Signs and portents of Iceland – Revisited

Steam and gasses coming out of a vent on the summit of Hekla in 2005. Photograph by Borkur SIgurbjörnsson.

When I planned to write this article about the current states of Iceland I only wanted to write about Katla and Öraefajökull. But, as things turned out a third volcano got my attention.

In the end this article will be about how hard it can be for a layman to see what is important and not when overwhelmed with the plethora of information that is available for our beloved Icelandic volcanoes.


On the importance of importance

Crustal depth map of Iceland, Mohorovic discontuity. Image by Andrej Flis.

Not all volcanic signals are equal. And not all volcanoes are behaving the same prior to an eruption. Some volcanoes like Bárðarbunga are incredibly noisy between eruptions, prior to eruptions and during eruptions.

For Bárðarbunga it is a completely ordinary thing to throw a moment-magnitude earthquake of M5 with a follow up consisting of a heavy earthquake swarm, without it being a sign of an impending eruption. If the same thing happened at a more quiet volcano of the same size and capability, like for instance Grimsvötn, it would be time to run for the hills.

It’s the same with uplift and inflation, what would be big news for one volcano would just be a yawn for another volcano. If for instance Askja would be inflating at the rate and persistency of Iwo Jima it would also be time to run like a bat out of hell. At some volcanoes a few centimetres is a tell-tale sign, whereas at some volcanoes hundreds of meters are normal.

Some volcanoes are consistent stinky gas bags at all times, while others barely fart even during an eruption.

This is why it is so important to know your volcano well before trying to understand what it is doing. And, this is also why new data can be misleading as it comes to light.



Myrdalsjökull Glacier in 2005. Myrdalsjökull is the glacier covering Katla Volcano. Photograph by Chris 73.

Katla is quite likely the grand volcanic master of producing fake news, only rivalled by Yellowstone. Anything, including figures on ice cream consumption, will inevitably lead to war time headlines belting out that it is about to explode and destroy life as we know it.

Before we start with the latest news I would like to state the state of this volcano.

For as long as we have been able to instrumentally track and record this volcano it has suffered from intermittent intrusions visible on both seismometers and GPS-stations. Problem here is that it has not erupted for a long time, so we do not know that well how an upcoming eruption would look like, due to lack of anything to compare with.

All we know is that it is at times a rather noisy volcano, and that it can withstand a lot of big intrusions without erupting. A qualified guess, based on historic records from previous eruptions, seems to indicate that the last couple of days prior to an eruption would be quite memorable. So, expect to see something close to what you saw at Bárðarbunga prior to onset of eruption.

That means hundreds or thousands of earthquakes per hour with, some reaching up to M5 in strength.

Now let us turn to the latest news item. A group of scientists published a paper on measurements of free air CO2 released by Katla. The results are really intriguing since the values was quite literally sky high.

As far as we know Katla is ranking in at number 3 in the world in this regard. This is an anomaly in and of itself, since figures that high are normally associated with calcium-carbonatite volcanoes, or volcanoes that are erupting through heavy layers of carbonatite bedrock.

There is obviously no carbonatite lava in Katla, so we have to look for some other solution to explain where all that CO2 is coming from. My personal guess is that it is a function of fresh basalt percolating through a slab of oceanic crust partially consisting of calcite.

Calcite crystal on bedrock from Brazil. Photograph taken by Rob Lavinsky.

We all know that as you heat crushed calcite to produce cement you get an enormous release of CO2, cement production is the second largest source of manmade CO2 after consumption of fossil fuels. We also know that oceanic crust contains quite a bit of calcites.

What we know is that CO2 normally is an indicator gas pointing towards that magma has moved closer to the surface, so the world press jumped on the band wagon that his meant that Katla would explode hugely in the next couple of weeks. Problem is just that Katla has probably had a tremendous output of CO2 for millennia.

But here is the thing, we do know that the magma has risen closer to the surface lately, so a small part of that CO2 is probably related to an upcoming eruption. What the world press forgot is that even though that part is true, we are not actually seeing any signs that an eruption is close.

In Katla’s case we would see a tremendous amount of earthquakes for a few days coupled with rapid intra-caldera inflation. At least if it will erupt in the normal way through a caldera vent. If instead we get a rifting fissure eruption the noise level would reach unprecedented levels recorded in volcanology and there would be very funky GPS movements to the NNE of Katla proper.

But right now all we can say is that Katla will erupt one day. That is after all what volcanoes do. What we can’t do is accurately forecast when the volcano will erupt, since there are no current signs for an upcoming eruption. But, we will be able to forecast the eruption in time to give warnings to the local residents.



Havannadalshnúkur and Öraefajökull volcano. Photograph by Kristinsteff.

This easy to pronounce Icelandic volcano is currently my main cause for concern. In case anyone from the press is meandering about I should probably state that as far as we know today, the upcoming eruption is quite some ways away.

At the same time the current level of activity has never been seen at this volcano. We do though know from historic records that this volcano is quite noisy prior to an eruption, so what we are seeing is nothing at all compared to what will be visible on the seismometers when the volcano erupts.

After being dormant and almost entirely quiet since the last eruption in 1728 the volcano started to show signs of re-awakening in June 2017. To explain what we are seeing I need to fire off an analogy to put things into perspective.

What we are seeing is a heavy weight boxing champion waking up on a match day. The signals we have seen is the first initial fluttering of the eyelids. We have not yet even gotten to the first yawn prior to the morning cup of coffee.

In other words, do not expect The Rumble in the Jungle starting tomorrow.

Gravity anomaly map indicating where there may be magma reservoirs. All known Icelandic magma reservoirs show up as negative anomalies with the exception of Grimsvötn. To date there is no explanation why Grimsvötn is a positive anomaly. Image by Andrej Flis.

The reason for me being so interested in Öraefajökull is it’s eruptive history, as far as we know every single eruption after the last ice age glaciation has been a VEI-5 eruption. In 1362 it belted out a major VEI-5 eruption, and the one in 1728 was a minor VEI-5, and as far as we can see all other eruptions has been around the same power level.

During the two last eruptions the residents near the volcano was wiped out completely, making this particular volcano Iceland’s deadliest volcano after Grimsvötn. And today there are once again residents living near the volcano. This is why this is a volcano that science is not allowed to misjudge prior to an eruption.

That being said, I am confident that science will accurately forecast this volcano in time to perform a mass evacuation.

The thing that makes this volcano so dangerous is that it is old, and that it erupts far apart in time. This has led to the magma fractionalizing at a high degree leaving a large reservoir of rhyolite. And as fresh basaltic gas-rich magma intrudes into that reservoir it will in turn expand.

Gas rich rhyolite has a tendency to erupt violently at the best of times, but in this case there is a lot of glacial ice, and as that melt it means that there is ample amounts of water at hand to further increase the level of explosivity.

During an eruption air travel between Europe and the United States would become impossible, and the weather in the northern hemisphere would be affected.

Summary map of eartquakes at Öraefajökull showing how the amount earthquakes increases over time. Image by Icelandic Met Office.

What I find interesting is how relatively smooth and persistent the increase in seismic activity has been. In June 2017 we saw 13 earthquakes larger than M1.2, and last month we saw 25 earthquakes larger than M1.2. Compared to Bárðarbunga this is of course a pittance, but the trend is clear. Magma is at a steady rate entering the magma reservoir and it is showing quite nicely on the GPS-plots.

So far we have not seen any large earthquake swarms, and we know that prior to an eruption the swarms will be continuous and contain earthquakes large enough to topple houses. At least that is what we know from written historic records.

Without giving any firm timeframe I say that it is a safe bet that we will over time see earthquake swarms developing, that in turn will increase in size, duration and energy level. Between these we will see the background seismic level increase.

So far this is almost a picture perfect start for the ramp-up period prior to an eruption. And as time progress we will be able to produce better forecasts, because all we can say now is that this volcano will erupt again in a not too distant geological future. And with geological future I mean anything from a month to a century in this case, with a month being highly unlikely at the current state of affairs.

Now that we have discussed two volcanoes that we can forecast days, or even weeks in advance, it is time to go to another extreme. A volcano only possible to forecast an hour in advance, at best.



Tephra layers in southwest Iceland showing the various tephras of Hekla. Photograph by Dentren.

In some ways Hekla is the most boring volcano on the planet, since it is not giving away a lot of information before erupting. At the same time this just increases the mystique for the true volcanic connoisseur.

Between eruptions all we see is a slow and steady inflation without any spurts of activity, it is almost like watching paint dry on a wall. You know it will be finished sooner or later, and that the result will be pleasing, still it is quite boring to look at.

The same thing goes for seismic readings from the seismometers. Now and then you get a couple of earthquakes, but not more than 5 or so a month. None of them are big, none of them are really significant in any volcanic way.

And let’s not even get into measuring gases. Your Friday ale is more interesting than Hekla when it comes to producing gases between eruptions.

In regards of Hekla we can’t even complain about not having an instrumental record of how it behaves prior to an eruption, because we have two good ones. And those two are stunning examples of pretty much nothing.

Progress over time of Hekla earthquakes since the major upgrades of the Hekla seismic network started in 2011. Image by Icelandic Met Office.

The only thing we know from those two is that roughly 60 minutes prior to an eruption there will be a small spattering of minor earthquakes, and then the mountain pulls apart and the gates of hell opens up in vivid colours and fury.

In the end what makes Hekla such an enigma, and such an interesting volcano to watch, is that everything has significance. A small earthquake can at any time get a few friends and fire and fury unleashes a few minutes later. Any earthquake can truly be The One.

Today the monitoring network around Hekla is 10 000 times more sensitive than during the eruption in 2000. This means that we see quite a lot more earthquakes, both smaller and at greater depth and in far more detail.

This time around we should be able to pick up what the signs and portents prior to a Hekla eruption are. Hopefully and perhaps.

One of the guesstimates is that prior to onset of an eruption we should see a few deep earthquakes between 25-30 kilometres depth heralding influx of fresh magma at depth. According to the guesstimate these should be almost directly below Hekla proper. And from 2011 the network has been sufficiently sensitive to be able to pick up those deep earthquakes. And on Tuesday two of them appeared.

Hekla in 1893. Photograph from the British Library Archive released under public commons.

Let me be the first one to state that there is a bit of conjecture that an eruption at Hekla would be heralded by those deep earthquakes. But Eyjafjallajökull 2010, Grimsvötn 2011 and Bárðarbunga 2014 was indeed heralded by such earthquakes. At those volcanoes it took between a year to several weeks before the actual eruption occurred.

We do though know two things about Hekla, it is an open conduit system, and we also know that the fastest speed with which magma has ever risen in Iceland is 2 kilometres per 24 hours, so if we would accept those figures the fastest possible time to eruption counting from today would be in ten days.

I am obviously not stating that Hekla will erupt in ten days, it might be in ten minutes or in ten years. I was just guesstimating the fastest possible time that fresh magma at depth could cause an eruption.

Those two small deep ones has since been followed by another 3 minor earthquakes at depth between 1.5km and 10.7km. The most shallow earthquake was probably caused by the weight of the mountain causing downwards pressure, and the one 5km is probably near the magma reservoir. But the most recent one occurred at 10.7km and may in some respect be associated with an ongoing intrusion.

Hekla is absolutely infuriating. I suffer from an almost perverse pride in my ability to accurately forecast Icelandic volcanoes based on scientific theory and raw data. Bárðarbunga was easy, I had that lamped a year prior to the eruption, and accurately predicted when and how it would erupt days in advance. For Hekla I might as well don a robe and rub tea-leaves on my bald head and go about chanting. So, all I have written above about Hekla is conjecture at best, but still a scientific conjecture based on raw data and a tea-leaf toupee.



If we now look at Iceland as a whole and try to see which volcano will erupt next time we have two well known candidates, and only one of them is on the list above. Grimsvötn is the most likely, but currently it is a bit far off from erupting according to data, so Hekla would be the best candidate for the next eruption.

But after the likely culprits of Grimsvötn and Hekla the field is surprisingly open. Katla is not a bad bet for an eruption in the next decade, but it might hold out a bit longer.

Öraefajökull is the big unknown, currently it is ramping up nice and slowly and should at the current rate also erupt in the next decade. Problem is just that the current and ongoing intrusion of fresh magma might stall, and it could take a few decades more until it is ready.

The conclusion might be unsatisfying, but in the end all we can say is that deep into the future Grimsvötn will erupt a couple of times, and that a couple of other volcanoes will pop an eruption. Perhaps not so bad after all, there is fire and fury looming in the distance as volcanoes do what they do best, erupt.


168 thoughts on “Signs and portents of Iceland – Revisited

  1. There are two kinds of major basalt lava flow eruptions in Iceland. Its what I learned some times ago.

    There are the large fissure eruptions that erupts tremedous ammounts of basaltic lava quickly quickly. They form tall fountains and large channelized rivers of lava and Aa flows.
    Like Laki and Holuhraun and Eldgja.
    Output is huge 100 s to 1000 s of cubic meters a second

    Then there are the large slow monogenetic shield building eruptions, such eruptions can be huge volume, but erupted rather slowly. Its an Artic version of Puu Oo. Such eruptions can last for many 10 s of years produces lava tubes and extensive pahoehoe flow fields.
    Examples are Trölladyngja and Skjaldbreiður.
    Output is low 1 to 15 cubic meters a second

    • We do not know if the shield eruptions erupt that slowly. Might actually be much faster.

      Regarding the post, let’s say:
      -Hekla can erupt at any time. A few deep earthquakes a week before. Could be in 2018 or much later
      -Katla closer, but pro needs some months of strong swarms and maybe sharp inflation. Could be 2019 or later.
      -Oraefajokull needs a bug bunch of M5 quakes, probably many months to a few years down the road. Could be 2019 or during the 2020s.
      -Grimsvotn still some more time to go. Maybe erupts 2020 or 2021.

      • If the shields are predominantly pahoehoe with lava tube systems then they should have formed at rates of 10 m³/s or lower which is what has been observed at Kilauea for that kind of flows. And the overall rate of the eruption would be lower because it was probably not continuous lava flow emplacement but in episodes like in Pu’u’o’o or Mauna Ulu.

        • I don’t know much about those shields so I don’t know whether that is the case or not

          • Considering that one of them allegedly created a 100 km long flow and several have made flows far in excess of 50 km long, I would say there is a pretty good chance of at least part of these shields eruptions being a lot more voluminous than steady effusion. A flow rate of about 30 m3/s is required to get a lava flow 100 km long, basing off the June 27 flow in 2014. I think it is reasonable to assume most of these shields had eruptions that went for nearly or over 1 century, pu’u o’o has a total volume of about 11 km3 in 36 years and so a 50 km3 shield like theistareykjarbunga or skjaldbreidur would have had an active life of maybe 200-300 years at pu’u o’o eruption levels. This is completely plausible but it is even more likely that eruptions were episodic and occasional maybe much higher effusive episodes occurred. In August 2011 one of the flank breakouts on pu’u o’o probably had a flow rate as high as fissure 8 but it was only for a short time and it’s flow was fairly small. Also on kilauea, in 1998 another flank eruption on pu’u o’o had possibly an even higher effusion rate and included fountaining from several vents including the active lava tube source and these combined to produce a channelized a’a flow to the coastal plain within a few hours, almost as if in attempt to restart its episodic high fountaining pre-1986 behaviour. This surge was never observed in action because it happened at night so it is much less well known to the general public (it is what I thought would happen way back in April though) but it shows that even consistent static and well observed eruptions can have unpredictable high effusion rates.

          • High rates tend to give very long flows but a good insulation like for example a lava tube can also allow it to reach very far away, for example the Kazumura lava tube of the Aila’au eruption is 65.5 km long and the full length of the flow is little more than that as there would also be the lava delta of Kaloli point. Since it was a lava tube flow it should have formed at rates not very far from 10 m³/s which probably were sustained for a long period of time.

  2. Small gathering of earthquakes on Mauna Loa. 36 in one area since the 9/27 with one of them a 3.4, most from 2-6 km.

  3. Talking about large effusive eruptions, it called recently my attention that the 1907 eruption of Alayta (In the Afar Depression) was remarkably similar to this year’s Kilauea eruption. What actually initially called my attention is that Alayta has a group of roughly 47 pit craters on one of its rift zones and then I was also surprised to find very wide perched channels (some comparable in size to the fissure 8 lava channel) in many of the eruptions coming from the east flank fissure area. I do not think though Alayta and Kilauea are volcanoes that can be considered to be very similar, you can probably randomly choose a basaltic shield and it will likely have more in common with Kilauea. In my opinion Karthala in the Comores would be Kilauea’s twin although smaller but I think Alayta, not being a good analogue to Kilauea is still very interesting as a volcano and to be compared to other basaltic volcanoes.

    That is one of the two main lava channels that are atributted to the 1907 eruption, the other one looks quite similar. It has the same perched morphology as fissure 8 channel and it also shows a bright area of pahoehoe overflows surrounding the channel, which means it was also sustained in time and had a fluctuating flow rate. The surface of the channel also shows crustal blocks which is how fissure 8 lava channel more or less looks now. These perched channels are more or less common when they are a few tens of meters wide, but channels of these characteristics rarely get to a few hundred meters wide, the two channels of the 1907 eruption reach maximum widths of ~200 m still far from the more than 400 m that the fissure 8 channel reached at some points but older vents close to the 1907 fissure produced perched flows even wider than the one of fissure 8 while Kilauea doesn’t show any other lava channel of these proportions. This channels only seem to form when the effusion rate is very high (~100 m³/s for fissure 8), when the lavas are very fluid and not very far away from the vents. The longer channel of the 1907 eruption was able to maintain this morphology for the first 9 km and then turned into the more blocky crust full-aa type. It also helps if all the effusion is concentrated into one vent as happens in many older erutions of Alayta.

    The area of the flow field of the 1907 eruption is of about 60 km² that to compare to other historical Afar eruptions is more than twice the extension of the 2011 Nabro flow and about five times the extension of the 2008 Alu-Dallafilla eruption. Holuhraun was 84 km² and Leilani 35 km² (but of course a lot of lava went into the lava delta). After doing some measurements with a point cloud of the fissure 8 channel I got thicknesses of 10-11 m for an aa lobe, of 16 m for the main channel when the lava was still liquid and 5-19 m for the areas to the sides of the main channel, the ones that got covered in pahoehoe from overflows. I decided to use a value of 12 m thick for the average of the 1907 flows based on its similarities to this year’s eruption and from that estimated 0.71 km³ of total volume, if this is right it would make it slightly smaller than the Kilauea eruption but remarkably close.

    • The structure of Alayta is also interesting.

      This shows the distribution of fissures (in red) and pit craters (the yellow dots). It is interesting that the pit craters more or less coincide with the point where the northern Alayta rift zone diverges into two, basically because two fault systems with two different orientations intersect in there. There are roughly 47 pit craters compared to the 15 Kilauea’s ERZ has, but the pit craters of Alayta are smaller than those of the ERZ. In terms of surface extension (cause I don’t have values for the depth of Alayta’s craters), the two biggest ones of the ERZ (Napau and East Makaopuhi) have around 1 km² while the large craters of Alayta average 0.1 km². It is still interesting to compare and I think supports the concept I had of the pit craters in Kilauea forming at intersections between intrusion pathways in an horizontal conduit.

      This is how I would schematize the fissure and pit crater distribution:

      I had to create some names cause there is no research done on Alayta as far as I know. Actually most of the recent activity is concentrated in the East Flank Fissure System and the most recent vents inside it are in its northwestern part. The NW rift zone has been almost dead for some time while the NE and S rift zones have produced large flows of recent appearance. I find it hard to believe that the East Flank Fissure System and Alayta share plumbing systems and maybe would be two different volcanoes. Also, Alayta has a very small shallow reservoir and seems to contain evolved magmas judging from the last summit eruption, there are no large calderas in the area either so this is a major difference with Kilauea and eruptions like Holuhraun and that is that the voluminous eruptions on the east flank are not likely feeded by any magma reservoir of Alayta or other neighbouring volcanoes. 1907 was probably a larger than average eruption for the area, including both Alayta and the east flank but there have also been other voluminous eruptions that are recent-looking, it would be interesting to know how often eruptions of this size happen there, but the only way to answer these would be very rough guessing.

      • Because this area is a desert it can make the flows look way younger, the brown surface is likely thousands of years old. However the black flows are all definitely Holocene. This area is a part of the world that really needs to be studied in detail, the world knows about erta ale and its lava lake but that us about it.

        Speaking of erta ale, is it still erupting? At the rate it was going it should have gotten to a pretty decent volume of lava by now.

        • “This area is a part of the world that really needs to be studied in detail”

          The hazard of developing rapid lead poisoning acts to curtail much of the research.

        • As of july it was still erupting.

          Interesting that pahoehoe gets covered in sand much quicker than aa, there are some flows of the east flank which look completely untouched in their lower half and as you approach the vents where there is more pahoehoe the surface looks much older and developed large sand pots. The 1907 flow might be getting its first very small sand pots in a pahoehoe area right next to the vents.

  4. Thx Albert What I was thinking is that the flow dynamics of conduit and the well may may be similar. I was thinking more of volcano’s like Askja, Vesuivius, that have a high content of gas and steam and a large eruption with a strong flow. My thoughts were once the flow tapers the remaining magma column would collapse. I didn’t look at how liquid flows in a pipe I am assuming that a similar effect would happen. The flow would tend to the towards the centre. Perhaps a layer of film too?

  5. Ok My comment plan for the press conference, not good. How rash. Nope I Have a better idea

    I just love Irpsit’s dream not that I would ever want that to happen. Nope we weave a tale that the asteroid is going to hit Iceland and destroy it. There’s a very good chance of rifts occurring after the impact and one of the biggest worries is the Thames Valley . Yes that’s right geophysicists have determined that a weak seam in the earth’s crust runs from Iceland to The Thames valley and People should be prepared to evacuate at a moments notice.
    Off course we would need cooperation with a few people. A astro physicist or two, Experts on geology, Anglican minister. You get the idea. They don’t need to be real. The Brit journalists won’t have a clue. Then we announce to the world that it is ending. Half an hour later we print a retraction.
    ‘ April fools’
    They do it every year here to us, fair game , hoax the hoaxers

    Ps keep the Shaleighlas just in case

  6. Erta Ale flank eruption is still going.
    Tube feed and feeds an expanding pahoehoe flow field into the deserts east of Erta Ale.
    This been going on for quite some time now.
    Slow and steady like a Puu Oo flow Episode.
    Check thermal webcams of Erta Ale here: the hot expanding pahoehoe breakouts is still active.The tube system is well insulated and Dont emitt heat.

    • If this flow has been going on for about 2 years at a flow rate of about 5 m3/s then it must have a volume at least getting towards at least 0.1-0.3 km3, I think there is a pretty substantial shield at the vent too, if/when the tube eventually breaks down then shield building might really take off.
      This eruption also took place in what looks like another caldera which is separate to erta ale proper, and the original lava lake seems to be recovered from when it drained during the start of the flank eruption so it is something to consider whether the flank vent is actually part of the same system.

      • The flank vent caused collapse of the summit of Erta Ale so the two should be connected if there is a lava lake then there is probably equilibrium in the system right now, the rates were higher when the collapse took place I imagine. This could be compared to the situation since 2008 in Kilauea with Pu’u’o’o and Halema’uma’u erupting at the same time.

    • Unrelated to the quake, I expect. It is far from the quake zone and this volcano erupts very often.

      • 2016 was its last outing. Of course, the press are linking it to the earthquake despite the considerable distance separating the two. The photos show the plume looking quite ashy.

        • Strong quakes could provide the slight edge to tip volcanic systems already on the brink of a new eruption.

          These things behave in a chaotic manner. One little push from a quake may just do its job.

          • They can, but it’s usually local earthquakes that have the potential to upset magma chambers. Soputan is over 500 km away from the epicentre of the 7.5 M quake so I’d put this one down to coincidence, especially due to Soputan’s regular eruption profile.

          • Sure, and that has been seen in Chile. But the volcano should be close to the quake! And this one wasn’t.

    • I’ve released your other comment, hopefully, this won’t happen again now.

  7. What’s the news on the Mount Soputan eruption in Sulawesi? The BBC reported it but I haven’t seen much else.

Comments are closed.