Could you drill into magma and trigger an eruption? A How-to Guide, Part 2
Eruptions have come up through wells before, but making this happen on purpose would be tricky.
This post continues Part 1’s discussion about whether a sufficiently motivated individual could bring about a volcanic eruption through deep drilling. I started by recounting accidental and intentional magma strikes during drilling, and what we can learn from those. This post finishes the story, first, by recounting an actual borehole eruption in Iceland and, then, speculating on what you could do to make it happen again.
Eruptions through wells
So, magma has been encountered during drilling and seen to make it’s way into and up a well (see cases from Iceland and Hawai’i described in Part 1). But it does so rather slowly, about several meters per minute, freezing in place long before it making it to the surface. In contrast, there is one well-documented case of a genuine eruption that occurred through a borehole. And the magma in that case moved much faster.
This all happened in Krafla (again) in Iceland. Between 1975 and 1984, there were a series of eruptions here called the Krafla Fires. The magma that erupted was basalt, which has low viscosity and so seems ‘runny’. And while the eruptions are spectacular, they are largely contained to lateral fountains called fissures, rather than enormous explosions. They also produce lava flows that make for some rather compelling drone videos as well as some truly frontier civil engineering.
A brief aside: why the Krafla ‘Fires’ and not Krafla ‘Eruptions’? From what I can tell, it’s the literal English translation of the Icelandic name of the event Kröflueldar, where the “eldar” (plural of “eldur”) literally means “fires”1. And, of course, magma is quite good at setting things on fire.
Anyway, several studies have since shown that this runny, low viscosity Krafla basalt is quite mobile underground, moving not just upward but also sideways. It’s quick too, moving through cracks in the rock at up to a meter per second, which is about walking speed. It’s also much faster than the other kind of magma, a rhyolite, which might rise at a meter per minute. But the main take-away is that this basalt magma is opportunistic — it likes to take advantage of weak pathways through rock to get around. That’s usually cracks and faults, but can also be a borehole.
Gudrún Larsen and colleagues describe some uncommon events in September 8, 1977, during the Krafla Fires. Earlier in the day, there had been some earthquakes and ground subsidence, before a fissure eruption started in the afternoon. This was an 800 m long crack in the Earth that fountained lava several stories high. It happened just to the north of the Námafjall geothermal field, where there was operating a 5 MW power station fed by several boreholes. That fissure eruption went for a couple of hours and then stopped.
Then, shortly before midnight on the same day, an explosion was heard at the geothermal field and an incandescent column was seen rising from borehole number 4. The sound is variably described as like “a hail of gunfire in a hurricane” although I get different translations from Google and various AI models. It lasted for about 30 minutes. We only know all this because Hjörtur Tryggvason who worked there got close enough to take some photos2.
The next day, they found the same borehole, now ruptured at the top, to be discharging a steam jet. But, aside from breaking some of the surface pipework, the eruption hadn’t much disturbed the well and it continued to supply steam to the station. There were a bunch of fresh volcanic rocks scattered around the wellhead confirming that it had indeed been an eruption. But not a large one — only 1.2 m3 of magma had been ejected at an estimated rate of 25 kg s-1. It’s hard to get a lot of rock up and out through a narrow 6” diameter borehole.
From cross-referencing against landmarks taken in photos the next day, they were able to estimate that this ‘eruption’ was 20 to 40 m high. A rough calculation you might have learned in high-school is to equate the gravitational energy of rocks at the top of the column to the kinetic energy of those same rocks when they first emerge from the well. If you apply that here, it suggests the magma in the well was travelling around 20 to 30 m s-1, or about the speed of a car on the highway. The magma was assumed to enter the borehole about 1 km deep, so this means it would made the trip to the surface in less than a minute3.
So, what can be learned from the Krafla Fires borehole eruption?
It’s not just rhyolite that can be encountered in boreholes, but basalt as well. This type of magma, being much runnier, moves around a lot faster than rhyolite.
This instance involved a good deal of luck in that, even though the original borehole did not find magma when drilled, some magma later found it.
Absent a well filled with heavy, cold drilling mud, but rather containing a hotter mixture of steam and water, the magma had a lot less resistance to flow on the way up.
So, could you get molten rock to come out a borehole on purpose?
From a technical perspective4, the evidence suggests this would be pretty tough but let’s work through the concept just for the sake of argument. I’ll assume we’re limited to existing drilling technology and focused only on volcanic regions. This means we’re not likely to be searching much deeper than 4 km. So, here are the problems.
Shallow magma is quite rare and hard to locate. After decades of geothermal drilling and many hundreds of wells, there have only been three confirmed magma strikes (that I know of). The likelihood of hitting magma with an untargeted well is probably less than 1%. So, drilling random holes at $10 to 20 million each and hoping to eventually hit something isn’t going to be a profitable strategy. Is there anything that could be done to improve targeting?
Well, you could obviously boost your chances by just redrilling a previously struck magma since you now know where that is. This is essentially the proposition of the Krafla Magma Testbed that wants to go back to the Krafla rhyolite to install a magma observatory. However, in all those prior hits, magma climbing the well froze after a few meters. It stands to reason you’d get a similar outcome on any return visit unless you did something drastically different with the well (more on that further down).
Alternatively, if self-preservation wasn’t a priority, you might go prospecting for a shallow magma that was already starting to show itself at the surface. One thing that happens to magma as it gets shallow and the pressure on it decreases is that the gas that was dissolved inside it starts to leak out (called degassing its a bit like when you open a soda and all the CO2 bubbles out and escapes). Other times, the magma will just push up the ground above it and so, just as you can guess that something is under the bed covers by looking for lumps, you can use satellites to look for suspicious bulges in the Earth’s surface. But while gas and uplift are both signs of shallow magma, they’re also indicators that it might be about to erupt anyway. Drilling there would be a particularly foolish way to execute an already foolish endeavour.
Anyway, supposing you did manage to find and drill a shallow magma, it would probably still freeze in the well on its way out. After all, from the magma’s perspective, it is being asked to leave its warm cosy home and transit a narrow, freezing tunnel that’s several kilometers long5. The heat seeping away causes molten rock at the edges and front to crystallise and harden, which creates even more drag for the magma pushing from behind. Rock freezing to the edge of the well makes it even narrower for the magma that follows, like putting on-street parking in a tunnel.
This freezing is also happening on a clock. A really nice review of magma dynamics by Rubin (1995) positioned the problem of how magma gets to the Earth’s surface as essentially one of thermal survival. If magma is rising at a given speed, say 1 km per day, and it has to travel 10 km to reach the surface, then we’d predict an eruption to occur after 10 days. But if it only takes 2 days for the magma travelling underground to freeze solid, then that eruption will never happen — instead the dike will arrest deep in the Earth. Basically, there are two ways to change this equation: either the magma has to find a way to move faster, or it finds a way to not cool down so quickly.
What makes magma move faster? We have at least one example of a very rapid magma, and that was the borehole eruption during the Krafla Fires. The main difference there was the type of magma, which was an especially runny basalt instead of a viscous rhyolite. So, one thing you could do to improve your chances of getting magma to the surface is to focus on finding and drilling shallow basalts. This was more or less the aim of the Kīlauea lava lake drilling, although the pressures pushing basalt out of a lava lake a likely quite different to those of a basalt several kilometer underground.
Alternatively, instead of getting the magma to move faster, you could slow its rate of cooling by drilling a larger diameter well (a wider tunnel). Some years back, we had a go at throwing numbers at the question of “how wide a well would you need to get a slow-moving rhyolite up to the surface?” We concluded that a hole 15 m across might do it. Although this is far wider than the holes produced by standard drilling technology, it is within the realm of Vertical Shaft Sinking machines used for excavating mine access shafts.
Those shaft sinking machines aren’t designed to go down 2 km through volcanic rock, but even if they were, what might we expect upon encountering a magma? The difference with this scenario compared to ordinary drilling, is that you wouldn’t have a heavy column of drilling mud behind you, pressing down on the magma and keeping its pressure nice and high. Instead, as you approached the magma from above, it would likely start to experience a destabilizing pressure drop due to the solid rock ceiling you are lifted off the top of it. As I mention above, rapidly depressurizing a magma like this can cause the gasses dissolved in it to come bubbling out. If these can’t escape in an orderly fashion, they’ll start to fracture and fragment the magma. The resulting feedbacks can cause more depressurization, more gas release, basically a chain reaction leading to an explosion. Whether that eventually lead to a sustained eruption (or just collapse your horrendously expensive vertical shaft) would be hard to predict.
So human-triggered eruptions are pretty unlikely?
I think the short answer to this is ‘yes’ if you’re talking about volcanoes. It’s a mix of several things, including how hard it is to find magma, it’s tendency to quickly freeze, and the lack of proven technology that can dig vertical shafts wide and deep enough.
But, that doesn’t mean there aren’t other kinds of eruptions that humans can, and likely have, triggered in the past. A good example is the 2006 eruption of the Lusi mud volcano in Indonesia, still erupting to this day, and which very credibly had an oil & gas drilling trigger.
Mud is a bit different to magma — it’s much less viscous and so flows more easily. It also doesn’t freeze solid. In fact, there are several fluids that are regularly encountered underground - water, steam, oil, natural gas - that could conceivably can flow back to the surface under their own impetus. These present the well-known threat of a well blowout, which is the topic I’m thinking to write about next.
At least, this is how ChatGPT tells it to me, but if my Icelandic friends would like to clarify that would be brilliant.
At least, this is who is credited in the initial account in The Naturalist. But then the Larsen et al. article credits a Hördur Vilhjálmsson, so I’m not 100% sure who was there observing.
Dartevelle and Valentine (2008) have done a more sophisticated study. They use computer models to consider mixing of the magma with geothermal steam and some non-steady behaviours. They also come up with an average velocity of about 30 m/s, similar to Larsen et al.
I’m not even touching the legal, safety and ethical dimensions of triggering a volcanic eruption but be assured that these are difficult.
It’s also full of cold, heavy drilling mud blocking the way, but I’ll assume that an attempt is made to pump this all out and clear the way as much as possible.