A while ago Jesper Sandberg asked a question about if Grimsvötn could sprout forth a lava lake. I somewhat sloppily answered that I did not think that it was possible. Jesper then asked why not, and I got somewhat flustered since I did not have a ready answer. I promised to come back with a good answer in the form of an article.
I thought that would be easy and earmarked the next weekend for a short article on the subject. Oh my, what a case of hubris that turned out to be. Week after week I had to postpone the article while I pondered things.
The problem is that we do not know a lot about lava lakes, in fact we know less about lava lakes than any other type of eruption. This is a bit odd, since they are the most openly visible type of eruption there is.
I have often stated that the edifice of a volcano, and indeed their eruptions, are just the effect of the cause that is the real volcano, and the real volcano is lurking at depth, often coming into being tens of kilometres down, or even hundreds of kilometres down. What we see is ruled by the unseen plumbing deep down.
In other words, just because you can see things in a spectacular fashion, it will not help explain what is going on at depth.
Lava lakes in general
Let us start with summing up what we do know about lava lakes. They normally form on large volcanoes that has a high lava temperature, Kilauea, Nyiragongo and Erta Ale are perfect examples of this. All of them are fed in turn by hot magma coming up from extreme depth via their respective mantleplumes.
Then we have Mount Michael, Yasur, Ambrym and the others who are fed by subduction and back-arc volcanism of a type producing hot basaltic lavas. The oddball is Mount Erebus that has a lava lake that is colder and contains more evolved lavas.
Obviously, there are other volcanoes that now and then produce more short-lived lava lakes. But the question was about persistent lava lakes of the Kilauea type, so let us stick to those.
A lava lake is a type of low-level eruption where the input of either heat or fresh material is hindering the lava to solidify, this is aided by the surface solidification of the lava that in turn acts as an insulating material hindering heat from escaping.
The solidified surface never creates a solid layer, instead it is reminiscent of disjointed ice-sheets drifting on a turbulent river, or like continents drifting on top of the less solid mantle. After a while these sheets get pulled down into the molten lava and melt.
This is roughly all that we know about lava lakes, I obviously condensed it a lot, but the gist is that this is the bulk of our knowledge that is true for all instances of lava lakes. None of this really helped me to answer Jesper’s question, so I had to sit and figure out more things about lava lakes.
Problem was just that as soon as I figured out something for lava lake volcanoes that are comparable to Grimsvötn, it did not fit Mount Erebus. A good fundamental thing must after all function for all instances observed, or your science is warped. In the end I literally discarded hundreds of ideas.
My favourite scientist is Einstein. Not because he came up with something incredibly convoluted and hard to grasp, instead I admire him for coming up with something that is so ridiculously simple that it encompasses almost everything. One single thing to rule them all, to misquote Tolkien.
It has been erroneously stated that only 100 people understands Einstein’s theories of Relativity. Nothing could be more wrong. In fact, the fundamental part is quite probably the single easiest scientific theory to understand. ‘What is observed will be different based upon your point of observation’. That is all of it, one sentence is needed to explain the fundamental part of the theories of relativity.
Yes, the mathematic proof behind it, and the effects down the line can be a bit hard to understand, but almost all parts of the theories of relativity can quite easily be explained by using common words. Testing the theory? Nothing could be simpler, all you need is a car, a friend, functioning eyes and an apple.
So, I discarded finding something complex, and went for something simple, the simpler the better. To be able to do this I went back to Eyeball mark 1 and watched hours of videos of bubbly lava lakes.
What I came up with are 3 observable facts that anyone that has watched lava lakes must also have seen. But, as where Newton just saw one effect of a falling apple, Einstein saw something very complex emerging from the falling apple under motion. In turn those 3 things gave me an idea of how to answer Jesper’s very tricky question.
Modes of convection
A lava lake can only be affected from 2 modes of convection. Either an open conduit is transferring large amounts of heat energy upward into the mass of lava in the lake, or lava is moving up through the conduit into the lake aiding to keep it in a molten state.
Both modes come with their distinct set of problems. For heat convection mode to work the conduit must be substantial, the reason for this is that rock (molten or otherwise) is a very good insulator. A narrow conduit would not be able to transfer enough energy upwards to keep a substantial lava lake like Halema’uma’u open.
Now, convection through lava injecting through a more normal conduit into the lava lake does not require a large conduit. But it comes with a massive problem that must be explained. If lava is entering the lake the level should be going up at a steady and observable rate, causing normal lava flows as the lake over-flowed.
We therefore must input an ad hoc solution here to explain why lava lakes does not have runoff lava rivers happily galumphing down the mountainsides in rivulets and happy little lavafalls. Let us assume that the excess lava is going down into the depth again due to having become cold and dense (for instance after having “frozen over” in the form of lava-sheets).
Problem is just that the conduit is busy pushing up warm fresh lava, and unless that conduit has a very odd topology, it is not possible for our cool lava to go down here. Instead we need a runoff conduit that is separate. Question is just why there is not warm lava coming up that way too?
Luckily for us this has been observed briefly during one of the Krafla Fires eruptions. During a smaller eruption two fissures opened up at the same time, one erupted lava, but the other one was dry. The lava from the first rapidly flowed into the other and disappeared.
If this had happened in a crater, or small caldera, and the eruption had continued, we could have gotten a lava lake with true lava convection.
Kilauea and Krafla both exist on top of rifting fissures, so it is not entirely impossible that two rift-conduits opened at Halema’uma’u, one from the magma reservoir, and one leading to somewhere. In the case of the Icelandic eruption the rift probably did not go into the magma chamber, so it would probably in the end have over-flowed.
But, in the case of Kilauea we have the potential that a rift opened towards Pu’u O’o and helped to feed that eruption. It does not explain other lava lakes where there is no sub-eruption going on.
This part requires further studies from young strapping scientists, and I will happily leave it to them. In our particular case, this is not where we will find the answer to Jesper’s question. Instead, let us discuss size and flowrate of ejecta.
The small thing about size
As you watch a video of a lava lake, or if you are looking at one live, you will be filled with a sense of awe and majesty over how big the phenomena are. One who was completely blown away was Mark Twain when he observed an old lake at Kilauea that covered a far larger part of the caldera than Halema’uma’u, it was probably one of the few moments in his life when even he had to struggle for words.
But what we are seeing is the effect of the eruption, and not what is actually coming up. A large eruption spanning hundreds of cubic metres would simply just overflow creating those happy rivulets and lavafalls as it fell down the mountain.
A monogenetic lava shield can only form if the flow rate is small enough, typically counted in the low tens of cubic metres per second or less, small flow is here building a large volume due to the long duration of the eruption.
It therefore gives that a lava lake eruption can’t have a higher outflow than a shield since it would create a lava flow or flood basalt. A flow in the tens of cubic metres would instead cause a monogenetic shield (unless of course the hypothetical return flow conduit is large). This leaves us with just a few cubic metres, and that is amply enough to keep the lava well above the point of solidification.
The Great Equilibrium Machine
The essential part that is a logical must, and is indeed observably so, is that a lava lake is a steady state eruption remaining at equilibrium throughout the eruption. This is also true for all known lava lakes.
Yes, the level of a lava lake will ebb and flow over the course of its existence, but it will ebb and flow around an equilibrium point. Now we are ready to start discussing our way to an answer of the question.
We now know that the eruptive rate is low and fairly constant, something that is made apparent from those volcanoes that are closest in comparison to Grimsvötn. Both Erta Ale, Kilauea and Ambrym, are large shield volcanoes.
They have been built up during numerous eruptions that individually are not that big but are numerous or have a long duration. Obviously, I am aware that Kilauea can do big eruptions too, but in regards of the lava lakes this is not important for the discussion.
This makes them prone to have eruptions small enough to produce an eruption at equilibrium between inbound energy and outbound energy. Or intruding lava and outgoing lava in case there is such a hypothetical thing (I do believe this is the case, but I can’t prove it).
They are also rift volcanoes, making them more prone to open two or more conduits in a small geographic area for the lava to circulate through.
Now let us instead look at Grimsvötn. It is a mantleplume volcano of prodigious size and power. It is also a rift volcano, but unlike the others the rifting is an inherent feature of global tectonism, that in turn is pulling the volcano in half at stable rate. The last part has a rather important effect on when it will be erupting.
For normal volcanoes you try to calculate how much pressure the lid of the roof of the magma chamber can withstand before an eruption will occur. At Grimsvötn the pertinent part is instead how much strain the roof can withstand before it is pulled apart.
As the lid is ripped apart a considerable over-pressure is released from the reservoir in the form of the eruption, and at the same time the pent-up local strain drops to zero, and as the over-pressure is gone the lips of the lid will weld shut.
In actuality it is obviously a combination of both intra-chamber pressure and tectonic strain that causes an eruption at Grimsvötn, but the strain part is not as evident at the other volcanoes, nor does it happen at the same freight-train geologic speed.
We normally say that Grimsvötn does not have eruptions below the VEI-3 mark. This is in a sense of it true, nothing below this size can erupt through the thick ice cover. Smaller eruptions would also be quenched by the ice and the water.
So far, we have though not seen any signal indicating that small abortive eruptions have occurred. We can also compare with other Icelandic central volcanoes. They also tend to have large eruptions with a time period in between them.
Icelandic volcanism, nor Grimsvötn does equilibrium. It is either all, or nothing. So, to specifically answer Jesper’s question, regardless if there is ice and water inside the three calderas of Grimsvötn, there is no chance that there will be a long-term eruption small enough to be at equilibrium.
Iceland has twice created false lava lakes in living memory, one during Holuhraun and the other during the Surtsey eruption. They where though not real lava lakes since they constantly over-flowed with lava due to the high eruptive rates.
As I reread this, I could feel how Albert would at this point be formulating a rather stern comment on what I wrote above about Einstein. I should therefore explain that I happily simplified things so much that I wilfully mixed together the two theories of relativity that Einstein wrote, and that I simplified things so much that it is impossible to gain any true understanding of the theories.
Obviously, Albert is correct in what he would probably have written. In my defence I can only say that I wanted to convey how a fundamental simple thing can affect everything. Mea Culpa!