.

        She is like an old drunk, she wakes up with a hangover, takes a swig of beer and projectile barfs all over the place before passing out again. Totally unpredictable and very violent.

        She also enjoys to have false ultra-short runups on a yearly basis. You get a quake or two, then the strainmeters start to go haywire. Then nothing. Do. Not. Climb.

        • Doesn’t help hold the record for the longest lava bomb kill?

          • In October, me and my wife went to Iceland for our honeymoon. We rented a car and drove around, planning one day at a time. At one point, we had been following instructions from one of the locals to find a beautiful waterfall, a bit off the normal tourist tracks. We were not using a GPS, but instead used a not so detailed tourist map, where the helpful local had circled a few interesting spots. It turned out that the circles were a few kilometers off, and we finally had to give up searching for the waterfall (which we later found out was Háifoss, Icelands 2nd tallest waterfall, which we would have found had we only continued 2 more km on a narrow gravel road). Anyway, we headed back for civilization and suddenly found ourselves in the middle of a stony desert, completely littered with old lava bombs. Twisted, grotesque black shapes, like an army of trolls just waiting for night to fall. I realized we must be very close to Hekla, but it was starting to get dark and visibility was low due to fog, so I never really saw her. When we finally found a hotel and more importantly, a wifi connection. I checked the maps and realized that those lava bombs were something like 20km from Hekla. I wouldn’t want to be anywhere near that lady when she decides to pop.

          • Hekla is the only volcano in the world I have been close to that gave off a visceral gutt fear. I have been on top of Pacaya during an eruption, I felt safer there then I did 10km away from a quiet Hekla.
            Somehow it projects fear and dread like no other volcano.

  10. I’ll have to go and shout at the mountain! But I’m going to keep an eye on her. Thank you for input – some good comments!

  11. I get more worried about Hekla than any of the other Icelandic volcanoes. It has thrown some of the biggest eruptions in Iceland’s volcanic history.

  12. The red points below are the new series of Bardarbunga earthquakes with M>2. The black dots are the ones from 2014/2015. The new quakes comes from the same ring, but tend to come from around the densest concentrations of the old quakes, not within. It is the area around the broken-up rock which is now breaking. There may also be a trend that the new quakes are located just on the inside, towards the centre of Bardarbunga but there is not enough data to make that a strong conclusion. Just something to keep an eye on.

  13. The advantage of a more frequent erupted system ,particulary if it has erupted often in recent history,is that there is some idea as to what to expect.With Bardarbunga,the Caldera itself has possibly not erupted since the end of the last ice age,so there is very little idea as to what to expect either in precursors or in the event of an explosive eruption .Fortunately the chances of such an eruption in the average human lifespan is extremely unlikely, even in a seemingly active system.

    • The thing is, you just said it yourself. We DO know roughly what to expect from Bardarbunga, because it HAS happened quite often in recent geological history. That being fissure vent eruptions similar either to a rifting fissure eruption in the dead zone, or to Holuhraun. We can’t discount those eruptions simply because they chose to escape from lateral fissure vents as opposed to the vertical lid.

      Realistically, the fissure eruptions are most likely the release valves that prevent a larger caldera-style eruption from happening.

      • I was not referring to an eruption in the Bardarbunga system as a whole,the most recent was 2014.I was specifically referring to an eruption from the Caldera itself,which could possibly based on other volcanoes in Iceland contain rhyolite,formed by the processes you describe over millennia due to the influx of basalt in a subglacial volcanic system.

      • “…likely the release valves that prevent a larger caldera-style eruption…”

        … or were the draining mechanism that allowed for a slower, less catastrophic collapse structure to form…

        Sort of like lancing a boil.

        • That theory is fine if the basalt that erupted was from a shallow chamber in the caldera,in essence the basalt was holding up the Caldera floor/roof.I think the general consensus was the Holuhraun magma was deeper source,so likely the shallow magma “chamber “in the Caldera is still in place.What was observed and interpreted as a sinking caldera maybe misleading,sometimes what appears to be a correlation maybe due to other factors.Counterintuitive.

          • Also there is no pressure equation here,an eruption of a system like this would be due to the “perfect storm”of various factors,almost an accidental failure of a fail safe system,that’s why explosive caldera forming events have periods of thousands of years even in the most active systems.The “pressure”in these systems is always there but rarely relieved explosively without a trigger,ie massive inFlux of magma or large seismic event and even after these triggers any eruption could take years to occur.

          • *preface* most of what is below is simply my personal opinion.

            Caldera eruptions do occur in various fashions, but I wouldn’t say they’re simply a random alignment of variables. And saying there is no pressure equation here is simply wrong. Pressure always plays a role, and is probably the single most important factor (both pressure, and depressurization to be more exact).

            I think the thing you’re maybe misunderstanding about Bard, is that there is likely a pretty constant supply of magma entering into the system, at least in a geological time scale. I personally tend to look at it like a big water pump, except… with magma. Whether the magma is being drawn out by tectonic forces, or if it’s being pushed up from depth by pressure from below is an argument I don’t believe we ever figured out the answer to, but either way, this isn’t a volcano that is freshly “waking up”. Bardarbunga has never been asleep.

            The scenario you present isn’t entirely innacurate, but it would require a large emplacement into a relatively cold, yet pressurized system. Bardarbunga isn’t a particularly cold system in all likelihood, and it also has the issue of any fresh basalt injection seeming to drain right out the side vents into a rift once the injection occurs. If we were to see a prolonged injection and inflation into a volcano such as say…. Tungnafellsjokull, that would more likely be the scenario where you would start to think about evolved magma being activated into forming a bit more of a sizable explosive eruption.

            That being said, I certainly wouldn’t write off a central vent eruption, but I think Bardarbunga depressurizes and releases built-up heat from fresh magma via fissure eruptions too easily, unlike Grimsvotn or some other Icelandic Volcanoes. But there are a lot more factors at hand than I and even the experts are aware of here, and it’s what you don’t know that can catch you off guard.

          • Might help… might not.

            Types of collapse calderas

            They categorize Taupo as a graben style caldera. There is one more style that they don’t list. I’ve only seen it mentioned in regards to a particular Japanese caldera that resembles, in profile, what I think a kimberlite eruptive pipe would look like.

          • I should clarify,”super eruption” in so far as the violence and scale for a basalt eruption.Erupting a similar volume to Holuhraun in the space of 6 hours and fountains of magma to the order of hundreds of meters in height,so compared to a largesilicic caldera eruption ,not a super eruption no.But of an extraordinary nature for its type.

          • Its interesting that you bring up Tarawera. There is a theory that I and a few others on here believe that Tarawera in a few ways could have been a failed super-eruption.

            Following this theory, the primary reason for which Tarawera “failed” as a supereruption is that the basalt intrusion escaped to the surface instead of stewing underground for an extended period of time. Regardless of the end result, it was a very good example of how these events possibly occur, and the type of basaltic injection of magma that may be necessary to set off such a massive event. It’s also an important note that one of the primary requirements of such large eruptions would probably be that a magma injection be constrained and concentrated, enough so that it can ignite a large block of crystal mush so that it becomes mobile again (which takes a TON of energy by the way).

            I guess the takeaway here, is that similar to how the basalt injections at Tarawera escaped to the surface, the basalt injections at Bardarbunga escape out into the fissure vents, preventing any large-scale activation of any of the potentially evolved magma that may hypothetically be sitting in an upper chamber.

            Another thing worth mentioning with regards to the speed in which these events occur, is that the depressurization of an over-pressurized system likely plays a huge role in how these things build. When you remove the cork on a bottle of champagne, the highly pressurized fluid beneath exsolutes and rapidly moves into the lower-pressure zone outside the bottle. This is a common analogy to volcanoes, although the critical piece that is often missing here is the removing of the cork (from a volcano’s perspective).

            In order to truly depressurize a mass of magma, you need a sizable vent open to the surface. In order to blast such a sizable vent, you need a LOT of pressure in the underlying magma to do such a thing. Also, moving a large volume of magma through a narrow conduit is extremely difficult, even if an opening is available. This could possibly be the reason why many stratovolcanos have strombolian or dome-building style eruptions: Simply put, the friction and viscosity of pushing a crapton of magma multiple kilometers through a narrow conduit puts up enough resistance that it kills the chance of an explosive eruption (And subsequent depressurization to the chamber below). Most of the time, the pressure simply results in a slow push out of the summit, or occasional mid-sized eruptions. Provided with a constant input, eventually the conduit will become so long and narrow, and the pressure from below so high that something will have to give. This is where the edifice failure or caldera situation becomes particularly relevant, as a stratovolcano will then start to form ring-fractures and destabilize.

          • So in laymens terms, in order to create a caldera event, you need a mechanism to release the pent-up pressure within a magma chamber. The mechanism to create this event can vary depending on several factors, but it typically involves opening up a hole big enough to reduce the effect of viscosity traveling through the conduit.

            With enough brute force, an eruption would break the entire grounds around the conduit. But similarly, if you were to hypothetically remove the lid over just about any decent sized magma chamber, you would get a massive explosive eruption.

          • cbus05,maybe thinking that Tarawera was a failed super eruption is looking at it from the wrong direction.In other words a super eruption is a failure in its self,maybe Tarawera was an effective bypass of a rhyolite emplacement that was not critical,maybe these occurrences are the case 99.9% of the time and that failure of the system ,a full scale eruption needed that dyke to occur at the precise time a large percentage of the rhyolite was at a critical state.Monitoring these systems is akin to waiting for an extremely reliable mechanism for which there is lttle understanding how it actually works, to fail.

          • Precisely Geyser. How you label it doesnt really matter, but the idea is that if the basalt injection didn’t cut through a rhyolite bed to the surface, something worse could have possibly occurred. With that said, it’s simply a theory – there may not even be rhyolite in the area around Tarawera.

            And also, as you mentioned, cutting through to the surface probably occurs more often than the other scenario where it would ignite a larger eruption.

          • Tarawera is perdominantly a rhyolite edifice.It had a rhyolite eruption around 1200CE.The Taupo volcanic zone is very rhyolite dominant with some andesite,dacite and basalt.White island in recent history has been andesite in its eruptions but it has had dacite episodes in pre history.Iceland is the opposite mostly basalt eruptions over 90% but it does have rhyolite as well but these eruption are far less frequent.Iceland is more frequently active than New Zealand it seems but when eruptions occur in NZ excepting the andesite stratovolcanoes they tend to be very violent but separated by at least hundreds if not thousands of years,that is probably the consequence of rhyolite,very explosive but relatively uncommon.So a rhyolite dominant eruption in Iceland would be and exceptional case but not impossible.

        • In order to have an explosive caldera, you need some specific items to be in place. Mainly, a well sealed roof so that large quantities of well gassed magma can stew. Also, something has to add the extra pressure, be it a fresh magma intrusion into rhyolitic mush, water, or a roof collapsing into the chamber.

          At least one of Taupo’s eruptions involved a sort of sequential trap-door style failure as the sections of the roof fell in. It’s probably likely that a lake existed there at the time, and that gives you a lot of water thrown into the mix for good measure. What I don’t get through, is that Taupo was evaluated as to have had little to no zonation in that eruption, indicating that it was either well mixed, or it went up all at once. Whakamaru, an ancestral caldera to Taupo, was known to show zonation in it’s eruption. In effect, Taupo drew upon the same feed system that drove Whakamaru since their apparent chambers overlap.

          Given that other eruptions in this part of the world have had very sudden onset, the influx of fresh magma can be a quite rapid event even on the human scale, so I guess it’s not out of the question that something could have quickly drove the Taupo chamber into a critical state. Iceland, I don’t know. Hekla is the only spookily quick system there that I know of. It usually activates when plate motion suddenly releases confining pressure on whatever chambers it has stewing. (my take on it) I imagine it’s possible for that sort of mechanism to happen elsewhere along the plate boundaries. Holuhraun was funky in that IMO tracked the dike as it wandered from system to system before finally erupting.

          • Not sure, but I think this was the Oruanui eruption that I read that about.

          • I would also like to add, “super” eruption is a media term generated to drum up traffic for a TV program. Though mostly lacking a definition, it is generally accepted to be up in the VEI-8 range. That then lends it to the vagaries of the VEI scale. VEI is based off of total ejecta generated from a system with not bearing on exactly how long it took to do that. Let any volcano have a productive energetic eruption cycle that lasts long enough, eventually it can reach the VEI-8 realm since that’s just a matter of volume with no standard as to how fast the tephra is produced.

            Generally, the VEI classification is taken to mean singular explosive events, but I haven’t seen that written down anywhere.

            As for “super” eruption, I try to avoid the term. Personally, I prefer “large caldera event” since it is a bit more descriptive and a bit more encompassing, though it’s not much better. (where is the break point between “large” and “ordinary?”) How would you define “ordinary?” One person’s “ordinary” could kill half a million people… yet still be run of the mill? It doesn’t seem workable to me. What if that “ordinary” caldera was in the middle of the Auckland Volcanic Field? Not a whole lot of Aucklanders would be pleased with that. And then there is Hokkaido, another island that is prone to large caldera events over it’s geologic history. They would probably have issues with the “ordinary” designation if they happened to be the recipients.

            I guess, if we really needed a break-point for large and ordinary, any caldera larger than the average caldera size would be by definition, “large”. No, I don’t have a set to do averages from, though there are a few lists out there that could serve as a basis for one. At least it would be closer to an empirical definition. than the term “super.” Plus, you don’t have to rely on spotty tephra volume estimates. Just look for the giant hole in the ground and calculate the elliptical area of it. Then get a histogram of those areas and make everything above the average “large.” I had a spreadsheet that I was working on (but have lost) that was close to giving me a generalized formula to calculate volume relationship of the field determined tephra output compared to the elliptical area of the caldera. Eventually I’ll remember what reference I was using and find it again.

            And for any Auklanders reading this, no, nothing I have read indicates that Auckland volcanic field is close to doing anything. One thing I can say though, volcanoes tend to behave as they have in the past. Though Wikipedia notes that “the field is likely to erupt again within the next “hundreds to thousands of years”” note that it typically does a Maar style of eruption in a localized area or just make a scoria cone. Dangerous? Yes, but not the “blast you into the stratosphere” big explosion style. White Island might do that… eventually, if it is anything like it’s cousins. In that case; “Bye bye Dino…”

          • There seems to be some confusion with caldera and stratovolcanoes with magma chambers and conduits,as the posted graphic of Taupo shows there is a considerable difference.Another neglected reason an eruption is the temperature of the magma,this may not make much differecently with more primitive magma but with the evolved magma the temperature is critical.Rhyolite can push domes to the surface (low pressure environment) and not erupt.Low gas content and low temperature,no violent eruption.But this viscous mass below the surface must have a huge energy potential compared to a more fluid basalt resulting in earthquake swarms.

          • No, no misunderstanding at all. Calderas are collapse structures. How whatever was in there got out is the main issue. I can see Calderas forming as magma is evacuated from the chamber via an opening or weakening rift line. Its also quite possible we are seeing the latter stages of that as bardabunga settles after the holuhraun eruption.

    • Interesting discussion. To erupt from the top, you need pressure from below. In gas-rich magma, that is supplied by the volatiles which try to get up explosively: as its pressure is so extreme, it just goes for the shortest path to the outside world which tends to be through the caldera or the top (but may be on the side,m as in Mount St Helens). If the caldera collapses, you can also get an explosion upward. In gas-poor magma, the pressure is less and is pushing in all directions. I will follow the path of least resistance, and that depends on the tensile strength of the surrounding rock, the weight of the rock above, etc. If there is a big rift next door, as in Bardarbunga, the tensile strength may not be very large and it will try to go there. It depends whether there is an easy path available to get to the rift. The 2014 eruption followed the contours of the mountain very well: at every point, the dyke followed the course corresponding to steepest descend on the mountain above. It was the weight of the mountain which pushed it out. A caldera eruption in Bardarbunga is not impossible, but history shows that it is rather easy for its magma to get to the rift and it normally does that.

      That also tels you that there is a link between the rift and the caldera. I have doubts about the argument that the caldera deflation and Holohraun eruption were separate events. They were too well correlated, up to their end which was the same within days or a week.

      • With my limited knowledge on this subject, this seems to me to be the most logical speculation on the BB setup underground and all of the evidence points to this as well.

        To me, with the amount of activity, energy and shaking around that happened from August to March….a few A-bombs worth…. and with nothing happening in the caldera thermally, other than possibly an slight increase in geothermal activity, it would suggest that BB caldera is relatively cold and dead with too Many easier paths for magma to be take tan upward through the middle.

        • How about most of the time the caldera is “cold” except when the threshold for release ofthe energy from rifting magma is exceeded thru other means.In other words the caldera is the final fail safe and becomes more active when too much heat enters the system from depth.The glacier plays a role as an extra moderator as does the lake in a temperate caldera?

          • If pressure increased in the system to that degree and I had to pick between the caldera eruption or fissure, I see 1000m high walls of lava shooting into the sky as the fissure at Holohraun rips open wider than it is today before anything happens anywhere else….I think an eruption at Gjalp is more likely the path of lesser resistance for any pressurized magma

          • Possible! The larger the excess pressure (which can be related to the magma supply rate or to heat carried up), the more tempting the ‘direct route’ to the top. In 2014, it is possible there was also a minor eruption at the caldera, underneath the ice. Some volcanoes start with a summit eruption, and as the pressure drops the eruption migrates down along the sides. I wonder whether the easiest way to get a caldera eruption in Bardarbunga is by global warming. Once the glacier has melted (which may be inevitable) the decrease in weight can generate more melt.

            Grimsvotn exists in the same environment but behaves different?

  14. Thanks Carl for a very interesting and exciting article. Nice to have you back. 🙂

  15. Thinking about Cbus05’s comment: “but it typically involves opening up a hole big enough to reduce the effect of viscosity traveling through the conduit” what springs to my mind is earthquakes. A caldera even under high pressure can stay plugged for ages, but an unfortunate tectonic / faulting earthquake might be that one item which can ‘crack the lid’, so to speak. Water systems in the caldera and phreatic events could also be a factor. Just my luddite’s input so feel free to ROFL! 🙂

    • Great discussion up above about large calderas.
      I will soon publish an article about a medium sized caldera that is able to become a large caldera over time.
      I think you guys will have fun then 🙂

    • I’m not really sure I’d say earthquakes are what you would need to create such an event. Earthquakes can make cracks, but I can’t think of any examples in which they single handedly opened up a volcanic conduit. With that said, they can definitely shake things up a bit – there is some analysis on Chilean volcanoes that puts a loose correlation of massive 8+ earthquakes (away from the volcano) and subsequent eruptions, but none of these are the type you’re thinking of.

      Think of it this way – in any given caldera lid, there are already tons of cracks. For a volcano that has already had a caldera event, these exist as the ring fault, yet most of the time they simply act as the region for which magma would percolate out to form new somma stratovolcanos.

      (Edit Add: Think Mount Tondano → a really good example of volcanoes on ring faults of larger calderas; Soputan, Sempu, Lokon-Empung and Mahawu)

      Going back to the champagne bottle example – think of the cork as the caldera lid. The space between the cork and the glass is technically speaking, a crack in which the champagne could exit through, but it is too well-sealed for this to ever happen. The same is true for most caldera lids. Provided enough pressure, some of it may push through the crack and leak out, but the crack is still for most important purposes “sealed”.

      Now, if new cracks were to form around an active stratovolcano that has not experienced a collapse event of sorts, that’s a different story as it would possibly signify that the central conduit or escape route is not adequately relieving pressure, and therefore the magma is searching for other ways to escape.

    • Carl – seeing that you’re relatively close by, I’m curious as to what your opinion is of Agua volcano. It seems to be a fairly un-studied volcano.

      • I am actually typing on a piece that involves that particular volcano.
        So, just wait a little while and there will be an article that covers all things about Agua (and more).

          • Likewise. The resemblances to Santa Maria (symmetrical, youthful looking, but no sign of life for 500+ years, then Ka-Boom!) are enough to be worrying.

          • There were lahars that came from Agua volcano in historic times that caused a catastrophe, and they weren’t even formed from an actual eruption.

  16. This is about where the dike turned, isn’t it?
    16.01.2016 14:02:39 64.595 -17.183 11.3 km 0.7 99.0 17.2 km ESE of Bárðarbunga
    16.01.2016 14:01:59 64.598 -17.215 16.0 km 1.2 99.0 15.6 km ESE of Bárðarbunga
    16.01.2016 14:01:48 64.599 -17.190 14.8 km 2.4 99.0 16.7 km ESE of Bárðarbunga
    16.01.2016 14:01:21 64.593 -17.209 17.4 km 1.1 99.0 16.1 km ESE of Bárðarbunga
    16.01.2016 14:01:10 64.597 -17.199 12.6 km 1.2 99.0 16.4 km ESE of Bárðarbunga

    • Yes, and I find this area interesting for more reasons than that. If you go back in time and check the earthquake maps, you will notice that this area has been active long before the Holuhraun event started. I’m not sure what that implies, but maybe someone has the answer.

  17. Operation Fork

    At 01:47, Icelandic time, on 10 May, HMS Berwick used its catapult to launch a Supermarine Walrus reconnaissance plane. The principal aim of the flight was to scout the vicinity of Reykjavík for enemy submarines, which the Naval Intelligence Division had convinced itself were operating out of Icelandic harbours. The Walrus was given orders not to fly over Reykjavík but – either accidentally or as the result of a miscommunication – it flew several circles over the town, making considerable noise. At this time, Iceland possessed no aeroplanes of its own, so this unusual event woke up and alerted a number of people.

    “Uncomfortable with the crowd, Consul Shepherd turned to the Icelandic police. “Would you mind … getting the crowd to stand back a bit, so that the soldiers can get off the destroyer?” he asked. “Certainly,” came the reply”

    Although the British action was to forestall any risk of a German invasion, none had been planned. There is only evidence of a German interest in seizing Iceland.

    • In my opinion the second world war was decided upon two large strategic mistakes on the part of the axis powers and one rather fumbling and brilliant (in retrospect) English attack (mentioned above).

      The first mistake stemmed out of Hitler being an infantryman causing him to be oblivious of the importance of naval warfare. If he had indeed taken Iceland in force he would have tied up the entire Home Fleet and the US Atlantic Fleet.
      This would have caused havoc among the Atlantic convoys to England since the submarines would have had shorter resupply routes and through long range reconnaisance flights and bomb flights out of Reykjavik.
      It would also effectively have stopped all convoys to the Soviet Union.
      To top it off, the Bismarck kerfuffle would most likely have ended up different with Bismarck running into port in Reykjavik instead of going for a French port and being sunk. That would have sealed the deal with keeping the Home Fleet tied up for years.

      Instead the Brits took the entire shebang bringing in a vital airport for the US airwings to use throughout the war. Liberated vital assets of the Home Fleet to convoy duty and the Bismarck ended up as a gallant but stupid sideshow in the war. And it also gave the ability for the Soviet PQ-convoys.
      It might have been a rather fumbled attack, but the result was most certainly so vital that it shortened the war considerable.

      Next mistake. Pearl Harbor, if the Japanese instead had followed the masterplan and attacked the Soviet Union out of Manchuria and sunk the Soviet Eastern Fleet the Soviet Union would have been in considerable trouble, far more than it ended up in. If one combines that with an Iceland flying the German Flag we would probably all have been speaking German right now.

        • You are of course correct.
          A very large portion of the planets population would not be alive today if the Axis powers had won the war.
          Either as direct casualties of the war, or died in concentration camps. I for one do not think that those would have stopped untill all undesired had been killed.
          That means that pretty much all of Africa, Asia and soforth had been murdered.
          Also, it would probably have been only a question of time before Japan and Germany had been at war with more casualties to follow.

      • In my opinion, the Axis powers would never have been able to fully take over the world. They made some mistakes, but even if perfectly planned, it would have been impossible for them to invade and occupy North America or South America. Their resources simply would have been spread too thin, and a north american campaign would be next to impossible with the technology they had at that time.

        That being said, it’s a darn good thing they never advanced further than they did. The horrors were already way too much even close to Germany. This was one of the few wars that was truly good vs. evil, at least in terms of the governments and leadership that represented the axis side of things.

        • I agree, they could not have taken North America.
          Not at least at once. But over time, who knows.

          I also agree that this was one of the few good vs evil wars. Come think about it, the only other one I can think of was when Vietnam whop-assed Cambodia and the Khmer Rouge.

          • The Axis powers would have had a difficult if not impossible time with
            North America and South America. In Oregon alone there was the coast militia that was made up of Loggers, Native Americans, Fishermen, etc. (some were all three.) who knew the woods and hills
            and were very well armed . Ditto the rest of the west coast including
            B.C. and Alaska. I won’t even go into the American west and Appalachia. Or the Canadian Maritime provences . The US. Navy had ships in the ways that would’ve required long range heavy bombers, something that neither Germany or Japan had. The Boeing B-29 was in development stages at the time of Pearl harbor, as was the Convair B-32 which was very similar, but not quite as complex as the ’29. The B-36 Peacemaker flew just prior to the end of the war and was capable of hitting both Germany and Japan South American invasions would have had similar problems.
            I would not want to be in an Amazon Jungle up against hostiles that knew the place and attendant disease problems also…
            However if Germany had gotten Nuclear weapons…..

        • There will shortly be a piece up on Gunung Egon that seems quite intent to erupt.
          It will be setting the background and giving facts from the recent report.
          The article will need some fleshing out in the form of events as they occur and links to webcams, seismos, gps-es and the usual works… So, let us crowd source for information!
          I will ad any links that is posted in this comment thread before publishing.

Comments are closed.