Geological time is interesting, because even the geological now is decades long, if not centuries. The current geological episode at Reykjanes didn’t start with Fagrafjall (it is no longer a valley, so let us drop the “dal” out of Fagradalsfjall shall we).
Nor did it start with the large and very noisy intrusion over at Thorbjörn a couple of years ago. No, it started over at a relatively unknown diminutive volcano called Hromundartindur in 2007, where a minor root-filling occurred accompanied by some 80 000 small earthquakes.
Another minor one occurred at Krysuvik at little later, but everything was still small and benign. Over the years other small and noisy intrusions occurred all over Reykjanes Ridge and Reykjanes Peninsula. None of which caused an eruption.
In retrospect it was a sign that the pent-up strain in Reykjanes was reaching the breaking point, and that it needed to stretch out and moult like a butterfly from its cocoon.
A lot of time was spent pondering on how much pent-up strain energy was down there in Reykjanes, because earthquakes was seen as the big problem. The numbers turned out to be quite mindboggling if one looked at it from a layman’s perspective, or from the perspective of the “oh, my god” hysterati*.
Because there are 3 M7+ earthquakes worth of energy down there, ready to be released in the form of earthquakes, it is incredibly important to understand that we are talking about energy that will be released in many smaller earthquakes over time.
Why 3 big earthquakes, is the first question? Well, because there are 3 zones that are separate, and each of them contain roughly the same energy. The first one is running from roughly East of Lake Kleifarvatn via Brennisteinsfjöll and Hengill, up to Hromundartindur. The second one from Krysuvik until the tip of Reykjanes Peninsula, and the third from the tip down the Reykjanes Ridge down the MAR.
The next part we need to remember is that the Reykjanes part of the MAR is filled with perpendicular faulting dividing the larger structure into many more manageable parts, so forget that an M7+ earthquake would crack Iceland in one go.
So, instead we will get many earthquakes ranging up to M5+ size, and that is fairly manageable, and something that we have seen during the Fagrafjall eruption.
Hysterati = people who professionally operate a YouTube channel for profit and uses clickbait pictures and title their dubious productions in a hyperbolic manner leading their viewers to believe that the world is about to end. If you are one of those looking for inspiration for a video, please feel free to be offended at this juncture, it was intended. If you instead make videos intended to educate people, no insult was intended your way.
The hen and the egg
Now, why did I just go on about earthquakes? Normally you will not have a volcano without earthquakes in one form or another. It is therefore prudent to talk about earthquakes and volcanoes a bit more, but this is a very large subject so I will stick to oceanic rift volcanoes in the particular form of Reykjanes for now. Do not overinterpret this to be valid even for every other oceanic rift volcano.
Let us use Fagrafjall as an example. It was a known rift that was perpendicular to the MAR as it transitions through Reykjanes, it was estimated that the total pent-up strain was equal to an M6+ at the rifting fault, something that turned out to be pretty much spot on as the energy was released one smaller earthquake at a time as the dyke formed.
But here comes the hen and the egg question. Is it intruding magma causing the rock to fracture and strain to be released, or is it the strain being released, and the fault pulled apart that suck in magma?
We know from the root-filling events, and other deep origin earthquakes that magma was indeed pushing upwards at the beginning, but in the beginning, this did not cause large intracrustal earthquake swarms. Instead, it slowly helped to add to the strain in the rift-system.
It also meant that there was magma available at many places under the Reykjanes peninsula. What then formed the magma? Well, that is easier to answer, decompression melt.
I think I am going about this in the wrong order, both spatially and chronologically. To answer why Reykjanes is a chicken & egg sandwich, I need to start from the beginning, if there is even such a thing.
The geological second
In Reykjanes 800 years ago, most of the strain was gone, the available magma had been spent, and life was about to turn dull for the volcano watchers of yon olden days. The eruptive phase of the 800-to-1000-year long Reykjanes rift cycle was over.
For almost 800 years the strain slowly picked up, but there was very little tectonic movement of the plates on either side of the MAR of Reykjanes, the plates were fairly locked in place. In the end of this period the strain was starting to be to large, and earthquakes and small swarms started as the plates started to move apart and against each other, the lock was coming apart.
As the plates moved apart the crust thinned minutely, just enough for the pressure to be lowered at the top of the mantle, and the mantle material started to transition into magma a little bit at a time.
Magma has a lower density compared to mantle material, and thusly requires a little bit of more space, so the magma started to exert force upon the crust above, causing small earthquakes, this is what we call a root-filling event.
Why upwards? Well, the magma is lighter than the mantle material, so it is trying to swim for the surface.
The root of the volcano at the MOHO level started in turn to put more force on the poor rift, and earthquakes started to happen above, creating voids filled with nothing, and nature hates nothing more than anything else, so magma was sucked up to fill in those voids. This of course lowered the pressure at the mantle creating more decompression melt.
The magma that was sucked up in turn caused increased strain above and towards the sides, so new earthquakes created new voids, sucking in more magma, more decompression melt… until Fagrafjall become a runaway train running over both chickens and eggs, the valley disappeared, and I renamed Fagradalsfjall into Fagrafjall.
At other places the going was tougher, and the overburden turned out to be to sturdy for the combination of suction/earthquake/decompression melt to push through. Yes, dykes formed, but they did not dilate enough to give birth to an eruption.
Fagrafjall turned out to be rather unusual in one aspect, it erupted in one go from onset of the earthquake swarm, via intrusion and dyke formation, dilation and to the squirty phase. Normally it takes two or more intrusions to break through.
The reason for this was that the overburden was rapidly weakened further by the pent-up strain, in other words, the strain was greater than the cohesion of the bedrock on top. At Torbjörn the situation was different, at 5km depth the bedrock turned out to be to solid and the intrusion sputtered and failed.
Eldvórp, Lagafell & Thorbjörn
The intrusion at Thorbjörn was large, at least counted as amount of uplift on a spreading rift goes, more than 100mm of uplift in the original intrusion.
During the time of the Fagrafjall eruption, and from then onwards, seismic activity continued at Thorbjörn, so it was clear that even though the intrusion was over, there was still both strain and pressure accumulating in the system.
So, I was not surprised when activity kicked off again a couple of weeks ago as a new intrusion started at Eldvórp, the next transverse fault over to the west. I was not even a tad surprised as Thorbjörn answered with an intrusive phase of its own just a day later. We had seen these intrusions pinball around before after all.
What was a surprising though was that the intrusions did not form dykes along their respective transverse faults that run perpendicular to the MAR, instead earthquakes propagated towards the middle at Lagafell.
This means that the new Dyke is following along the MAR itself, whatever now the consequences of that might turn out to be. I do though not believe that a MAR being magmatically lubed up by a dilating dyke is a good thing for Grindavik.
The runaway train
A decade ago, I was really bored one day, so I compiled all the data that I could find for the Reykjanes area bedrock and used finite element method (gobbledygook for heavy math) to try to find out what was the average* point when you could assume that the buoyancy of the magma would overcome the overburden at a rift fault.
It would have made for a fine paper, but alas, I am to lazy to be a good publishing scientist. Instead, I used the average number that I got, 2.7 kilometres, in passing mention for years, until I used it when forecasting the Fagrafjall eruption, and was resoundingly laughed at for it, until Fagrafjall proved the math I had been too lazy to publish, by becoming a runaway train at that depth.
Why am I now mentioning this? Simple, because of this earthquake:
|10:48:50||63.868||-22.443||2.8 km||M2.7||99.0||3.3 km N of Grindavík|
There is also something else that happens at this average depth, and that is that it is where nucleation of volcanic volatiles will start to happen at an increased rate. In other words, this is where volatiles will start to free themselves from their magma prison.
Obviously, I factored that in when I did my heavy math lifting, and it biased the average number quite a bit and skewed the results towards 2.7km. Take this titbit as you wish.
Average = I bolded and italicked this in the meagre hope that itinerant readers will really take note that this was the average value across all of the spots at Reykjanes that I could find relevant data for. The depth was differing across the systems, so average it is. And I also marked it clearly to differentiate it from depths that may differ at other rift zones in Iceland, and in the world. In other words, it is a very good and well built-up mathematical ballpark. The mileage of other volcanoes may vary. M’kay?
We are here left with known volcanic formations and rifts. 50mm of uplift in this intrusion, on top of a previous 100mm of uplift in a previous intrusion, we have a dyke that has formed along the MAR, and that is dilating across it, and magma that is now seemingly very near the point of no return.
Now, before you become way to happy if you see an un-checked and un-verified M0.1 earthquake at 1.1km depth, I will throw a spanner in your happiness.
There is a reason I only look at M2+ earthquakes in these situations, and either check them myself or wait until the Icelandic Met Office has checked them. I am most often waiting for the IMO, since I am famously lazy.
A small earthquake is much harder to locate for the system, and to locate properly manually. So, the smaller the earthquake, the bigger the range of errors will be. And this is especially true for the depth locationing.
The accuracy is also depending on the density and quality of the network at hand. Even at the best of times anything below M2 will be 1 or more kilometres off dephtwise at Reykjanes, but an M2+ will be about 500 metres, or even less off from the real location.
So, below 99.0 confidence rating (indicating manual check), and below M2+, do not get overly excited.
Anyway, we are now very near the runaway train that is covered in chickens and eggs. It would only take a comparatively small push to get it going now.
I hope this will not happen because unlike Fagrafjall this would directly impact the lives of many locals since any lava would move towards the town of Grindavik.
If an eruption would occur in the next few days or weeks, the current probable location for a vent to open up is near Lagafell, but this is speculative.
Due to the location, it pains me to say that it is now touch and go.