The Plume of Ballareldar?

Stunning photograph of the eruption. Photograph by Haussman Visuals.

To me the part of a volcano that is visibly erupting is the least exciting partPerhaps a better way of stating it is, that it is only the effect of the cause. This is obviously not true to most people on the planet, so I think I owe everyone an explanation.

And that explanation is especially important since we need to look deep into the volcano, to understand its future.  

Like most people I can obviously spend hours looking at lava bombs being hurled, and lava slowly filling valleys. But, getting to know the hidden innards of a volcano, and understanding their functions, is making the experience even better. 

So, let us take a journey through the volcanic features of the Ballareldar from the bottom up. During this journey I will try to impart the wonders I see, and why this cute little tourist eruption is one of the most scientifically important eruptions ever witnessed. 

To do this we must employ many tools of the trade, petrochemistry, geophysics, chemistry and garden average physics, to be able to look below the ground we walk on. As I go on, I will try to tell you what we know, what we can assume, and what are open questions to science. 


In the beginning there was the mantle 

The Geology department at the University of Iceland did wonders in the opening stages of the eruption taking samples and analysing them at breakneck speed. I wish we had later data at hand, but I think that is saved for some juicy future articles by the scientists in question. Which is fair enough, and they have let a few tidbits out that is highly intriguing. 

What we do have is still enough to make me feel like a kid visiting his first candy store. Because things are sufficiently “out there” to make my eyebrows lift quite substantially. 

Regular lava in Reykjanes is indicative of Mid Oceanic Rift Basalt (MORB) origin, coming up from the Mohorovic discontinuity between the crust and the mantle, it is normally partially evolved (fractionated)and has ample amounts of inclusions indicating that it has resided in the crust for a while. 

Normally you will see a medium amount of sulphur in the lava, and it will be fairly cool compared to the plume derived lava coming out nearer to the Icelandic mantle plume, and the plume derived lava is among the most sulphur rich lavas on the planet. 

Image by the Geology Department of the University of Iceland.

If we look at the current magma being ejected as lava, we find that it is, first of all, unusually hot. The confirmed temperature is 1190 ºC from the first observations, but I have seen later unconfirmefigures up towards the 1220 ºC. 

Higher temperatures than 1190 ºC is by far not impossible, remember that the official temperature was taken early on when the magma had been cooled and partially quenched as it passed between the cold sides of the 7km long and 15 km high dyke leading from the deep feeder conduit near Keilir, all the way to the surface. 

As the surrounding rock is heated by the passing hot magma, over time the cooling effect will diminish and the temperature of the lava will go up a bit. 

As far as I know the previous temperature record holder in Iceland was the Holuhraun III eruption at 1180 ºC. Here we had an origin that definitely was from a well-formed mantle plume, yes it had partially resided inside inside of Bárdarbunga and had travelled for a long stretch across a very long dyke. 

But a big part of it was fresh material from inside the mantle that was newly arrived. On top of that the flow rate in the dyke was large enough to heat it very well indeed, decreasing the cooling effect considerably. 

The telltale low TiO2 and high MgO. Image by the Geology Department of the University of Iceland.

The temperature of the lavas erupted during the Ballareldar is high enough to be seemingly congruent with plume origin.  

If we look at the regular lavas erupted at Reykjanes through the eyes of groundmass glass, we see that it usually contains about 250ppm of Sulphur, but the current lava is erupting an average of 1140ppm of sulphur. 

And we do know that the Icelandic plume produces record breaking amounts of sulphur as the associated volcanoes erupts. 

Here it is easy to think that what we are seeing is a tendril of magma that has squeezed itself merrily along the underside of Iceland until it arrived below Reykjanes during the last 800 years. Looking at the evidence so far, it is not a bad idea. 

But we need additional data to prove or disprove our little plume origin hypothesis. This is the point where petrochemistry shines. 

Rare Earth Minerals chart. Even the lantanides are unusually absent. The Geology Department of the University of Iceland.

The first we see is that the lava is rich with olivine, a crystal that is called Peridot and Chrysolite when used as a gemstone. It forms in the upper mantle, so now we know that at least the magma is from the upper mantle. 

Olivine comes in three distinct flavours, the magnesium flavour called foersterite (peridot) that can be green or transparent. It can only form above 400km depth, below that you get wadsleyite. 

The other common one is the reddish-brown fayalite (chrysolite) that contains iron, this forms at lower pressures than forsterite, so as such it is not pointing towards deep mantle origin. 

I will just briefly mention the third flavour, the whacky Manganese olivine named tephroite. From a volcanologic standpoint, it is the least understood of the 3. It can also have any colour visible to man, since it has a propensity to make love to pretty much any other metal. It is the penultimate slut in geology, making it into a darn good precursor when looking for mineralisations to mine. 

The Ballareldar lavas are rich in magnesium olivine (forsterite), this means that the origin of the magma is somewhere between 15-400 kilometres down. We also know that many Icelandic lavas are forsteritic, so it seems like we have once more proved a plume provenance. 

Now we need to compare the Ballareldar eruption (2021-) and the Holuhraun III eruption (2014-2015. The first thing that we see is that Holuhraun III has less olivine (forsterite) than Ballareldar has. 

If we look at the weight percentage of TiO2 at Holuhraun III we find that it is at 1.75 to 1.9%, whereas at Ballareldar we see a figure of 0.9%. On the other hand, we see weight percentages of MgO at around 6.7% at Holuhraun III versus 8.8-9% in the Ballareldar samples. 

Did we just find a spanner crashing into the spokes of the wheel of our hypothesis? Can we save our our pet theory? 

Yes, sadly our pet theory dies here in the warm embrace of TiO2, this is due to us knowing that the distance from the Icelandic plume center does not indicate decreasing TiO2, or vice versa. Plume derived forsteritic basalt does not drop in TiO2 with half. Bummer! 

At best we have a partial influence of the Icelandic plume, but sufficiently small to not explain the sulphur and the temperature as such. 

Here one could come up with the crutch-theory that it is another unknown plume at work. That is amply gunned down by geophysics, since we know from tomography mapping of the mantle, using measured differences in the travel speed of sound indicating temperature variations in the mantle. In simpler terms, we have a fairly good map of where there are plumes, or not, in the mantle. 

There is obviously no special plume under Reykjanes. At this point we will have to wait for new data from young strapping Ph.D. students. 


New data 

The gassy belly of the beast. Image by the Geology Department of the University of Iceland.

This is written a couple of days later as an addendum. I had already edited in the article when I found new data from the geology department at the University of Iceland. Problem is that the new data made mince-meat of what I had written above. 

My first instinct was to do a complete rewrite of the article, so that it would no look like I used the southern end of northbound donkey as a brain. Instead, I am leaving out the first part as it is, as an example of how new data is driving scientific discovery and creates the need for new models and hypothesis-formation. 

I love the smell of fresh science in the morning, well that and coffee. So, without further ado we will boldly go where no person has gone before. 

Let us begin with what is the same. The sulphur content is same at the high levels, and the release of SO2 is keeping steady at 2000 to 3000 tons per day. The variations closely follow eruption flow rates, so we can safely say that it will not increase nor decrease over time in any significant manner. 

Several people have asked me lately about the noticeable increase in “smoke and gas” from the vents. And yes, there has been an increase in the visible gas volumes at the volcano. Problem is that there is no increase in release of CO2 or SO2 from the volcano, and this is to be expected since the lava flow rates are constant while the Sulphur content has been consistently high. 

So, why then are we seeing more gas? There are two reasons for this. The first is that it is likely that water vapour has increased due to the magma moving through a number of aquifers, and that a few of those contain super-critical fluids. 

I have however not seen any data on water content, so this is speculative. The second reason is simple: from an actively erupting vent you have sufficient thermal uplift to chuck the gas straight up and out of the way as a visual hindrance. 

That is why we see more visible gas from dying colder vents; they do not have the energy for effective thermal convection. 

In short, the gas increase is mainly more a question of altitude than attitude. 

Petrochemical differences over time. Image by the Geology Department of the University of Iceland.

Now, let us talk about the differences. MgO has increased from the previously high number of 8.8-9 percent, now it is 9.7-10 percent. This means that there is more forsterite in the mixture. This in turn points towards greater depth. 

Now, let us turn to the TiO2, it has increased from the low number of 0.9% to 1.5%. These two increases in TiO2 and MgO indicates a deeper origin. 

This indicates that the original magma most likely was of Icelandic Plume origin and that the plume head is slightly wider than previously believed. It also points towards some process depleting the magma during its long and slow movement towards Reykjanes from the plume core under Kistufell. 

One solution that is likely, is that TiO2 due to it’s higher melting point trends towards attaching itself to the bottom of the crust in a process called underplating, whereas the MgO does not. 

Now, here we arrive at a monster of a question. Was the eruption caused by arriving deeper material that first pushed up the depleted magma under the eruption site? Or, has the eruption depleted the supply of depleted magma and new deeper material is going up to fill the gap? 

If it is the latter, we are most likely seeing a smaller version of the process that created the Icelandic plume to begin with, eruptions causing a void creating lowered pressure increasing the melt process at depth.  

At the Icelandic plume this process has been running for 14.4 million years now, so it has burrowed itself deep and become a true monster among plumes. Whereas Ballareldar is too small in the greater scheme of things, and it will putter out when the eruption dies out. 

I should here point out that we do not know which one of the two options given above is true, I lean towards thel atter idea of burrowing. But, as per usual, until a strapping young Ph.D. Student has done the heavy lifting and done a garnet study we will not know for sure. 


Final words 

What I would like to see is a study of garnets in lava. Various garnets form at different depth in the mantle, so have a garnet study would be helpful to constrain further the depth of the formative melt. Want to get a doctorate in petrochemical volcanology..? Go garnets, go! 

I had initially planned to write about the dyke, and the future for the Ballareldar. I had also planned to write about the name BallareldarThat will though have to wait for part two of the article since I got rolling with the petrochemical part of life. 

So, in part two we will leave the mantle behind and become crusty indeed. 



characterization_of_the_1st_and_2nd_day_of_volcanic_products_from_geldingadalahraun_2021.pdf ( 

Microsoft Word – trace_isotope_report_v1r2.docx ( 

MS Template ( 

640 thoughts on “The Plume of Ballareldar?

  1. Remember that crack near Suðri that was emitting lots of steam during the very early days of the eruption? One tourist posted video of walking right up to it and watching it puff?

    Well, it’s emitting steam again, or vog perhaps, occasionally visible on ruv1. Watch from 02:13 for about a minute to see one instance.

      • Norðri still puffing a bit an hour later, but no discernible trend in activity.

        I’ve found something odd though when scanning back through the various cam youtube feeds: on the mbl cam, at the very least, it seems that night never fell! There was sunset, then twilight, then sunrise, with the darkest point at about 02:00. Even more strangely, the ruv cam shows the expected pitch black sky at that same hour. Suggests the mbl cam night mode is doing some serious light amp or something.

        Meanwhile, on ruv, something apparently ate a “slot” through the side of the fissure 5 cone pointing southwestish. In the center of the slot is what appears to be a ramp leading underground. Each time there’s a major episode of fountaining, a brief lava flow goes down that ramp.

        So, does fissure 5 now, because it’s become so busy, have an underground parking garage? Where is all that lava going? Because it doesn’t seem to be filling that slot up and (re)building a rampart wall there.

        The lava field northwest of fissure 5 is still emitting some vog, visible on ruv1. Is this a sign of continued activity feeding that area beneath its crust? What would be the source for that, perhaps fissure 3? There’s no visible activity from 3’s cone, however. On the other hand that spot appears to be uphill from fissure 5, and 5’s lava seems to be going to only two places, Meradalir and down that new parking ramp which points toward fissure 1 not 3.

        • Its a spatter flow, where the cone material is semisolid it can flow away but its really viscous so it moves very slow, its basically a basaltic lava dome flow/coulee, really gives a contrast to the fresh lava that is as fluid as olive oil in these sort of really hot eruptions.

          These spatter flows happen a lot at high fountaining cones, Pu’u O’o had lots of them at one point from when the tall fountains were directed over its cinder cone and flooded it wit hlava, it cooled but still flowed down the cone turning into massive rootless blocky lava flows. I also saw a video of one that formed on a cone at Piton de la Fournaise.

        • Not really strange. This time of year in Iceland the sun isn’t setting completely, it goes below the horizon, but not very far. So you don’t get a complete blackness of night, there is still a tiny bit of light illuminating the sky and that’s what the camera is picking up, depending on it’s sensitivity and settings. The actual midnight on Reykjanes is around 1:30 so that should be the darkest hour. Then, in a few weeks, it’s going to be light all night.

        • Thanks Chad, that’s what caused my southwest slow flank collapse on 1st May leading to this phenomenon. I’d not known this was a thing.

  2. Heavy vog production from a small region near fissure 3 and between it and 5. New vent opening?

  3. I think this eruption is going to behave a bit like Pu’u O’o did, now that a single large vent is open this spot will probably be the easiest option to erupt from. This geyser-liek fountaining might persist for a long time, or become more obviously episodic with longer intervals and larger phases of high rate effusion. The effusion rate averaged out is still somewhere around 7 m3/s, but because each fountain erupts after a few minutes of quiet it is effectively erupting 3 min of lava in a few seconds or so, the effective effusion rate of the actual fountains is soemthing in the hundreds of m3/s range for a brief time. That woudl probably serve to erode the vent pretty will, its likely very open now with the dense lava above it serving as the obstruction instead of an actual narrowing of the conduit or fluctuating gas levels, its basically a lava geyser in the most direct sense possible.

    Other vents opening now will probably require the cone to grow a lot bigger and the vent elevated a long way so that pressure favors a new fissure instead of a fountain. This is of course exactly what happened at Pu’u O’o, though it took at least 5 tries before one of these flank vents actually succeded in terminating the fountaining stage there.

    • Do you thing North or south is more likely then?

      The last to the north of the current vents looks to be higher, whereas south is down the valley somewhat, but that is also closer to the end of the dyke. So pressure would need to build up to fill in the dyke in such a way it is not released through the current vent.

      I’ve haven’t seen much of the valley where the paths up to the eruption were for a while, but this seems the most likely area to me. (Unless it goes sideways and that smoking patch to the side everyone was discussing yesterday is a precursor).

      Also, while we are looking at Geldingadalur, is anyone analysing the earthquake swarm near the big lake? Is it me or (despite no movement on GPS) does the pattern start to look similar to what happened here before the eruption?


      • I think the elevation of the vent now is probably why the other vents stopped, this one is lower, but now its also bigger so even if it gets to the same height this one will still be preferred now I think. Probbaly the actual vent will need to be at least 100 meters above where it is now, the cone might be twice that tall in total by that point, and it will take years at this rate. That long in the future the other vents will be cold, a new dike is equally likely in either direction, though the south is lower down at present so maybe that way. Really though its hard to say how this place will look in a year, or 3 years, but its probably going to look a lot different.

        There will need to be way more quakes at Kleifavatn to indicate an eruption there, Krysuvik behaves a lot like Krafla when it erupts, massive initial intrusion, rift filling with small eruptions, then big lava flood eruptions when the rift has no more space for magma to go. That initial intrusion will probably go under Reykjavik too, its goign to be quite obvious… 🙂
        Having a big basaltic fissure eruption go through a lake liek that would be a big deal though, the lake isnt going to last long but will put up a mighty fight, something like Tarawera I expect, or Veidivotn…

  4. On a separate subject I believe one of the “forecasts” in the VC head to head for Grimsvotn to erupt was 21 May 21. Is any review of that head to head planned now we are nearly there, and the cumulative pressure is still “only” at 4.5 when it needs to get to 5 (we think)?

    It seems unlikely Grimsvotn will erupt in the next 6 months based on the graph trend (que ashy clouds spurting out of the ice in the next few hours with that statement down ;o) ) so an updated view from the experts on what is going on under the ice would be nice.

    • There is a law in Iceland that says two volcanoes cannot erupt simultaneously.. So Grimsvotn has taken a year-long holiday. That prediction will be revisited at some point. I am happy though with the current eruption!

      • Ok, 21/05/22 noted as the date to watch. The vote for this eruption was 6 months to a year, no? Perfect timing to keep to the single eruption rule 😁

  5. RUV1 08:27 fissure 4 is also showing some vog production, caught during one of those pan&scans that camera does.

  6. The burning area is still burning, the smoke is visibly browner today than yesterday, but the area doesn’t seem to have expanded much since yesterday.

  7. Think there are weak fumaroles on the hill facing the webcam:

Comments are closed.