After a couple of rather hectic volcanic weeks in the world of volcanoes I thought I would write a little something about two of these events. Namely Stromboli and Grimsvötn, the reason I am cherry picking those two is that they have a common theme, water.
Unless you have been to Stromboli you will not have a good grasp at how minute a volcano it is. The island itself, which basically is the volcano, is small compared to other single volcano islands. The two vents used to be tiny, and the small strombolian puffs of lava that it emits are miniscule.
How miniscule? Well, at least 2 000 years of recorded non-stop activity has not even made a dent in filling in the scarp after the flank collapse.
This has led people to assume that the volcano is a safe and tame volcano that you can climb and sit down watching its gentle fire fountains. And for fairly good reasons I might add, I have spent 24 hours sitting fairly near the vents studying them.
Problem is in the main driving mechanism. Most people would say that it is the magma moving upwards, but that would only have caused a minor trickle of lava slowly oozing over the crater rims. What produces the fire fountains are the opposite of fire-like lava, it is water.
Stromboli’s magma has an unusually high water content, making the volcano into a steam powered gun hurling lava into the air.
So, what happened was not necessarily that a larger amount of magma entered the system, instead a blob of magma with an even higher amount of water entered the system, and as the water expanded 900 times in volume it emptied out the entire barrel of the volcano in one big hurl.
Depending on how much of this water rich magma entered the system, the new modus operandi can continue for quite some time. So, it is a good idea that the authorities are performing at least a partial evacuation.
The videos of the eruption contain a couple of interesting clues just prior to the big hurl. First you have an explosion as the vent is blasted apart by gasses running before the big lava ball, and you can also see how the increased pressure is deforming the mountain causing a rock-avalanche to run down the mountain due to a partial flank collapse. Yes, it was not a pyroclastic flow as many states, that came later.
If you look at the images of the ash column you can see that it has two distinctly different colours, one is the regular ash from a blast like this, the other is eerily white. The white part is steam from the propellant force.
It will be interesting to see if this was a one off, or if this will be a new pattern for Stromboli.
Words like minute, small and minor, has never been associated with Grimsvötn. Basically, only two things are the same between Stromboli and Grimsvötn, that they are both volcanoes, and that both are heavily influenced by water.
If changes in Stromboli’s water content is a local problem, any changes in Grimsvötn’s water treatment is potentially a global problem, and most definitely a problem for Iceland. And something has definitely changed dramatically in Grimsvötn, and I will give a try at explaining what that means.
Up until 2011 Grimsvötn basically erupted the dregs left after the global-impact eruption of 1783. In 2011 the ejecta was juvenile material, and that means that it was hotter and richer in volatiles, compared to before.
If water is the main volatile ingredient in Stromboli, this is not so at Grimsvötn. Instead the volatiles at Grimsvötn is predominantly sulphuric gasses. This means that the lava itself is not overly explosive, and that without the water in the sub-glacial lakes, and the glacial ice itself, Grimsvötn would during eruptions produce fire fountains 700 to 1700 metres high.
A while ago Andrej made a couple of plots trying to see how the system had changed as represented by location of earthquakes around the upper magma reservoirs. The plot below shows the potential rim faulting based on the seismic data. The colorbar shows the terrain elevation for the plot on the right side.
What we found was dramatic to say the least. The influx of large amounts of high-temperature magma had enlarged the chamber significantly around the southern magma reservoir. The red square actually shows the area that is used by IMO for the cumulative seismic moment plot for Grimsvotn.
Prior to 2011 the reservoir basically followed the caldera wall, but now the reservoir has expanded 1 kilometre further in an area running from southeast to southwest. Normally this would have been performed by intrusions into the bedrock and it would have been a very noisy process.
Instead it seems like the hot magma has relatively quietly melted the reservoir walls. The reason that this was even possible is due to the ambient temperature of the surrounding bedrock.
The implication for the available eruptible magma is staggering. We are talking about a further 10 cubic kilometres of eruptible magma that does not show up on the GPS-stations as inflation. Not that there is a lack of inflation over at Grimsvötn, it has inflated well enough in the usual manner after 2011.
This increase in heat and new magma volume has had other drastic effects. It has increased the amount of geothermal energy release, probably with as much as twice the previous wattage per square metre. It has also changed the locality of the release of the energy.
Previously the main energy release was inside the lake, causing constant glacial melt that in turn caused large glacial water outbursts in the form of Jökulhlaups. As the water accumulated below the uniform glacial cover it after a while lifted the glacial ice and the water gushed out through a crack in the caldera wall.
Now instead the main location is around the east, south and western caldera edges. And the amount of energy released is so large that it has melted the glacier enough that rocks not seen for decades, or even centuries, are visible. And there are even a few spots where the glacier has melted sufficiently for the sub-glacial water to come to the surface.
This means that the melt either can go directly out under the glacier without causing lift of the caldera glacial cover, or that the water can expand freely and move up through the ice free (sort of) parts of the lake. In other words, the days of the large Jökulhlaups are over for now since the water pressure can’t build up enough.
The lack of large Grimsvötn Jökulhlaups is a great boon for Iceland, for as long as that now lasts. Not having to worry about the water content of 100 amazon-rivers gushing forth every now and then, makes life a lot simpler.
The drawback? Let us start with the most famous bet in volcanology.
Albert vs Carl Grimsvötn Forecast Bet
This bet was made under the mutual assumption that we could use the previously well-functioning model of forecasting, namely lazily looking at the Cumulative Seismic Moment curve handily supplied by the Icelandic Met Office and combine it with earthquake plots and GPS-trajectories.
This has worked well for the last couple of eruptions and would presumably have worked for every eruption after 1783 if we would have had the recorded data for that period. More data would probably have made our forecasts even more accurate.
Albert came up with 2020 as the most likely year, and I came up with 2019, both of through quite a bit of data wrangling and basic math. The price would be that the loser would have to buy the beer the next time we got around to having a few rounds in person.
The bet was of course made in good fun, and as a means to prove that you can forecast an eruption based on stringent scientific modelling, if you have enough data, and that the volcano does not change too much.
Both me and Albert caveated ourselves heavily on the “change” part, since we well know that if any larger changes occurred, we would both be thrown off by a curveball changing the basic parameters that we made our forecasts on.
I am a stickler for scientific honesty, and Albert is an even larger stickler. And I am quite happy with conceding that my interpretation is made invalid by the changes to the geothermal release and the changes to the upper magma reservoir.
The reason for this is that the method we both used is basically a function of pressure causing the roof of the reservoir bursting at a specific point, and that we could forecast a likely moment using Cumulative Seismic Moment analysis as a tool of measuring pressure increase.
That model did not take into the account the effect of hotter magma melting the sides of the magma reservoir, and presumably at the same time melting the roof of the reservoir, as evidenced by drastic increase of geothermal energy release. At least for me I have lost the bet due to nature throwing curveballs.
Even if I hit the date, it will be for the wrong reasons, so no beers will be claimed by me.
So, now that we are both probably wrong, we can do the fun part of science, trying to find a model for the new conditions that nature have produced. I am therefore proposing that we reset things and make new forecasts as soon as we have put the new possibilities into our respective models.
This will be a bit ad hoc since it will obviously take me some time to figure out what the effects will be.
The increased volume might indicate that the risk for another Lakí style eruption has increased, but I do not really believe this since there should not be enough strain yet at the fissure swarm.
Instead I think it is increasing the risk for an eruption even larger than the 2011 eruption. And since the reservoir has expanded beyond the caldera rim, a larger eruption could imply a future partial caldera event. But this is so far just a small possibility, with a 2011 sized eruption being the most likely outcome.
There is also the possibility that the eruption will be voluminous, but not overly explosive, since the eruption will quickly build a cone or a fissure hump sticking up out of the water and ice. This is the solution that would make the Icelandic people happy, since this would limit the amount of ash that they get dumped on their front porches.
One thing is though clear, we will not have a small insignificant eruption when it happens.