The Hell Machine

Guest post by Chad

The three musketeers, firing away

As is now abundantly clear, a new eruptive cycle on the Reykjanes Peninsula has begun. While an isolated eruption did happen in the ocean off the end of the peninsula in 1783, there has not been an eruption on land since 1240, today 781 years ago.

The eruption ongoing now has ratherbroken the regular cycle of eruptions at Reykjanes. Normally those begin at Brennisteinsfjoll and go west over a few hundred years, but this vent is on Fagradalsfjall, a mountain right in the middle of the peninsula. This area would not be expected to activate for at least a century yet, but here we are, watching the camera religiously as this new volcano grows before our eyes.

The magma is being generated at a great depth, nearly 20 km or twice as deep as normal eruptions in this area begin, and by mantle that is nearly totally melted to result in a lava that has a very high concentration of olivine. This sort of basalt even has its own name, picrite, and it is the closest thing to the ancient komatiite lava that erupts on the Earth today. This has been interpreted as a possible case for a shield volcano to form at, something which has not happened here for many millennia, but I think there are rather more profound implications for this. The magma is fluid and extremely hot, well over 1200 C, and that is enough to allow both physical and thermal erosion of the interior of the conduit, which will serve to only allow more magma to flow, in a compounding effect. Direct mantle eruptions are rare, but one that is currently ongoing if hard to observe is going on under the ocean near the island of Mayotte near Madagascar. It is not particularly intense, but since it began in mid 2018 some 5 km3 of lava has erupted to date.

The magma for eruptions like this is created by decompression melting as the pressure drops, a self feeding cycle that is limited only by the rate of mantle flow to the area. It is a self sustaining machine, the Hell Machine as I like to call it.

Hell Machines are a rare sort of eruption. Like said above they are perhaps best categorised as some sort of hybrid between a shield volcano and a lava flood fissure eruption. The eruption rates may be sustained at a relatively high rate for a long time, well above the rate seen in the formation of shield volcanoes but also never getting to the colossal rates seen in the big fissure eruptions, at least not for any long duration.

Mayotte is an example of a Hell Machine that is erupting now, it is in the deep sea and it may be that most such eruptions are, but there is a historical eruption like this that occurred on land and it offers some tantalising clues.

On the 1st September 1730 an eruption began on the island of Lanzarote, in the Canary Islands. It was relatively large, creating a cinder cone that was ultimately named Caldera del Cuervos, as well as over a few months the similarly sized Caldera del St Catalina and Pico Partido cones. This constitutes what would have been a typical if somewhat larger than average eruption in the Canary Islands, but we now know this first eruption managed to set off the Hell Machine. Eruptions continued into the next year, though it had declined from the initial high rates as expected, the flows were still fed at an impressive volume to sustain over a year. Submarine eruptions well off the west coast and going up all the way to just inland beginning in June of 1731 marked the completion of the hell machine, it was now in full working order, and eruptions marched eastwards back onto land to flood even more of the island. Observations after this are poor as few inhabitants stayed beyond this point but the eruptions continued into 1736, erupting some 5 km3 of lava, constituting the large majority of the volume of all eruptions in the Canary islands in recorded history. Lanzarote has no active central volcano, and has not possessed one for millions of years, the eruption of 1730-1736, or Montanas del Fuego as it is known locally, was fed out of the mantle directly by extensive melting under the base of the crust. Perhaps most notably, all the eruptions from the late Pleistocene and Holocene on Lanzarote before this were on the northern part of the island, the area where the Montanas del Fuego eruption occurred had not been volcanically active for possibly several tens of thousands of years or more, a certain similarity it does share with Fagradalsfjall.

Lava fields on Lanzarote. Source: Becerril et al. ,Assessing qualitative long-Term volcanic hazards at Lanzarote Island (Canary Islands), 2017 Natural Hazards and Earth System Sciences, 17, 1145-1157

The Signs

The first sign of a hell machine is that the eruption is persistent and erupts out of the mantle directly, something clearly the case already. An eruption should also be a rifting event, where the rift is kept open by new magma, and this also seems to be happening, as a dike has formed within the crust going from less than 2 km depth down to at least 7 km, with possibly many feeding points below this that are not resolved. The very high temperature of the lava will also serve to enlarge the pathways, something that may take longer than a week to become evident but which is expected.

So far our eruption is quite small, even tiny, but it has a big future. The first obvious sign could be that the vent increases in output, or that it stays constant but another vent opens. We may already be seeing this, as the vents are rather a lot larger than they were to begin with, and the output appears to have increased. Deflation is not observed along the dike either, which would suggest the eruption is being fed by continued decompression melting in the mantle, a process that may increase accordingly with the eruption rate and set the hell machine in motion.

What will become of the Reykjanes peninsula if this occurs? Perhaps the most direct consequence is that land within perhaps 10 km of Fagradalsfjall is potentially at risk. There is no danger to Reykjavik but Grindavik may suffer if not from the lava then from the gas emission. The effect it will have on the normal Reykjanes cycle is also unknown, most typically the section of the transform fault at Brennisteinsfjoll is the first to go but this time around the section going through Krysuvik was set of by the recent intrusion, leaving Brennisteinsfjoll in a precarious situation. Decompression melting at Krysuvik may well have already begun now in preparation to its awakening, and Brennisteinsfjoll could wake up rapidly following its eventual quake. Eruptions here will resemble those from the Reykjanes Fires, but taking a back seat to the lava flood that is just beginning now at Fagradalsfjall. We are watching history creating itself.

Perhaps this will not come to fruition, it may turn into a shield instead, or even stop. But with all the signs it is looking like the future is grand for the new little volcano.

Chad – 2021

866 thoughts on “The Hell Machine

  1. that reddish haze of last night… that’s a view that only foxes can get despite all this lanterns are hangin around…d..whoua! I wish to be there! iceland against sofriere is like an teacher touch against all live experience…

  2. HEY! Why have the Geldingadalir vents been replaced by a blue screen with some sort of countdown on it?

    • Happens from time to time at night, I suspect an empty battery, weather hasn’t been helpful for the solarpanels lately, so there’s no more power.
      It could also be for other unknown reasons, but it has happened before and the view will be back at some point.

      The “countdown” actually is the clock showing the local time, and if you enable the sound you’ll hear icelandic radio.

  3. Noticing a lot of comments on other apps where people aren’t interested in the St Vincent’s eruption because it is less photogenic. I find that sad as they are two very different types. One is pretty but so far harmless, the other has potential for great harm if pyroclasic flows hit populated areas or ash causes debris flows. Both have potential to learn from unless you only read picture books.

    • I hear you Ellen. I love to watch the eruption in Iceland as it is relatively harmless with minor impact for the people who live nearby. La Soufrière is a different thing altogether as could be seen in the past (e.g. 1902) and hopefully a bit less so during this eruption. It feels more like voyeurism to dwell on the harm it is doing to the people of St. Vincent and Barbados.

      The La Soufrière eruption definitely requires a different tone when we talk / blog / comment about it. This is where our fascination with volcanism has to be more serious in tone and content…

    • Its just that I much rather watch volcanoes that give plenty of warning before slowly destroying a rock field. Otherwise I just worry to much about the poor animals or people on the slope of the volcano.

      Also most of us don’t have a degree in geology and the fissure eruptions in iceland are a good beginner level.

      And some people just have a more photographic interest in volcanoes than a academic one.

  4. I see there is reports of steaming of the Upper parts of Rabacca Valley in regards to La Soufriere eruption. I took that as a potential vent is to open or result of a Lahar-the latter I suspect may be so.By all reports it’s bigger than the 1979 eruption. Not a people friendly event like the current activity in Iceland.

  5. Howdy all. Long Time no Sea… 😀

    I am one of your dragons. I lurk around in the back channel conversing with the other Moderators. My claim to fame here on VolcanoCafe is that “I plot stuff.” One topic that we have been kicking around, is just how large is the La Soufriere eruption. Albert threw out VEI 4 range as an estimate, I had no idea since I hadn’t been watching. I felt remiss in that, so I spent the evening rummaging around the VAAC reports. Now, first a note. VAAC Reports are SPECIFICALLY for the purpose of providing warnings and alerts for the aviation industry. They are not intended for any other purpose. What I do from time to time, is to use them for an unauthorized or sanctioned purpose. To estimate the eruptive rate of a volcano. Since my estimates serve no purpose other than for us to get a handle on what size an eruption is, and that I make NO CLAIMS as to their usefulness in a scientific or official realm, I think I may be able to slip under the radar and get away with it. The first thing you need to do is to weed out some of the errors.

    First of all, VAAC reports are for giving appropriate aviation warnings. These guys are going to error on the side of caution. Warning boxes are of no use for what we are doing. The only data you want are actual reported plume heights. Note the yellow areas in this sample.

    This is the only data that is of real use to us. FL100 means “Flight Level 100” Essentially, hundreds of feet. TGMCoy may be able to elaborate on this since aviation is his field. (If he is still using that moniker). In order to convert that plume height to cubic meters of dense rock equivalent, refer to Mastin et al appearing in Journal of Volcanology and Geothermal Research. doi: 10.1016/j.jvolgeores.2009.01.008. The long title of the paper is “A multidisciplinary effort to assign realistic source parameters to models of volcanic ash-cloud transport and dispersion during eruptions”

    Essentially, it is a method of getting usable data from what you can find. Their method is derived from work done by Sparks and is generally in close agreement with it.

    What I do, is gather all the VAAC reports I can find for a particular eruptive sequence and convert all of the time stamps and data points to the number of seconds from a reference point. To fill in the gaps, I run a linear interpolation from point to point on a per second basis. This is a potential source for error, so keep that in mind. {note, I am not a glutton for punishment, so I used a software tool to do this… specifically DPlot by Hydesoft Computing.} → (Full Disclosure; this is an unpaid endorsement. He is based out of my home state and I’ve had great customer support from him.) Upon completing this step, I now have an eruptive rate estimate for each second of the eruption. I then run an integration step (same software) to obtain the cumulative DRE of the eruption on a second by second basis. This is what it looks like.

    See that 4/10/2021 15:55 time stamp? (midway along the curve) that is where the column was reported above FL350, an eruptive rate of 639.68 m³/s.

    At over 1.2 x 108 m³… I think it’s safe to say we are in VEI 4 territory now.

    Enjoy. Feel free to discuss.

  6. I think that cousin up hill and the last born vent starts to act sincronously…and the lava field survive the night bright and shine…volume or temperature increased?

    • Volume, spatter out of the cones 1 and 2) has been quite large the last few hours, this indicates a larger flow.

  7. I’m expecting the St Vincent eruption to stay in VEI 4 range. The rate dropped off on the last report I saw. VEI 5 is a long way away.

  8. A bushplane just landed on the MBL side of vent 3. Sure didn’t need much distance to stop…

  9. RUV: The webcams are broken
    Unfortunately, no footage from our webcams is currently being sent to the eruption sites in Geldingadalur due to a failure. It will not be possible to repair them until tomorrow.

    posted 12/04 00:55

    • I wonder if the plane has something to do with this. Two people walked down to the left and not long after I started seeing moving shadows around the MBL cam.
      There could be a repeater station up on that hill. The fact that both of the RUV cameras went offline at the same time indicates there is a single point of failure between the two, and a repeater station would fit that quite well. On top of hills are good spots for them, too.

      And now the same two people are walking back down from the hill. At least one back towards the plane.

    • Hard to say. It could be deep degassing getting ready to enter an energetic phase… or it could just be magma moving about.

    • But it is worrisome. It’s already been popping 40kft

      Does it have more?

      • Good question. Past eruptions suggest it might.
        Also “tremor increasing” isn’t a sign of things settling down.

        Does anyone on that island still have electricity and/or an internet connection?
        If it doesn’t settle down soon maybe they should evacuate everyone.

        Bringing to mind that terrible hurricane in the Bahamas a few years ago. Dorian.

      • Hello GeoLurking!

        As there is a discrepancy between FL advisories and Volcanic cloud Monitoring NOAA/CIMSS in height is that simply because the upper level emissions in the Volcanic Cloud map is related to very fine particles and gases?

        The latter have consistent measurements of ash/dust at 18-20 km height but FL advisory has been to FL550 only once(?) during this episode.

        As seen here (last four hours);

        https://volcano.ssec.wisc.edu/imagery/view/#sector:West_Indies_2_km::instr:ABI::instr:ABI-MESOSCALE2::instr:MODIS::instr:SEVIRI::instr:VIIRS::sat:all::image_type:Ash_Height::endtime:latest::daterange:240

        New blast started ~50 min ago.

        https://whirlwind.aos.wisc.edu/~wxp/goes16/grb/meso_ircm/meso2_60.html

        • There was bit of overshoot visible in one of the satellite images. I suspect that’s the reason. Vaac is going to warn for anything that is a hazard to flight. The reason I stick to “observed” data is that is a firm time-stamped data point. After emission… vaac is going to report the hazard wherever it drifts to and even uses wind models to determine hazard regions in both position and vertical dispersion.

          • That is a webcam from Taal volcano in 2020. Someone pointed that out earlier. It’s clickbait.

  10. The graph seems to indicate a fairly steady eruption rate for the 24 hours or so between about 18:00 April 10 and 18:00 April 11……but the satellite data indicates discrete but large vulcanian explosions separated by a couple of hours each. The best way to estimate the mass of each explosion may be to look at the horizontal growth rate of the umbrella clouds…esp during the first hour or less, when the wind diffusion is a small factor compared to the gravitational spreading of the plumes. The umbrella clouds do look rather energetic, some have growth 10-20 km upwind before stagnating at the upwind edge. They seem to grow to more than 50 km in diameter in less than an hour, and there have been at least 10 of them. After looking at the satellite data, this eruption seems more impressive to me than I first thought.

    • The graph looks like that because it is an integration. My next plot I’ll provide the non integrated plot as well. It will be while though….I have a few things to deal with and the freaking Owls kept me awake all night. Now the damn rooster is going off.

    • That’s the one we call Greip (unofficial name suggested by VC). Search for Greip and you will find old articles and comments.

      • Correct! Just could not remeber the name. Any case, Greip just got a new magma injection this amateut thinks

  11. Heard from a friend who was at Geldingadalir a few days ago, that there nearly was a bad accident/darwin awards nominee. A woman wanted a selfie and stood really close to the blackened a’a front, when suddenly there’s a surge underneath and her leg gets stuck, she had to quickly wiggle her foot out off her boot and make a jump for it before her entire leg was burried.
    Don’t know if she had to hike all the way back with only one boot or if she got a ride from the authorities.

    • Is surprising that there have not been more accidents like this or worse. Even ignoring the volcano it is a mountain with no trees in a blizzard, which is just as dangerous if unprepared.
      It seems though that all the places where people can walk right up to the vents are inaccessible now, so probably safer than it was before.

      That being said theres no spattering at Gollum visible on the one working webcam so it might have a low lava level, 5th fissure soon.

  12. Geldingadalir cam was operational again at 11:48:30 for a short moment before going to black, so I suspect there are people working on it atm.

    • The MBL can is turned to show the flow into Meradalir from Fissure 2.
      Getting stable again, but I assume overloaded.
      They should offload to YouTube.

  13. Is the ash from La Soufrière heading west to Central America or east to Africa and Europe? I’m sorry, I know that’s a really simplistic way of putting it but I’m a novice at this. I’m just curious as to whether the ash will be joining La Nina in doing things to our weather in Western Europe in the coming weeks.

  14. Does someone know what Yellowstone’s prehistoric hotspot track was (earliest i’ve read was it being around Yukon in the late Cretaceous, but what did it do before that), and is it likely that either one of the two hotspots (Yellowstone or Iceland) produced the Siberian traps, since both seem to date around Siberia 250 million years ago, i’m not sure?

    • Neither produced the Siberian traps, I doubt that the plume creating that eruption still exists today. Oldest plume that is still active might be Kergualen, which was created maybe in the Jurassic, and it is showing its age. Hawaii might be next, it is at least 90 million years old. Reunion might be a similar age, but no ocean crust north of India still exists. Iceland is about 60 million years old, related to the opening of the north Atlantic and Baffin bay. Yellowstone is probably formed in the Cenozoic, its probably older than the Columbia basalts but not by much.

      Realistically the Siberian traps were probably a failed rift through Pangea, from the enormous trapped heat in the mantle.

      • YHS is a lot older than the Columbia River basalts. A 29.5 Ma caldera of it has been found at Crooked River in Oregon. There’s also a belt of adakite in Oregon which can be tied to it as well. Back beyond that it formed the Siletzia and Yakutat terranes starting 56 Ma.

        There’s even a possibility it might have been involved in the Laramide orogeny by being the formative agent of the theorised conjugate Shatsky rise which is one possible mechanism for the orogen.

    • I think Yellowstone is associated with the Columbia River flood basalts. Nick Zentner has a couple of really nice lectures on this (‘Flood Basalts of the Pacific Northwest’ and ‘Supervolcanoes in the Pacific Northwest’). See 58:20 into this video:
      https://youtu.be/VQhjkemEyUo

  15. It is suspected that quite a bit of the communities/buildings/structures in the way of pyroclastic flows have been destroyed #LaSoufriere

    The new and old domes have been destroyed. There is now a crater /opening at the summit of the mountain measuring about 825mx715m in dimension #LaSoufriere
    https://twitter.com/heidibadenock/status/1381581784064925699

    • “Pyroclastics likely to reach sea. Fisherfolk should stay at least 3 miles from coast near volcano.” – Professor Richie Robertson

  16. Even though it looks like the mbl camera has turned into a potato, it does look like the flow from the original twins is going towards the exit in the back. And while it is daylight and hard to see where the flowing lava is in the front camera, it seems to be supporting the notion.

Comments are closed.