Icelandic volcanoes are all about long term projections. They do not do things in the short run, before any eruptions you will see either several intrusive episodes like at Eyjafjallajökull, or you will see signs of steadily mounting pressure prior to an eruption.
For some Icelandic volcanoes these two processes goes on for decades, or even several centuries. In a select few volcanoes this process takes from a few years to a few decades. In both cases the processes leaves tracks that we can follow through instrumentation.
For a handful of Icelandic volcanoes we have instrumental records that are beneficial for predicting their near future. These predictable volcanoes are Krafla, Bárdarbunga, Grimsvötn and Eyjafjallajökull. The special case where we do have a record but are still not able to make long term predictions is Hekla, which is a truism I will get back to later in this article.
Even though we lack long term data prior to eruptions for Askja and Katla we do think we can make accurate short term predictions ranging from days to weeks. And since both of these can cause massive problems this is a good thing.
So, without much of a further a due, here is comprehensive list of what these volcanoes are up to now and which ones are close to erupting. I am obviously leaving out the volcanoes that may in the future erupt that we lack good records of.
After a 50 year long deflationary period there are signs that magma has started to slowly accumulate again at root level. No large scale intrusions has so far occurred to the upper magma chamber, nor is there any signs on the GPS-station that magma pooling has reached a level even close to what it was prior to the last eruption. This is currently a volcano that is definitely several years to decades away before an eruption can occur.
Recent field measurements taken by the Icelandic Met Office and the University of Iceland gives at hand that two processes are occurring at the same time, there are both root filling of magma at depths of 15 kilometers and post eruptive deflation in the large magma chamber.
People tend to see the spattering of earthquakes at 0 to 7km as signs of an upcoming eruption, but these earthquakes are caused by subsidy of the caldera “plug” as it lowers and it is an ongoing process that started during the eruption itself. Over time the amount of M3+ earthquakes has steadily dropped and over the next few years Bárdarbunga will become ever more quiescent. It is though good to remember that Bárdarbunga has always been a very noisy volcano.
Current estimate is that Bárdarbunga is anything from 3 years to several decades away from the next eruption.
After the last large eruption in 2011 Grimsvötn went into a period of almost dormancy due to lack of system pressure as a significant amount of magma had been erupted. After a period of 3 years we started to see signs that system pressure was picking up, especially if we look at the Cumulative Seismic Moment (CSM) plot. According to this plot we are about 2 to 3 years away from the next Grimsvötn eruption.
GPS trajectories at Grimsvötn normally support this steady inflation model. During an eruption you get a marked deflation, that post-eruption slowly inflates again at a steady pace, before the process repeats itself. The GPS and CSM curves pretty much used to march on hand in hand.
But, sometimes even the most predictable volcanoes can do things that are a bit out of the ordinary. In Grimsvötns case this has to do with the volcano having 3 distinct magma chambers. Two of them are reaching into the caldera; these are the NW and NE magma chambers. The NE magma chamber has not erupted in instrumented history and it is believed to be both old and severely damaged after a large eruption during the Saksunarvatn Tephra eruptions.
If we instead look at the NW magma chamber we find a true volcanic monster. It is of equal size to Bárdarbunga, but the rate of eruption is several times larger, indicating a higher magma intrusion rate. It stands for all known observed eruptions from Grimsvötn, among them the 1996 eruption of Gjálp that was equal in size to Holuhráun and the 2011 eruption that was even larger. Half of all ash and lava produced in Iceland since 1783 is estimated to have originated from this magma chamber alone. This is what we know as the Grimsvötn volcano proper.
So, it should be safe to assume that this is the most likely spot to erupt in Iceland? No, not right now since the signs are not entirely there. Up until the eruption at Holuhráun there were almost no signs of increased pressure on the CSM and the GPS-trajectory flat-lined after only two years. And during the Holuhráun eruption this magma reservoir even showed signs of relaxing as intra-volcanic pressure caused by Bárdarbunga diminished.
Now there are probably some people saying that I have been talking about Grimsvötn being close to an eruption, and also that the CSM and GPS-trajectories are moving nicely in a direction pointing towards an upcoming eruption. And yes, they do and I did.
The problem is just the where part of where the eruption will occur. Let us start with the trusty old CSM plot that so far has not failed us when predicting Grimsvötn. It is a sum of released seismic moment in the form of earthquakes. The bulk of the earthquakes that has started to lift the curve of the plot have not happened inside the caldera or along the usual NW magma chamber radial fissures as they have previously.
Instead the earthquakes have occurred around the Southern magma chamber that resides outside of the caldera under what by some is seen as a twin volcano of Grimsvötn called Háabunga, or at two radial fissures associated with Háabunga. One of the fissures goes towards the Hámarínn Volcano and the other towards the Thordharhyrna Volcano.
So, what does the GPS say? Well, first we need to discuss the location of the GPS station. It is placed on a Núnatak on the southern caldera wall just south of the tip of the NW magma chamber and just north of the southern magma chamber. During a normal eruption it will move about 30cm north and west and drop about 25 centimeters. Then it will steadily move up and to the southeast roughly the same amount before the next eruption occurs. Keep this in mind, because now things will be turned around.
If we look at the GPS we see the regular pattern up until just prior to the Holuhráun eruption with steady up, south and east components. This stalled right prior to the eruption and even reversed a bit due to intra-volcanic pressure drop.
After the eruption we get a rather unprecedented pattern where it looks like two opposing forces are fighting for dominance. There is no sign of the steady nice movement that we have seen between every instrumentally recorded eruption. And around volcanoes new things are pretty much always bad things.
This leads us to the activity of the last two months. In the beginning I wrote it off as ice-buildup or a technical failure. But, as time passed on and as the GPS-movements surpassed any known ice-buildup recorded I got highly intrigued. And since there were no signs of instrumental failure I started to see the figures as possible correct. Also, both the GFUM and GRIM stations gave the same data.
I still waited to see if I could detect any movement from nearby stations that would support the GFUM/GRIM-station data in what is called a teleseism (distant effects from volcanic intrusion). These are normally delayed in time, and so it was here. The detectable signals are small but they are there if you remove seasonal variations and the behavior of the individual volcanoes they belong to (Bárdarbunga, Vonarskard and Hamarínn).
If we put all of these things together we see a minor intrusion at the northern end of the Háabunga magma chamber pushing GFUM/GRIM stations 40mm north, 30mm west and 90mm up in 2 months time. The high up-component is probably due to the station being between two inflating magma chambers, think of it like two inflating balloons pushing towards each other, anything in between will be lifted upwards at express speed compared to if there was only one balloon.
Before we start talking about Háabunga there is one thing I should point out, it is likely that the intra magma chamber pressure battle will cause the NW magma chamber to erupt again instead. This chamber is after all far more fractured due to regular eruptions and far more prone to failure than the chamber of Háabunga.
Háabunga is potentially bad news since we only know two things about it. The first thing we know is that the magma chamber is equal in size to the NW magma chamber that causes the Grimsvötn eruptions. The second thing we know is that this is the only magma chamber at the Grimsvötn/Háabunga massif that has not gone caldera. Mind you, I am not saying that it will go caldera, not at all.
We do not know how a Háabunga eruption would be, we do not know if it would be massively explosive, or if it would be benign and effusive. We do know that there have been radial eruptions towards Hamarínn and fissure swarm eruptions down towards Thordharhyrna, but there all of our knowledge ends.
We also do not know how large the signs would be before an eruption occurs, logically we know that since it has not erupted for at least 1 000 years it is likely to give off pretty noticeable earthquake swarms prior to erupting, regardless if it will erupt upwards or via a fissure since neither the top of the chamber nor the sides should be fractured.
Right now things are hanging up in the air, either we get a “traditional” eruption at Grimsvötn or we get an unknown eruption at Háabunga. Either way I think that Háabunga is here to stay as a potential eruptive place and that it is only a question of time before it will erupt.
The latest computer model I have made is giving a potential point of eruption at anything between 3 months and 3 years. I will follow this particular volcano even more diligently in the near future.
Eruptions in Hekla are caused by two things. Steady pulling apart of the Heklugjá fissure as Iceland drifts apart and steady state intrusion of magma from the MOHO. As the tension and pressure becomes too much for the overburden to contain the fissure opens up and we have an eruption.
The constant pulling apart of the thin hot crust on top of the fissure produces very little seismic signals between and during eruptions, and since the intruding magma barely contains more volume than the volume the separation creates there are few pressure increase related earthquakes.
Neither does Hekla produce large changes in the GPS since the fissure pulls apart at almost the same speed that the magma intrudes. The only thing that really moves is the top of Hekla that steadily drops between eruptions due to the weight of the edifice pushing the entire volcano downwards at a rapid pace through the thin hot crust above the fissure. Still there is evidence of inflation, but visible as East and West spread mated with no down-component.
Hekla is since 2011 the most instrumented volcano on the face of the planet, and this makes it possible to see things that were not possible to see during prior eruptions. Even the 2000 eruption occurred almost in the dark. But now things happens in the light of glowing instruments in the dark of the night.
Now we can see quite a few minute earthquakes occurring, and here is what is a real breakthrough. We can see a definite trend that follows the pattern of the other well known seismically quiet volcano on Iceland (Grimsvötn).
If we look at the new plot that the Icelandic Met Office has started to produce we see that there is a steady and marked trend towards more earthquakes at Hekla of a size larger than M0.8, and we can also see that the curve of increase is logarithmic (same as at Grimsvötn).
Since we do not have a complete pre-eruptive sequence for the 2000 eruption we can’t predict accurately when an eruption will occur, we can just say that we are now rapidly closing in on one.
Let us now play a bit with this curve and pull it out into the future, in 3 years time we will have more than one earthquake larger than M0.8 every week. And we do know that this did not happen prior to the 2000 eruption, so it is probable that an eruption will occur prior to that.
So, by using the traditional model for Grimsvötn upon Hekla and the known record during the last couple of years prior to the 2000 eruption and then applying all the new equipment it is possible for the first time ever to give a medium term prediction for Hekla. And that ends up with being anything between 1 hour and 3 years with a 70 percent confidence. I should point out that there is not a lack of instrumentation that gives the low confidence number, it is the lack of instrumented record.
As almost everyone has noticed Katla has been a bit festive lately. There have been repeated earthquake swarms of unusual magnitude sizes. Only problem here is that all earthquakes have been shallow indicating movement of hydrothermal water and not intrusions of magma. This could though over time change, but there are currently no signs of an eruption being close by, at least not more signs than before the seismic hubbub started.
The deeper earthquakes have all occurred at the Gódabunga magma intrusion. The magma here originates from a separate magma conduit outside of Katla and may constitute a formative volcano between Katla and Eyjafjallajökull. The area has for the last 25 years been the most seismically active region in Iceland and the level of activity is currently low for this volcanic feature.
The question about Katla is more why it is not erupting, because this is an unusually long interlude for this particular volcano. My personal opinion is that the repeated intrusions into Gódabunga has somehow siphoned off magma from the MOHO that otherwise should have gone into Katla, if I am correct this will probably have consequences for Katla in the long term.
Well, I did mention this volcano as one that we have a good record of and that we should be able to predict prior to an eruption. Only problem is that this particular volcano works on such a timeframe that we will most likely not have any use for that information in a couple of centuries. So if you wish for a volcano prediction for Krafla you will have to dig me up in a couple of centuries.
In the end not much and everything has changed in Iceland over the last year. Yes, it is still the extended system of Grimsvötn that is the most likely to erupt soon. But, it is likely that it will be an intriguing new part of the Massif that will be erupting.
At Hekla we have learned that we now can say that it is likely to erupt soon with a greater confidence, but we already knew that it was probably to erupt soon anyway. The really new part for Hekla is actually that we now with great confidence can say when it is not going to erupt, and that is brand spanking new.
And for those who see a lot of earthquakes at Katla and Bárdarbunga it is time to start looking closely at where the earthquakes occur, and what the reasons behind them are. Otherwise you will miss out on all the wonderful trends that are real out there and that are proven to be indicators of upcoming eruptions.
Some of you may have noticed that I omitted to mention the strainmeters located at Hekla. There is a reason for this. Lately there has been an almost religious belief about those spreading in various sites, a belief that I fully intend to debunk in my next article.