The Life Aguatic

Agua in the background. Photograph by ZackClark, Wikimedia Commons.

We got a letter from one of our readers this week pointing out a seismic series near Agua volcano and Ciudad de Antigua in Guatemala.

“I’ve been living in Guatemala for a while, staying in Antigua. On April 27, the area around, and on, Agua volcano started to shake. We felt at least 8 quakes that day. They were very short (1-2 sec) and different from what the usual earthquake feels like. The locals never felt anything like that before. These quakes lasted for a week. Please see the official report I’ve attached to this e-mail.

Later I’ve checked the official reports and they showed some more quakes in the same area (April 7-8, also attached). There’s no new data after that.

I know nothing about volcanoes, and I’ve been trying to find out more about what’s going on but there’s little information. Different web sites didn’t even mark those quakes on their earthquake lists and maps, just few. But the official report looks scary.

Are these quakes related to Agua volcano and how? Could they be an alert? How dangerous it is to stay in Antigua?

Maybe these are silly questions for someone with the knowledge. But I’m dying out of curiosity (and a bit of too much adrenaline). I hope someone can explain/comment it. I’d really very much appreciate some info.”

Before I try to answer these questions, I will give a brief background of the surrounding geologic setting.


Background without Steve Zissou

Agua photographed from Pico Mayor, Acatenango. Photograph by Kevin Sebold, Wikimedia Commons.

Guatemala is heavily influenced by the subduction volcanism in the country, creating a range of volcanoes running the length of the country, and creating the volcanic highlands.

The subduction process has created a seismic hotbed with numerous large earthquakes occurring at a high rate, destroying 3 out of 4 capital cities in the country. The one not destroyed by an earthquake was destroyed by a lahar from Agua.

This means that large parts of Guatemala are crisscrossed with tectonic faultlines. In short, things tend to move about quite often in Guatemala, something that is amply evidenced by all the church ruins in Antigua.

If we turn to the volcanoes around Antigua that most people know about you have 4 of them. These are the Acatenango-Fuego double volcano, Agua and Pacaya. Currently both Fuego and Pacaya are erupting at the same time, something that is pretty common.

But, and this is a big but, these four volcanoes are dependent on a fifth far larger and less well known volcano. The name is Amatitlán, it is a large serial VEI-7 caldera mostly known for the lake that it contains.

In most ways it is different from its neighbours, firstly it does not look like a volcano at all, it is more of a hole in the ground. It erupts rarely, but always spectacularly. Either it will extrude a lava dome the size of a small mountain, or it will just have a large explosive event creating a new vent, or a combination of both things.

As far as we know Amatitlán has erupted 27 times in the last 250 000 years, giving an average of almost 10 000 years between eruptions. All of the known eruptions have been large. Another way to look at it, the entire Mayan civilization up to this date has existed in between cataclysmic eruptions from Amatitlán. Thankfully no eruption seems to be around the corner from this volcano.

The reason that there has been no large eruption is found out on an en echelon fissure swarm running outwards from the caldera, containing what can be perceived as distal flank vents. Pacaya is the first one of these, and it is a true flank vent situated on the caldera rim of Amatitlán.

The next over is the dormant Agua, that is believed to have erupted 10 000 years ago. And at the end of the fissure swarm we find Fuego and Acatenango.

As long as Fuego and Pacaya is erupting, the pressure at Amatitlán will be kept below the eruption threshold, so in a way Acatenango, Fuego and Pacaya works as large pressure relief valves.

Amatitlán is also highly seismically active, but I will not go on about this here. For those interested I am leaving the link to a previous article I wrote about it.


The current seismicity and answers

82 earthquakes. Legend by INSIVUMEH.

The current earthquake swarm is not that large for the area, containing only 5 M4 events, and the largest is only an M4.2. And this is not a lot for the area.

The origin is unlikely to be caused by regular tectonic activity. First of all, the earthquakes are not following a known faultline, they are also widely dispersed over quite a bit of area without any discernible pattern indicating a faultline.

Even though the earthquakes are small to moderate they would be quite noticeable for the population of Antigua since the field of earthquakes run out from the volcano of Agua in under the city, and towards Amatitlán.

As I am writing this, I had hoped for one of Andrej Fliis wonderful plots, but it has run a bit late, and I needed to write this article this weekend, so I freely admit that I am lacking a crucial piece of evidence right now. I will rectify this later as the plot is done with an update.

This means that I can’t ascertain fully which of the below two things are happening.


Flank weakening

Stratovolcanoes in Guatemala tends to be rather steep and unstable since they are mostly built by magmatic rubble. This means that quite often parts of the flanks of the Guatemalan volcanoes tend to fail, and large amounts of volcanic debris start to flow for large distances.

Both Pacaya, Acatenango and Fuego has done that, but so far Agua has staid put. Flank failures are among the most dangerous events that can happen at a volcano. Most people know about the explosive failure of Mount St Helens in America, but not everyone is familiar with the far larger explosive flank failure in 1902 at Gagxanul (Volcán Santa Maria) in Guatemala.

There are though quite a few things against this being related to any future flank collapse of Agua.

The main reason is that the current activity is pointing toward the volcanic highland side of Agua instead of towards the coastal lowlands. On the highland side the edifice of Agua is only 2000 metres, but towards the coastal plain it is 3500 metres tall. This means that the gravity is constantly pulling the mountain towards the sea, and that it is buttressed against the highlands in the other direction.

So, currently I will rule out that Agua would catastrophically fail.


Magmatic intrusion

Rare image of two facehuggers in their natural habitat. Blue Sky Bar, in the light of Fuego.

Normally when you track a volcano prior to an eruption you look for deep small earthquakes indicating that magma is arriving from depth. Sadly, the national seismic grid in Guatemala is not good enough to track these minute earthquakes.

The equipment at hand has for natural reasons been aimed towards known active culprits to detect what they are doing, especially since they tend to kill people. This leaves the less active, or dormant volcanoes without coverage.

This is not in any way any criticism against the geological authorities in Guatemala, it is just a recognition of a fact. That being said, Agua being situated between the grids aimed at Pacaya and Fuego is reasonably well monitored for being a long dormant volcano. In fact, it is better monitored than 90 percent of all US volcanoes.

Anyway, time to stop rambling. There is another way to detect magma, and that is to look for shallow emplacements in the magma reservoir, as magma arrive a magma chamber will start to creak and groan as the pressure builds up. And those creaks and groans are the earthquakes often associated with volcanoes.

If it is a more active volcano the earthquakes are often reasonably situated under the volcano, and you will get a ball shaped series of earthquakes around a void that is the magma in the chamber.

In long dormant volcanoes you instead tend to get more confusing shapes of earthquakes since there is no longer any easy way for the magma to enter the magma reservoir, or the magma must build a new magmatic system.

At El Hierro in the Canaries we saw an array of dykes and sills form as magma intruded as sheets in the bedrock following the point of least resistance.

Since we can rule out this being faultline tectonics, and it is unlikely that it is related to a future flank collapse, we can therefore ponder the idea that this is a dyke/sill system forming as magma is intruding towards Agua.

For a non-volcanologist this may sound alarming, but it is not especially alarming. First of all, volcanoes like Agua tend to go through cycles of dormancy and renewed activity. During a phase of renewed activity, they can have dozens of eruptions spanning a few hundred years, and then quietly go back to dormancy. After all, erupting is what volcanoes are supposed to do.

There is also another thing to remember. Volcanoes do not erupt out of the blue. They tend to be quite noisy prior to an eruption, and the amount of noise tend to increase significantly the longer a volcano has been dormant.

Gagxanul is a spiffing example of this, it took at least 3 months of incredibly strong seismic activity before it finally erupted. In all probability Gagxanul had several seismic crises prior to the eruption, but they where not recorded since nothing came out of them.

What we are seeing now is probably just a small sheet intrusion that will not amount to anything on its own, at least in the near future. For Agua to be able to roar back into life we will have to see swarms like this lasting months, and at a far higher seismic level.

So, to answer the most pressing question for our reader. Is this dangerous? No. At least not in the immediate future. If Agua would really come back to life, we will know about it weeks in advance and the authorities (that are quite good) will know about it weeks or months in advance, and they will do any necessary evacuations prior to the beginning of any serious fireworks.

Currently I am not worried at all, and I have most of my family in the vicinity. I hope I have dispelled any worries for our readers about visiting, or living, in Antigua.

And for any of our readers pondering visiting Antigua, please go to the Blue Sky Bar and climb up to the roof bar along the narrow creaking cast iron staircase (mind your head if you are tall). As you sit there you can have drinks mixed by some of the worlds best bartenders while looking at Fuego erupting in the night with stars shining above you. Don’t mind the two smiling idiots looking at each other all the time, it is probably just me and my wife.

Carl Rehnberg



140 thoughts on “The Life Aguatic

  1. Carl is the red glow in the bar… nightglow from Fuego?
    Carl and Carmen illuminated by volcanic action

      • Wow then it was a pretty intense eruption then!
        Huge rubbly channelized viscous Aa flow made its way downhill right?

        • More like large fountaining and rubble rolling down everywhere. Not that much channeling if I remember correctly. I was a tad occupied with other things that night…

          • Fuego is crazy steep ridicilously steep
            Iconic stratocone but Shishaldin is far more perfect and cartoonish.
            Fuego may do a santa maria flank failure one day and that will be directed towards the sea and not towards major cities
            It will still be deadly when it happens and very very dangerous
            many many km3 sized St helens event.
            Unlike St heles Fuego is far more mafic and hotter
            After the collpase a Pacaya like thing may grow as a New Fuego.
            Still the chance for that is rather small I think?
            Fuego needs lava flows to build its base and stabilize it
            Fuego have done lava flows before

            The 2018 event was a small flank failure that decompressed hot gas rich magma I think

          • Carl another musing
            The highland Guatemala north europe looking conifer forests:
            Are they Ice Age refugia? that escaped the lowlands after last glacial maxium ended
            Or are they just native and evolved there?

      • That’s odd…the glow from Fuego in the top left corner is apparently from a point BEHIND Carl and Carmen; so how come it is illuminating their faces? Is it an illusion done by mirrors. or am I losing my remaining marbles?

        • It is actually attributable to the bad lighting in that angle, the glow is more on my head from behind, exactly as your remaining marbles told you.

    • IF Pacaya and Fuego are pressurize valves of Amatitlán
      Then thats an indicator that Amatitlán haves a rather large quite fresh magma supply
      Pacaya recives the freshest magma in Amatitlán system as its the most basaltic and hottest.
      high for a subduction zone
      But there are other subduction zone volcanoes like Etna and Ambrym that haves large supplys

      • Not sure that Etna is a subduction volcano as such, its origin is rather more complicated than that

      • hmm, italian volcanoes in general have a high influx rate. so do some volcanoes on the asian side of the pacific plate. but yeah, most high influx volcanoes are not subduction sourced. but yes, amatitlan is a very active caldera system, probably more influx than the aira and aso systems in japan or any of the italian volcanoes. i don’t think any of those systems, aside from perhaps aira, can keep two stratovolcanoes erupting nearly constantly. icelandic volcanoes would have a similar rate of influx perhaps, but they don’t have constant pressure reliefs. not big enough, the crust is too thin, and the magma there is mostly basaltic.maybe hawaii is also a good comparison for influx rate? idk, i’m just musing to myself

        • As long as these larger systems are happily venting they are not much of a concern, it is when they gum up for an extensive period that things can become truly “interesting”.
          We do not have to go far from Amatitlán and we get an even larger system, Atitlán. This one has coughed up a VEI-8 eruption, the 2800km3 Maria Tecún Tuff. It has not erupted for quite some time and is quite busy inflating, dunking village after village under water as the entire caldera is tilted on it’s side. Privately I would be quite happy if the side vents there started to erupt continuously.

          • To further expound, the current Atitlán III Caldera was formed during the Los Chocoyos Event at 350km3, Atitlán II was a mere high end VEI-6, Atitlán I was the MTT Event, the second largest eruption in the last 10 000 000 years.
            It has one active volcano, and one dormant. Volcán Toliman has not erupted for about 2 000 years, but Volcán Atitlán (not to be confused with the mother-caldera volcano) erupted 116 years ago.

            As such it is the longest known interlude between eruptions, but during that time we know that there has been steady inflation under the side of the caldera below Tolimán and Volcán Atitlán.
            We know this from the Lake tilting, and that was going on even during Volcán Atitláns eruptive sequences. There is an entire Mayan city below the water at 30 meters depth.

            As such it might be the only volcano on the planet that makes me a tad nervous. That being said, I would happily go live there. See the part of me writing above that we would know quite a bit in advance before anything went poop at such a large volcanic system.

        • 1600 years ago pacaya collapsed, which carl probably knows way more about. Since then it has formed mackenny volcano inside that crater, and as of this century the collapse scarp has been pretty much entirely refilled. Its not really as simple as this but the collapse is about 2 km wide so it has made a cone 2 km wide and about 1 km tall in 1500 years. That is about 1 km3 in 1500 years, with the real value being maybe closer to 2 km3 to include the obvious variables that I dont know. This is then roughly 1.3 million m3/year supply to form pacaya on average, which is very impressive for a subduction volcano and especially because a lot of it is in the last 100 years which changes the short term averages a lot, but compared to hawaii it is quite tiny still, kilauea gets 2 km3 of magma in a decade, and has erupted about 4 times that much since 1983.
          This really makes a lot of sense though if you actually compare the size of hawaii to literally any volcanoes on earth that are on land.

          I didnt include fuego in the numbers because I think calling it a flank vent is a little less certain than it probably is for pacaya.

          • Note the difference between supply and output and that 2 km3/decade is the supply to Kilauea and Mauna Loa.

            Hawaiian volcanoes are very good at keeping the conduit open though that usually doesnt prevent the summit from inflating and intruding the rifts. Geologic evidence indicates that Kilauea erupted from 800 to 1500AD continuously or nearly continuously from the summit by sending one lava flow after
            another down the flanks

        • Grimsvötn also haves a ridocolusly massive supply
          But most of its input never erupts
          Whats the yearly supply for Grimsvötn now again ?

          • Iceland as a whole gets around 1km3 per year. Grimsvotn and its rift system captures perhaps one third of that.

          • Grimsvötn haves a very large supply
            Of around 0,5km3 every year from passive spreading decompression and alot from the Mantle Plume combination.
            The hotspot supply may resemble Kilauea but its likley much lower in supply. There is 5 volcanoes under vatnajökull and 2 outside that shares of the large deep supply and there is the passive rifting that steals and decompresses alot of magma. There is 7 volcanoes in Iceland that feeds directly from central plume supply. Grimsvötn, Bardarbunga, Kverkfjöll and Askja are purest plume magmas in Iceland in compostions.
            And there is the Iceland Mantle Plume directly under there.
            Iceland is pulled apart 0.9 centimeter per year at Grimsvötn. When it spreads the magma decompresses that rises the system mainly via the mantleplume, but partially alot from MORB-processes.
            Grimsvötn is the heart of the Icelandic Hotspot.
            The Hotspot and Mar spreading both decompress and makes magma.
            Each of the vatnajökull volcanoes may have 100 s of km3 of liquid magma thats not eruptible.

            This tells how a 500 cubic kilometer deep chamber can form under the Grimsvötn volcanic system. Under an area of 200 km long 60 kilometers there is around 800 cubic kilometers of basaltic magma.
            Temperatures deep down is around 1190 C and its around 1510 C in partial melting zone in the astenosphere source, even deeper down.
            This magma reovair feeds hot open pathways and Grimsvötns much smaller upper magma chamber and Thordarhyrna, Kverkfjöll and partly the Dead Zone. When these upper magma chambers drains we gets the 10 to 20km3 events and calderas in the central volcanoes.

            Most of this magma can never erupt and is doomed to become gabbro.
            When Grimsvötn dies in millions of years there will be a giant gabbro batholith like thing there.
            The very large rifting events coud partialy tap from this upper parts of this magma resovair. But most of this cannot erupt.

  2. Some tremor, I think this is from the deep area near Pahala since it shows there and not much else shows on this instrument.

    This one shows the quake for reference



  3. Watching weather channel. How come they always say stuff like “at least 4 people were killed”? What? Are they disappointed? … ‘It could have been less, but AT LEAST this many expired.’

    It has been said that “Words have meanings” and that is what goes through my head every time I hear one of those shallow, vacuous twits use that expression.

    As for the storm front. Not much of anything energetic near here. But since weather systems are chaotic by nature, you always have to pay attention. As soon as you are dismissive of them, they get you. That’s part of how karma works.

    Note: Samuel Clemens noted his displeasure with the overuse of superlatives, stating; “When you catch an adjective, kill it.”
    I have violated this idea by the use of the “shallow, vacuous twits” comment… but I wanted to express a twit that is both shallow, and vacuous. Features that are easy to find in the “news” media. Note: Anyone wishing to turn my phrase against me should bear in mind that I am not a reporter, or a journalist. Yes, I can be considered an idiot, I am after all a member of the Homo Stultus species. (Homo Sapiens has been extinct for long time) And though I am male, and live in Florida, I am not the semi-famous “Floridaman,” though I could eventually become him. (I’ve got the gender and locale correct, and I am just about as stupid as everyone that lives around me, so I have potential. About the only thing in my favor is that I actively seek to NOT be Floridaman.)

    On the plus side, at least we don’t have the Guy from Ohio… but the 15 to 18 year history of Fort Walton Beach winding up with a random mutilated goat carcass being found on city property has me weirded out a bit.

    • Not sure what the problem with “At Least” is.

      From the example given, all that’s being said is “We know that four people have died, but we think that there are more”

      As far as I can tell “At least four people died” is just a useful way of compressing that sentence, and doesn’t really imply any sort of enthusiasm.

      Of course, I haven’t seen the offending clip, there may have been something implied by the tone used, but the words are essentially innocent.

      • “At least” is just sloppy writing. The alternative meaning is just as valid as the innocuous one.

        My interpretation, though it assume’s the worst in journalistic opinion, fits quite well with the “she can tell you about the plane crash, with a gleam in her eye” line from “Dirty Laundry”

          • Not to mention, I get to sit and watch Vodka Bob on local news every evening denigrating anyone who gets in trouble with the Law. That’s why I found his DUI arrest so gut wrenchingly hilarious.

        • in my experience, journalists are normally professional and the science journalists I have met were always meticulous in asking about the uncertainties of the research. But when it comes to presenting, newspapers and other outlets too often ignore those journalists and go for the sensationalist approach. I regret that but understand that it is what sells. It can be harmless, though doesn’t do much to educate people. It can be harmful if it shows presenters (for instance) outside in hurricanes, which encourages people to do the same. And it is very harmful if research on topics that affect our future is misrepresented, ranging from climate change to food production and antibiotics. I see that in both sides where both the ‘green’ and (for want of a better word) the ‘red’ only present the part of the research that supports their opinions. Am I angry? No. Concerned? Yes, a lot. Problems become unsolvable if they are either denied (‘climate isn’t changing’ is one that denies the blindingly obvious) or overstated (‘More hurricanes will destroy’ – research does not show that hurricanes are increasing in number). If you find that sea level will rise by 70cm this century, that is something we can make adjustments for. If you report that ‘it will go up by several meters’ (the sensationalist news) people will just walk away and nothing will be done.

  4. Oh, and for any passing news twits, if you use Carl’s numbers from above to work out a recurrence interval so you can claim something is “overdue”, then you are a bona-fide idiot. Volcanoes don’t work like that and DO NOT follow schedules.

  5. Nice and to the point article.

    A small correction “Both Pacaya, Acatenango and Fuego has done that, ” should be “have done that”

  6. I was doing some reading (including Volcano Cafe) on the recent quakes between Hualalai and Mauna Loa, and what they were not ” flexural earthquakes”? No soccer today due to the storms. Recent set of quakes since the 5.3.
    2019-04-14 20:56:53 1.9 14.4
    2019-04-14 15:26:20 2.8 7.2
    2019-04-14 09:50:28 2.2 19.5
    2019-04-14 04:46:29 2.7 16.8
    2019-04-14 04:28:59 2.3 12.6
    2019-04-14 04:12:57 2.3 13.4
    2019-04-14 03:20:36 3 14.5
    2019-04-14 03:09:02 5.3 13.2

    Now while I am reading these papers (linked below) I realized that we may have a couple of flexural earthquakes on the current map, 2019-04-03 15:02:44 2.2 37 and 2019-04-09 05:26:42 2.4 40.2.

    The biggest problem is there is no signature for these quakes…
    “Most earthquakes felt beyond the Island of Hawai‘i are presumed flexural earthquakes based on their estimated locations. Some historical examples include the magnitude-6.8 Lāna‘i earthquake on February 19, 1871; magnitude-6.8 Maui earthquake on January 22, 1938; magnitude-5.2 O‘ahu earthquake on June 28, 1948; magnitude-6.2 Honomū earthquake on April 26, 1973; and magnitude-6.7 Kīholo Bay and 6.1 Māhukona earthquakes on October 15, 2006. ” From Volcano watch.

    “[3] The depths of the two major events place them in the
    mantle. Mantle earthquakes around Hawaii have been
    attributed to flexural stresses induced by volcanic loading
    [e.g., Pritchard et al., 2007], although magma-related
    processes also contribute, especially in shallower regions
    [Klein et al., 1987; Wolfe et al., 2003].” from “Flexural stresses beneath Hawaii: Implications for the October 15,
    2006, earthquakes and magma ascent”

    “Why do some Hawaii earthquakes occur so far offshore?”

    Flexural stresses beneath Hawaii: Implications for the October 15,
    2006, earthquakes and magma ascent

    Enjoy, I started with the Volcano watch article which is in layman’s terms. The linked PDF from GEOPHYSICAL RESEARCH LETTERS, VOL. 34, L23305, doi:10.1029/2007GL031305, 2007, does have some information that many of us will “get” but I am not getting it all.

    Good night

    • It is probably not deep enough to be in the mantle. HVO has said “The location and depth of this event suggest it is likely related to flexure or settling of the crust beneath the weight of the island.” It seems close to the bottom of the lava pile lying on the crust, but it is not obvious to me whether to attribute it to the crust or the lava pile! The quake introduced a significant tilt change at Mauna Loa, so settling is an option.

    • Flexural quakes is what I referred to as unequal sinking or subsidence of the island in the last post and it seems that it is the most likely option particularly if HVO backs it up. That depth is in the crust not sure though if at the base of the volcanic pile or in the underlying oceanic crust or in the contact between the two.

      • I agree. But the sinking can shifts cause in either the crust (the usual) or the lava pile on top. I was wondering because the quake sequence follows in a linear structure pointing away from Mauna Loa, and because the event caused a notable shift in the tilt at the Mauna Loa summit.

    • Meanwhile, in other Hawaii news, JOKA, the GPS of Heiheiahulu (Kilaueas MERZ/LERZ) is aiming for the sky, I am guessing it will slow back down cause at that rate things would get fiery this summer again.

      • Interesting that this is the first place that the inflation spike shows up. I was expecting something at the summit and also still not for another few months at least. Maybe the hypothesis of the deep rift feeding the area has some solid grounding now. The summit does seem to be rising based on the tilt signal but not by a huge amount, not enough to show on the GPS, and it has done this a lot before a DI event happens so its too early to say right now.

        It would be good to know if something similar happened after 1975, the only information regarding ground deformation before the 1977 eruption that I could find was more about the summit. It also talked about the formation of a dike in the area but that already exists now so things could be quite different.

        • There is one GPS at the summit which has been inflating (UWEV) northwest of Halema’uma’u, not sure what it means though. Appart from that, there has been no recovery of the summit while the ERZ has been inflating but it is Kilauea after all… it wont follow the patterns of other volcanoes not even its own patterns a century later.

          • All the other GPS between UWEV and the east rift are on moving south flank blocks, so this is probably offsetting any inflation signal but UWEV is on a more stable block and could well be showing the true signal in which case there has been minor inflation for some time. The deep chamber is probably below a lot of the movement so the magma could be getting that far but no further up into the shallow system. The south flank moving probably takes up some of that but I think more of it is going into the recent spike at JOKA station. In the past I used the volume of space created in the 2018 quake as a guide to how long it might take to repressurise but it actually looks like nearly no large scale deflation occurred on the east rift itself, with all the rift space being refilled by the magma leaving the summit and still it was enough to feed fissure 8, the volume of the summit caldera being about the same as the amount of lava erupted.
            The east rift after last year is probably a continuous open system with uninterrupted hydraulic connection to the summit, so it could simply be that it is easier to get the already pressurised part of the east rift down from puu oo to do things instead of filling the summit system against gravity. The 2018 dike was probably substantially larger in width and total depth than the average dike at kilauea, probably quite deep at the bottom (6 km?) but also shallow in others. There were probably places the magma came within 1 km of the surface without actually erupting all through the distance between fissure 8 and puu oo, at the highway it is obvious but likely similar things happened elsewhere including under heiheiahulu and maybe under the 1977 vents, as well as just uprift of puu oo where the tiny fissure opened at the start.


        This might also be of interest, particularly the part relevant to the start of puu oo and the high rates of deformation before it. 40 cm a year between 1976 and 1983, a total of about 3 meters before it stopped and puu oo started shortly after. That is compared to about 50 cm from May last year to now and which stopped some time ago. Maybe despite obvious immediate similarity the two situations are not quite so comparable after all, which really just means we have to wait and see what comes from this and that my prediction from a few weeks ago is already wrong :I

        • Situations are comparable to a certain point but never going to be the exact same, looking at the trend of 1950 till now, the deformation since the eruption and what we know about its prehistoric eruptions I think the next decades will be dominated by ERZ activity and that next Kilauea eruption is most likely to start from the rift near JOKA unless some massive intrusion affects the summit.

          • Well it does seem like a ‘massive intrusion towards the summit’ could be on its way but it might still be a few years away from actually erupting yet so most likely you are right with the next eruption happening near pu’u o’o or downrift of it a bit. The recent uplift at JOKA is definitely something to watch.

          • I would recommend a bit of patience. There were GPS signals all over the island just before and after the quake. Things seem to be settling back to where they were before (even the Mauna Loa tilt). Give it a week to see whether the signal continues or goes back to where it was before. It is still too early to decide.

          • I would say Kilauea has switched from the eruptive to intrusive state (possibly including failed or near-eruptions, the last was at Halema’uma’u in 1996) for now like in the 1940s and to an extent in the 70s and early 80s. I would imagine the next rift zone eruption will be similar to 1977.

          • Yes, it is a bit too early to tell if this is just a spike a short inflation or the start of an increased influx of magma into that area. But in any case I really think JOKA is dangerous as it is very close uprift from Leilani and other populated areas.

          • An eruption at JOKA would actually flow north to pahoa exactly the same as the 2014 flows that invaded the area in question, so yes it is a rather scary location, not to mention the risk of another downrift intrusion into leilani again. That probably wont happen but an eruption east of heiheiahulu at any point could send lava flows that way anyway, and could be a persistent threat if it becomes long lived.

            Most likely though it will be a pretty small eruption, like 1977, which was in the same area. It might not give much obvious warning though maybe a few hours before a vent appears. If I was one of the people living near the highway 130 cracks though I would be watching this as that could still be potentially reactivated. If this movement is real (it seems too far from the quake near hualalai to be related in any way especially on a GPS signal over several days) then if it carries on at that angle it will probably go for a few weeks before erupting based off similar episodes at puu oo, but really unless this stops soon we are looking at an eruption this year. Again still too early to call but it is definitely a change in any case and I think we will get some info from HVO on it if it keeps up much longer.

          • That was what I was most worried about, the upper fissures of Leilani getting reactivated, even in a small eruption it would deal a lot of damage again. Worse would be a satellitic shield though…

          • Also to mjf, I really dont think kilauea will go into a period of intrusive activity, as I said before there doesnt seem to be any large scale deformation along the east rift apart from the actual quake, which would have occurred anyway at some point. The quake is a permanent change, and not caused by loss of magma, so if you ignore the drop it caused you find most of the east rift has already recovered any deflation that resulted from magma last year. Also important is that unlike 1975, the effects of last year became physical, nearly all of the magma that erupted in leilani was replaced by physical collapse space inside the new caldera, in 1975 all the space was a diffuse band of low stress through most of the east rift. 1975 was also not triggered by an intrusion, it just happened on its own.
            If you factor in the difference between the volume of lava erupted in leilani and the volume of the caldera it is actually a rather minor amount, only about 0.2 km3, which is about a year of supply. We are getting close to the 1 year anniversary of when things all started so maybe it shouldn’t be surprising things seem to be happening now.

            Last year was probably an event that was in the making for decades, all the activity since 1960 had filled the east rift with magma most of the way already and then all it took was a little extra push to go flood basalt mode.

          • Well I meant briefly and maybe not actual discreet intrusions, but certainly input of magma at a more significant rate than during most of the Pu’u O’o era without being able to sustain an eruption at low rates again like at Pu’u O’o. In other words, the system will chug and not start until it is able to tick-over and be idle at a stable level, even longer to get moving and get of 1st gear, and of course decades until it cruises down the magma motorway and then gets over-confident and ends up with a speeding ticket…. Ok, maybe not the last part!

          • The last part is what happened last year…

            Having a bit more data now, it does look like this isnt a massive spike we are seeing at JOKA station, but all the signs are consistent with magma accumulating slightly south of that area at some depth, over 1 km. The fact a spike of any sort happened is evident to the potential instability of this area. At the moment the deformation is not high enough to trigger alerts at HVO but last year was preceded by 3 years of slow inflation that everyone thought was completely harmless until it suddenly wasnt. This deformation on the east rift is about 30 cm in the 8 months since the inflation became obvious, which is not a completely trivial amount, actually the same as the summit inflated between 2015 and 2018, and comparable to what it took to make new vents at puu oo in the past. If it keeps going up at this rate it will not be on green alert for much longer thats for sure.

            The one thing that could be a real issue is HVOs assumption that any eruption outside the caldera will be really obvious well before it happens, that was true in the past for big eruptions like 1960 but after puu oo formed with its long and robust magma conduit it is not likely the case anymore. Effectively puu oo was the new end of the mauna ulu conduit after 1975 extended it to the middle rift, and JOKA is at the end of the puu oo conduit that extended to about highway 130 after last year. An eruption here is probably just as likely as a summit eruption, and would be just as sudden, probably within a day from the first signal to an eruption.

            I might as well right a post or two on this stuff now, been putting it off too long…

          • Actually to properly do this it will definitely be a minimum of 2 posts and probably at least a week before the first one is done.

            I want to wait and see what happens with the middle east rift GPS signal before I make a conclusion as what happens there changes the outcome of this quite a lot as you can imagine…

  7. While it is pretty obvious that hualalai isnt going to erupt soon, I still did some research and measurements on the 3 most recent eruptive episodes.

    The first eruptive episode (white) that is clear on the USGS 2005 map happened in around 900-1000 AD, and was quite large, it covered a total area of around 100 km2 mostly on the southeast side towards mauna loa, approximating that the average thickness of the flow is 10 meters gives a rough volume of 1 km3, though without knowing the thickness of the flows this number might be significantly over or underestimating. Despite the size of these eruptions none of the flows reached the ocean.

    The next eruption (grey) happened about 200 years later. This eruption was also on the east side of the volcano, and only occurred from one vent where a big cone called wahapele formed, which seems to have later been consumed by a collapse but the remaining parts of the cone are still about 50 meters tall. These flows reached the ocean. The total area of this flow is about 39 km2, but the flow near the vent seems to be substantial, maybe as much as 30-50 meters around the upper part of the flow field, down to about 5 meters at the ocean. The overall average thickness of the flow is probably somewhere in the area of 20 meters, maybe slightly higher, which would give the eruption a total volume around 0.8 km3. This eruption probably lasted for at least a year and likely exhibited lava lake overflows and lava tubes similar to pu’u o’o, though most of the lava is a’a so it was probably not as stable.

    The most recent eruptive episode (dark grey) was around 1800, and probably lasted at least 5 years. The initial eruptions made the kapulehu flow, which flowed north to the ocean and covers an area of about 35 km3, the flows seem to be about 5 meters thick on average, as typical for a’a flows of that type. This, along with another vent that was just downslope, gives the flow a volume of about 0.2 km3. The second eruption, maybe a few years later, was the huehue lava flow, that is underneath kona airport. This flow is covering an area of about 17 km2, but it is very thick, the on land part is maybe even as high as 30 meters in places, and the flow probably extended the shore quite a lot. The total flow is probably about 0.5 km3, so the total volume of the most recent eruptive episode was somewhere around 0.7 km3, maybe slightly higher.

    The wahapele eruption happened 200 years after the previous eruptions, but it was not until 600-700 years later that more eruptions happened. The most recent eruptions were similar to the eruptions about 1100 years ago, but the wahapele eruption was a single vent eruption of probably long duration. This is a small sample size but it is possible that the next eruption on hualalai might be an eruption from a single vent that makes a sizable cone, rather than a rifting fissure that covers large areas of the volcano. It is also possible that it might be generally quite soon, the chance of hualalai erupting this century is probably fairly high.

    • As we have pointed out to Jesper a few times.
      Could we have the Hawaiian comments centred in one comment thread in articles that is not about hawaiian volcanoes?

      I know that you two love Hawaii, and we do not mind.

      That being said, if either of you two (or anyone else), would like to put together an article or five about Hawaiian volcanism we would post it. I think that would be a much better use of your writing skills. 🙂

      • I agree, there’s more than enough content in your collective comments to publish a book on the subject let alone several articles.

        We’re not trying to stifle your enthusiasm, please don’t think that. We run a blog on such a broad spectrum of topics, at times, as well as the ‘hot volcanic action’. As part of this, we do have to maintain a semblance of order to the comments as it can make finding info hard for our occasional visitors, not to mention us dragons behind the scenes trying to moderate content. I did float the idea of a Hawaiian bar (separate area) in my mind, but I don’t think this is a good idea on reflection. I’d rather our visitors see your comments than miss them in some cocktail stained area of the blog.

        So, to throw the gauntlet of challenge at you runny lava lot. Collate your thoughts, data and hypotheses and put them in article form. Send it over to the VC email address and the dragons will take a look and publish it on the blog. Don’t be afraid to ask and don’t be afraid to share your thoughts and knowledge. We only have one rule, that’s be nice, everything else is fair game.

        • And also, we do need good content on subjects that our readers enjoy, and we do know that many readers like Hawaiian volcanism.

          • I don’t. The Coffee is over priced. 😛

            Grandkid brought me back some of those tourist grade Kona coffee packets. I’ll stick with Columbian. (I’m a retired CPO, I want my coffee simple, and strong.)

          • O no. A coffee barbarian. Kona coffee is excellent if you get the right stuff (many packets are heavily diluted with non-Kona) and grind it correctly. There are better coffees in the world, but not so many.

          • Colombian coffee is quite good Albert, I do not know about coffee from Columbia though 🙂

          • I guess the burn factor depends if the coffee comes from the state named Columbia, or from District of Columbia, the burn factor is far higher from the latter. 🙂

    • Seems the valiant Pompiers managed to mitigate the spread and save the bell towers. Also, a LOT of statues were recently removed for renovation, and one of the famous stained-glass windows appears unscathed…

      • But a disaster nonetheless. The ancient cathedrals have often been damaged and rebuild, but they are a major part of our cultural heritage. I hope it will be rebuild the way it was, and not as a modern re-invention.

          • Yes, I know you would have liked to fill the whole place with lava. I prefer to keep historic monuments at lower temperatures – and am still in shock by the damage. I feel for Paris. It won’t demolished: it is too important for the city. The French will find a way to fix it.

        • I have a degree in history and i seriously cried yesterday when I saw the news…Notre Dame de Paris is in every sense a world heritage site, the cultural loss will be immeasurable. I am a very rational person, every part of my thinking is based on science and facts, so I really don’t care personally about the clerical loss, BUT it is not just a church, it is significant in so many ways there a hundreds of scientific books written about it. This is a disaster of epic proportions, even though a lot of the structure and artworks have been saved.

  8. Thanks for the article Carl. Once again my Guatemala-loving wife was pleased to see a mention for her favourite country! That was a good, informative read.

  9. Grrr… rough day. On top of that, my wife found out a couple if weeks ago that I make a pretty mean Low Country boil. So, guess who’s elected to make supper…

    Fortunately, I can knock out a boil in 30 minutes.

    • Aka:”Frogmore stew”. (No frogs)

      My variation is almost a gumbo. 3 more ingredients and it would be true gumbo.

      No, I have no cajun lineage. My exposure to it comes from my Aunt who used to be head nurse at Avondale Shipyards. R.I.P. (Her, not the Shipyards, I think they are still running… though bought out.)

      Her death was brought on by the World’s leading cause; Complications with growing old. I’m still a bit pissed off by it. Her sister was no help in the matter at all and was effectively a vulture preying upon her. (Yeah, that’s another Aunt… but not all relatives are worth a crap. Karma got her… big time.) ← No details, but the events as they played out would have made the hair on the back of Carl’s head stand on end… if he had any. He was weirded out enough when I inadvertently correctly answered a weekly poll question about 5 minutes before he posted it. (no magic, I was just being a smart-ass and happened to get it right. It was a volcanic question and I threw out a volcanic related answer… Actually, the odds were stacked in my favor… sort of.)

      No, no discussion from the back-channel, I had absolutely no knowledge of the question before-hand. Just pure dumb luck. I don’t even think I was a dragon at the time.

      • OT ok, just a fast list of just the spices You use…. stay safe with all the bad winds.. Best!motsfo

    • As for the rough day, I encountered a living, breathing human example of vapidity.

      No details, this was a public figure and I don’t want to potentially mess with the integrity of the contract, but I’ve had many students who were easier to reach. Even Vampire Kid. (No, Vampire Kid didn’t make it through training. He got processed out due to legal issues unrelated to his time in service. Fastest Academic Review Board I ever convened. Took all of 5 minutes.)

      Base Legal → “He ain’t coming back”
      Me → “Aye.”

      Filled out the paperwork, grabbed three instructors who wandered by, “Sign this, here’s why” → Done.

      Everything else was up to Base Admin.

      So you see, the system can work in an expeditious manner if the situation requires it. My running problem with him was keeping him awake in class. By regulation, We were not allowed to throw things at students. As for the “Vampire” notation.. that was his own invention. The “kid” bit is my own nomenclature, nothing in his student record indicated that he was Capra aegagrus hircus.
      Homo Sapiens Stultus, yes. (just like the rest of us, even Floridaman™)


      • Speaking of Floridaman™, his most recent exploit (according to the news today) is that he was hit by his own boat after falling overboard. Somewhere near Destin.

        And, I have no further information on the sunbathing guy that was run over on the beach in South Walton county by an Enforcement Officer perusing an unleashed dog that apparently had no permit. As far as I know, no critical injuries. This was within the last two years or so. No, he was not a manifestation of Floridaman™… but the Officer was. (The lack of critical injuries is probably due to the nature of our “sugar sand.” Though I DO NOT recommend trying out that scenario. If he sand is even partially damp, it has a resiliency near that of concrete.)

        Gurgle’s “bad latin” construct: ecce viri stulti

        Hey, at least I admit it. (Honestly, I don’t know if it’s bad or not, I’m just going by reputation.) After all, I am one as well.

      • Underground coal fire melted the bedrock and it came out as paralava
        Paralava, or parabasalt, is a pyrometamorphic rock that is vesicular, aphanitic, often pahoehoe ropy and with clear flow structures, formed from melting of sillica poor shales,

    • Very interesting, I agree with carl though, if it was a volcano it would be way bigger and we would have seen it coming. It could even be an electrical fault,

      • I pondered the part of the electrical fault. But as was the case with Tor Zawar, there are just no underground cables in either Pakistan or India. I have though seen similar things like in the video above in videos from Centralia.
        Another viable option is that this is above a burried petrol or gas tank. But, coal seam is the best fit option.

        • It really looks like lava ( molten rock )
          Pahoehoe that is with the ropes
          Glassy and black and yellow orange hot in the bubbling spot
          1180 C
          Underground coal fire/ gas
          melting bedrocks
          I wonder what part of India this is

  10. On another US related note, it has now been three months since the last activity at Cleveland, the only US volcano to erupt this year so far. This means that for the first time in decades, there are no ongoing eruptions in US territory and there’s only Cleveland and Veniaminof on yellow alert, and none any higher. What a difference to last year with 5 volcanoes erupting, or 2014 when for a while there were 7 volcanoes on elevated alert levels.

    • One last point regarding end of eras (apparently a common theme at the moment), Bagana has also finished its 18-year eruption period. We’re down to 6 volcanoes with ongoing eruptions that started before 2000, and a further 6 that started between 2000 and 2013 (unbelievable Jeff). If we weren’t already, we are most definitely in a global volcanic lull now! Activity is at its lowest within my 21 years I think. 2010s= effusive (bright) and also dull (not playing down the 3 major volcanic tragedies in last 9 years). I suspect the 2020s will= quietness quite quickly quits (they thought it was over…). (thought tsunami over now I promise!)

      *Jesperturtleramblingitis intensifies* 😀 😛
      (Bonus points for noticing the “round” references).

      • It is 19 years into this century, and there are still 6 volcanoes that are in continuous activity since before 2000. I think that is pretty impressive.

        It also appears I am not even slightly on the younger side here, I am 20 and I thought everyone else was way older… :>

      • 2010s also includes the two biggest explosive eruptions since 1991, and a total of about 6 km3 of lava if you count what slowly came out of pu’u o’o.

      • Would be interesting to know which volcanoes are on the list? These are the ones I can name out of memory.
        1. Fuego
        2. Stromboli
        3. Mount Erebus
        4. Gagxanul/Santiaguita

        • Carl! Mount Erebus is a volcano I wants to visit..
          Thats the only volcano in the world .. that haves a high viscosity lava lake about 100 times stiffer than Hawaiis lavas. That alkaline evolved phonolite lava lake is the strangest stuff I can ever imagine in my head.
          That lava lake behaves completely diffrent from the more fluid basalt lakes.
          Erebus is proof that even more viscous magmas can have lava lakes under correct conditions.
          Petrologicaly that volcano is intresting too.. with similar compostions as Kilimanjaro.
          Erebus is anorthoclase tephritic phonolite and phonolite with huge crystals
          Erebus magma rocks seems sometimes to have green light grey tone .. or even dark green – grey.
          Alkaline Continetal Rift volcanoes in the west antartic rift system.

          Visiting Erebus rim coud be dangerous as that stiff lava lake sometimes explodes with gas slug slow.
          I seen enromous phonolite spatter bombs thrown on the rim in old photographs etc… scary
          Despite being souch a cold cruel place, I wants to visit Erebus

        • Ongoing eruptions (defined here as a period of eruptive activity with no pauses of 3 months or longer) that started before 2015:

          1. Yasur (1774)
          2. Santa Maria (1922)
          3. Dukono (1933)
          4. Stromboli (1934)
          5. Erta Ale (1967)
          6. Erebus (1972)
          7. Fuego (2002)
          8. Nyiragongo (2002)
          9. Suwanosejima (2004)
          10. Popocatepetl (2005)
          11. Ibu (2008)
          12. Reventador (2008)
          13. Semeru (2014)
          14, Villarrica (2014)

        • Carl what will your next VC post be about? You writes very good
          If something intresting pops up in Carls head write it 🙂
          The Fuego article was excellent
          And first time I realises Fuego and Pacaya feeds from the same source

  11. just a suggestion: i for one, really enjoy Turtlebirdman and Jesper talk about the Hawaiian Volcanoes but it does get a little jumpy for my poor brain to follow all the talk so maybe we could make a separate thread, a J/T thread where all the Hawaiian discussion goes….. i’d hate to lose any of their points if they bury them in some ongoing thread that perhaps i don’t choose to read because the topic just doesn’t interest me and i’d miss their points…. All the Best!motsfo or maybe the dragons all ready have and no one is telling me about it, like Facebook stuff…………. and what else are You guys not telling me about…… 😉

    • I think that was discussed but isnt going to happen because a lot of viewers are interested. The idea is more to have everything in one comment, which I assume also generally applies to other topics too but those dont usually get as much discussion 🙂

      • The best way to do this is by putting them in a post! Turtle, you could easily combine a few of your comments into a post. No problem that they have been here already as comments: ask motsfo (little old ladies always know best..). You can send your musings to the VC email and we’ll upload and if needed edit.

      • It might take a while but I can probably do that, will probably have to see what happens with the JOKA station before making a final decision though.

        • I just wish we could see more of the readings from the GPS units that don’t display any data. Especially those east of JOKA.

          • They probably dont show any change since August or July last year, most of the 2018 dike is probably too cold and viscous to allow circulation through it like what is probably happening at puu oo and nearby. I think if the whole dike was still completely molten and hot then fissure 8 would have become the successor to puu oo and still be erupting now but it hasnt and inflation is happening uprift which means it isnt getting much past there. Actually there are multiple shields in the JOKA area and further up but none beyond there, the LERZ only does large and short lived eruptions that are like last year, while the rest probably more often does shields if an eruption becomes big.

          • Dikes don’t cool that fast. The dead zone in Iceland is still quite ductile based on the heat models I ran even though it’s been since 1773.

          • The dikes in that part of iceland can also be a lot deeper than in hawaii. The maximum depth of a dike in kilauea is about 3 km up until about where tge highway is, below that is the deep rift which is technically magma but it is crystal rich and full of olivine, and eruptions on the upper part of the rift down to about where the current activity is at JOKA all probably feed out of this for the most part. The area downrift and beyond is not so active and I imagine the magma becomes non eruptible quickly, only able to erupt again if a new dike comes along. Last year showed this well, and 1955 erupted magma intruded in 1924, the magma from fissure 17 could have also been from 1924.

            In the dead zone in iceland you get the option of dikes that are 70 km long and over 20 km deep, quite a massive size difference, as well as such eruptions being caused by an actual rift zone rather than unstable flanks and foundation like hawaii. This plate boundry also brings the mantle into the equation, as such big rifting events as 1783 or eldgja probably rift all the way to the mantle and that will introduce a potential new source of magma as well as a much larger source of heat, this is not an option in hawaii and that makes a massive difference. There is also the ocean and water table, which is quite high in puna (explaining the minor or not so minor plreatomagmatic activity of most eruptions there) while based on the recent studies by HVO it is likely there just isnt much of a water table at all above the dead zone, not enough to do active cooling.

          • Dikes stay hot for a long time. Centuries, certainly. However, eruptions don’t re-use them. An eruption starts by creating a new path. The Kilauea rift zone is probably nothing else than old dikes.. so anywhere the new dike goes it will need to push out old magma. I assume that existing paths are not re-used because it is harder to push through liquid than through rock. If you split rock, an open path forms. If you spit split liquid, nothing happens.

            Dikes don’t open to the mantle. That requires a conduit. Dikes form above the locking depth, which is typically a couple of kilometers.

          • Yay Albert!

            Additional to my comment, the majority of heat loss is lateral to the dike. I even ran one with an “overburden” of water to simulate faster vertical heat loss, the results were the same. Decades of a slowly thinning melt zone… even centuries.

          • For eruptions as violent and energetic as the skaftar fires i would doubt 3 km is enough to prevent immediate eruption, and again it is not the dike itself forcing into the mantle but rather the rift opening through tectonic activity at the same time, likely through increased pressure at the hotspot during a surge. Likely not all eruptions in that area involve a deep set connection, but likely very few eruptions there were ever as big as 1783. Descriptions of 1783 from as far away as the coast basically describe a line of plinian eruptions at least a few km long, and then a few days afterwards was a similar sized row of lava fountains that were at least a km high, over 2 in a few cases. The fountains were visible from north of vatnajokull…

            The previous eruption in the dead zone was the 1477 eruption from bardarbunga, and while it was mostly explosive it was not significantly more voluminous than holuhraun, and made no large lava flows. This eruption was probably a more shallow event as you describe, a dike that intruded more passively and encountered groundwater that caused explosive activity.

            1783 was probably one of the biggest rifting events in the holocene epoch, the scale of the event basically doesnt have a contender except for eldgja and thorsja, which were similarly much larger than anything else their parent volcanoes have done. Basically what I am saying is that assuming all eruptions start the same is rather problematic when it is quite obvious that not all eruptions are actually the same. Your model works for kilauea last year, and for holuhraun, and it also probably works for all other eruptions at bardarbunga or kilauea, or pretty much every other basaltic volcano. It doesnt work very well for eldgja or skaftar fires because those eruptions were very different from just scaled up fissure 8. Holuhraun and fissure 8 had very roughly about 1 km elevation difference from their parent volcanoes summit, laki is only a few hundred meters lower in elevation than holuhraun but was on a completely different scale that goes far beyond gravity driven eruptions.

            In the picture below is the relative comparisons of fissure 8 and holuhraun when they were at their most intense (based on lava fountain height) compared to skaftar fires lava fountains… The black line is 1 km more or less.

          • I respectfully disagree.. it is not based on similarity, it is based on physics. Below 5 km or so, you can’t get empty spaces in rock because the pressure is too high. Magma can push through in two ways: as a narrow conduit, or as a slightly inclined sill (typically inclined at the same angle as the surface). Splitting it open vertically is much harder because of the pressure gradient: any liquid in it would be squeezed up and out. When you see rifting, you see a surface (or near-surface) phenomenon. Deep down things behave differently.

            Bottom line: if you propose that the rift has a connection to the mantle, that has to go through one of the allowed pathways, either a separate conduit or a sill. Neither are present in Lurking’s dead zone, and so you would need to source it through Vatnajokull (for Laki).

          • It still takes a long time for rocks to flow into spaces though, if you made a cave 6 km deep it wouldnt collapse immediately. The rifting at tectonic plate boundaries is not the same as rifting in other locations, and the boundary in a place like iceland is probably full of magma, not eruptible magma but something akin to the deep rift on kilauea (which is also below 5 km deep for the most part). The base of the crust also tends to mimic tge topography on its surface but exaggerated, the base of the dead zone probably has long ridges and valleys going lengthwise that can connect with the surface in extreme events like 1783. This doesnt mean the magma actually came from there, but it does allow for a dike to exist below where that would normally be possible. The appearence and behavior of eruptions is still important in determining the characteristics of the system feeding it. Fissure 8 and holuhraun occurred at the end of dikes of similar length and were even about the same size, skafter fires dike was same length and grimsvotn and bardarbunga are very similar volcanoes, but skaftar fires was an enormous eruption that was 10 times bigger than holuhraun or fissure 8, and grimsvotn itself also erupted alongside the eruptions on its rift zone and continued to do so for years afterwards. For a long time this was cited as proof of connection (which is still true in a sense) but after watching kilauea and bardarbunga do almost identical eruptions in the last 5 years and what we now know about those volcanoes, summit eruptions definitely arent a characteristic of a volcano that is losing magma at all. 1783 was 300 meters lower above sea level than holuhraun, but bardarbunga is also 300 meters taller than grimsvotn, and in any case for that to make such a big difference I think is just straight up impossible, especially on that scale. 1783 wasnt even just a holuhraun sized eruption that lasted for 10 times as long, it actually lasted only about a month longer, but while holuhraun was 500 meters long, skaftar fires was over 20 km long… fissure 8 and holuhraun also stayed within 1000 m3/s, while eruption rates during the opening stage of the skaftar fires were above the maximum a basaltic eruption can be before even the 1200 C ultra hot mantle plume tholeiite will be fragmented into fine ash. At the very least it is obvious there is more to this than just extrapolation from recent observation. Quite possibly the only way to understand 1783 would be to observe a repeat…

          • if you made a cave 6 km deep it wouldnt collapse immediately. I should have been clearer. You can’t open a cave there. Yes, it would collapse immediately, just like taking a brick out of a stack collapses the stack. The brick above doesn’t spend a few minutes thinking about it before falling down. You can’t open spaces at that depth. You can have faults and rocks can move along the faults. But you can only open the fault by pushing something in at higher pressure. That works for a sill or conduit, but try to do the same thing with a dike and the pressure gradient squirts it out upwards – and closes the gap again. Look at it another way: magma always takes the path of least resistance. In a dike, that is never down. There is probably a way around it by moving extremely slowly but that is not what you are looking for.

            Why the enormous eruption rates of Laki? I believe this was due to the width of the dikes. Instead of meters, it was tens of meters.

          • Holes have been dug to greater depths and dont collapse, even at 14 km deep in the kola superdeep borehole. I understand what you mean but solid rock is more structurally sound than uncemented brick towers so it isnt exactly that clear. If through some means you made a void within the crust at 6 km depth it wouldnt instantly implode it would exist for a certain time (maybe still only seconds, but it could be days in something really strong).

            It also still makes little sense why a dike that big would not erupt earlier. If the pressure was lower along the rift than going up it makes sense to get a dike but not that big before most of it goes out before that, which results in an eruption getting smaller away from the origin. Eldgja was like this but skaftar was the opposite. This is also probably why you almost never get silicic eruptions from flank vents at any real distance from a volcano, the easier option is to go up. Even in apparent silicic fissure eruptions like mono/inyo craters or cordon caulle there is a magma body underneath the rifting area not a single source with radial vents.
            This isnt exactly comparable to laki but it still has some important similarities. Namely that you have to get magma to move sideways 50 km at 3 km deep but when the magma erupts it shoots the same distance in the air. 3 km would contain a normal dike but not something like that for so long.

            Looking at this from the other way too, the dead zone is an extension zone which will change the depth/pressure gradient. The depth of the (no cave pressure point) is almost certainly deeper there, and could intersect the lower crust and topography of the upper mantle and change thing up a lot more than the simple model suggests.

          • Kola reached about 12.5 km. It effectively created a conduit, filled with mud at 250 atmospheres. Not an empty cave/dike! The mud was needed, as they wrote, to maintain the integrity of the cave wall: it doubled as the coolant for the drill.

          • I should say that my use of the locking depth is not quite correct. This is the depth below which the rock is ductile enough to allow movement. Above it, the fault is locked and will move in earthquakes. It is not the same as the depth below which cavities can’t exist. Although there is a similarity, locking depths can be much deeper depending on the area.

          • Looking at this completely from a different angle, there is an interesting situation. The dead zone is being pulled apart sideways by the plate movement, and it is also more or less on top of a mantle plume, this would serve to pull the crust apart quite well from all depths. This in turn would lead to deeper potential magma activity as well as a higher melt fraction in the mantle underneath from the lower pressure (not the same as a magma chamber) which causes large intrusions that have potential to rift to the mantle and if they do it results in eruptions like 1783. Maybe this doesnt happen elsewhere because the rifting is older and magma has made a deep rift structure that is mostly olivine crystals but still molten enough to prevent really huge deep intrusions and associated flood basalts. Like at kilauea, large intrusions would not normally go through this stuff (actually probably never in iceland) and most magma movement would be on top of this deep rift. Here it would not get the sort of runup to erupt on the massive scale of 1783 (longer runup means longer degassing and more pressure) and it would also pick up olivine crystals. It also probably coold down slightly, enough to not remelt the crystals. Holuhraun had a lot of olivine which supports this.

            Eruptions southwest of bardarbunga are also not really that similar to skaftar fires or eldgja if you look closely. Most lava flows in veidivotn are relatively limited in area despite fissure length, often no bigger than holuhraun despite erupting out of a fissure that is 50 km long, only thorsja is truly huge out of all the lava flows and it extends well beyond the vent area, but it might have been more like holuhraun, a long duration gravity drain eruption that lasted for years at a huge rate, basically fissure 8 but 20 times bigger, erupting for about 8 years at that rate. This is a massive eruption, but it is not much to look at compared to the rapid and very vigorous eruptions to the southeast like skaftar fires and (probably) eldgja.

          • Lots of unknowns! But seeing that Eldgja, Thorsja and Laki all erupted in the same area but from three different volcanic systems on two different volcanic complexes, suggests the main factor in common is the dead zone itself, not the volcanoes that feed these mini flood basalts. What can we deduce? Not a lot, since we have too little data.. But what the dead zone has is (1) an active rift, rifting at about the same rate as Vatnajokull itself; and (2) no private magma supply (we know that because of the lack of doming in the area). Maybe I should write a ‘speculation alert’ post on this. As for the duration of the eruption: we used to think that Eldgja lasted 7 years – now we know it is 2, the same as Laki. I expect Thorsja was the same.

        • Skaftar Fires was very large ( 14 times Leilanis volume in 8 months )

          But now imagine the very largest fissures for CAMP and Siberian Traps
          The mind totaly boggles….
          ( more than 1500 Laki volumes per flood basalt sheet in some cases )
          ( But mostly many 100 s of laki volumes per flow )
          and many majority flows flows just 10s of Laki volumes per flood basalt flow )

          Old dykes from Mackenzie Large Igneous Province
          ( Mackenzie dike swarm beacuse the flood lavas are eroded away )
          These basalt dykes are in excess of 500 km long
          And one dykle in Karro Ferrar flood basalt province is over 1000 km long I think tracked compositionaly and geological mapping.
          Souch dykes are proof magmas can travel and intrude in enormous ammounts in the crusts. And are a clue thats sometimes very large LIP events reached terrfying propotions as eruptions. Siberian Traps is still very visible despite millions of years of erosion suggesting the LIP s volume may been much larger than previous estimates.

          Souch events where oversized versions of Skaftar Fires
          But unlike Laki these mammoth flood basalts where not drawn from central volcanoes and specialy not CAMP with tectonic factor over the superplume.

          At the largest cases the fissures coud have reached over 500 kilometers long
          As long as Iceland widest point or more
          Souch fissures can produce sheet Aa flows the size of the nordic countries in a few weeks to months mind boggles
          Souch things happens only during major superplume sourges.
          East Africa Superplume
          and Maybe Baikal may be places for future large FB basalts.
          Antarticas continetal plume driven rift coud also be a candidate
          But East Antartica is far less powerful than Africas Plume.

          Africa Superplume
          The Africa Superplume thats the most powerful mantle plume for the moment
          MAY be cooking up Deccan Traps sized event in the future once the plume head finds a weak spot between the Tanzania and Keanya Cratons as its already beginning to do.
          The whole East Africa is uplifted by this Mantle Plume
          But no mega uplift with 100 s of meters yet luckly over a large arera.
          But the whole East Africa is experiencing slow plume uplift and slow rifting.
          But its hidden beneath one of earths oldest and thickest parts.
          The plumes head parts have managed to get between the cratons softer orogenic belts and caused the rifting
          While the plume center is under Africas oldest and hardest parts

          Its an odd tought of some day lions, leopards and elephants and zebras running away from a 50 meters high glowing Aa wall that advances at good walking speed
          And Kilimanjaro and Mt Kenya becomes islands.. in a glowing lowland sea of fast moving lava… Kenyan peaks transformed into safe Islands in a glowing hellsea that slowly advances towards the easten sea coast.
          And sheet after sheet flows out ultimately burying the African peaks
          The crust depresses too under the enromous load.

          East Africa may become one huge igenous plateau IF African Superplume reallly decides to do a major flood basalt.

        • Laki was very likley an intrusion directly from the mantle plume and major rifting decompression
          As you say Its diffrent from the gravity driven caldera collpase magma chamber drainage eruptions at Kilauea and Bardarbunga

          Still Laki ( skaftar fires ) is acossiated with grimsvötns huge rift resorvairs

          • I disagree. Bardabunga piping magma out through Holuhraun is probably the best example of what happened for Eldga and Laki. For Eldga, Katla was the magma source, and for Laki, Grimsvotn/ Thordarhyna.

            After I get done with my psuedo storms I may dig out Heat3D and model the intrusion’s heat loss using only magma injected above the locking depth. Unless directed otherwise, I will use a depth of 2km just for the sake of argument.

            In my previous model runs for it, I assumed an open pipe to the mantle.

            (This will probably be tomorrow sometime.)

            As for my psuedo storms. They are real and bonafide storms, but the fronts seem to loose energy by the time they reach here. In this case, the sun went down and the front slipped out from under the upper level jet stream that tend to amplify the vertical instability. Mississippi caught hell, but here its just thunder and rain…and over in Mobile Alabama some very pissed off cops are on a manhunt (In the storm) for two guys that shot at them near Tillman’s Corner. { No body injured… yet }

            Heh… and per the news, one of them was described as looking like a “wife beater.” No, that’s not stated as why they were after them, he just looks like one. {A trashy, grungy looking appearance I assume} They had fled from a traffic stop and crashed, then took off into the woods. One person is in custody, noted as unrelated, just a random homeless guy that freaked out when he saw the cops coming and ran. A canine unit took him down.

            The physical description of the Caucasian male point towards him being on the slight of frame size. Basically a sort of short skinny kid around 26/27. The “wife-beater” reference probably means the style of white T Shirt he was wearing. Persoanlly, I tend to wear something similar, though quite loose since this is Florida and it gets obscenely hot here from time to time. I picked up the habit of this style of shirt from playing racquet ball during a yard period in Portsmouth. The gym was next door to our assigned barracks and we could go over and beat the crap out of each other for free. Decent entertainment and somewhat healthy.. though you do come out sweating profusely. At the time I had a wicked serve. I had even weighed down the tip of my racquet so I could impart more energy to the ball. In all likelihood, this was probably cheating. But we were just playing ourselves and everybody knew what my racquet had on it. One of our group was a small skinny ET who was quite nimble at getting across the court, making seemingly un-natural saves. Even with my increased fire-power he could nab the return. My 210 lb frame was just not fast enough to match him on speed alone. It took conniving on shot placement. Our particular group was responsible for updating the ships technical publication library to match new and deleted equipment. (Including any and all updates since they were published.) Essentially, we were librarians… blasting out Metallica while we worked in our little corner of the warehouse. (Other teams were working on getting the required supply-support stock in order) {all part of ILO – Integrated Logistics Operations}

          • By the way, if anyone happens to know the actual geothermal gradient for the Dead Zone, that will make the model run much more accurate. Previously I had used 50°C/km but have no reference data to justify that assumption.

            Additionally, I will be using 1100°C for the intrusion magma temperature since that is the closest to published values. (Unless someone has a better temperature and reference to go with it.)

          • If anyone has actually read my older points it is pretty obvious there is something inherently different between what happened in 1783 and a gravity driven eruption like holuhraun. The two are not analogous. In both holuhraun and kilaueas eruption last year the eruption rates were on average somewhere between 100 and 200 m3/s. The lava fountains at max height were also in both cases comparable (150 meters for fissure 8, 350 meters for holuhraun at max but usually about 150). The eruption rate at the start of holuhraun was higher, but fissure 8 was more consistently higher (1 km3 in 3 months vs 1.2 km3 in 6 months). If you look at eruptions on the other side of bardarbunga you get a much faster eruption but the volume is not much different. The last eruption southwest of bardarbunga was similar to holuhraun, somewhat slower though. The large eruption of the frambuni lava field north of bardarbunga and to the west of holuhraun was also in this size range. Even the massive tephra volume of 1477 only amounts to an erupted lava volume of about twice that of holuhraun.
            At both kilauea and bardarbunga there was also no summit eruption while tge flank eruption happened, in both cases it was theorised to happen based on prior events but these old observations are considered incorrect now. These are two volcanoes that are completely unrelated but the same thing happened. Actually at kilauea it was very obvious the magma had left the area entirely down to a decent depth, and even today most of it prefers to go down the east rift instead of the summit even with no open vent anywhere. The same may well have been true at bardarbunga in the months afted holuhraun.

            Contrast this to 1783. Even in the very beginibg there is a difference, the initial eruption was not effusive, it was like 2011 but on steroids, a line of plinian eruptions some 5-8 km long in the middle of the zone, an eruption style that would only typically happen at a central volcano. Even after this the eruptions were violent enough to spray the heavy basalt lava well over 1 km into the air along the entire fissure that was active at that time. Eruption rates were thousands of m3 every second. Even months afterwards the eruption rate was higher than holuhraun ever was. The eruption was a bit over a month longer than holuhraun, but 10 times more voluminous. In addition, grimsvotn central volcano erupted alongside the massive rifting event, and possibly so did thordarhyna, if these volcanoes were losing magma to the rift eruption they would not be doing this, let alone erupting at VEI 4 level. This means at the very least the rate of magma input was high enough to not only offset the rate of magma flow to the eruption at laki but also fuel another eruption at both the central volcanoes and also keep that eruption at grimsvotn going for 2 years after the skaftar fires eruptions stopped… The only way I can see this happening is either the eruption was fed out of a larger source than just grimsvotn itself or the sources describing the event are themselves wrong. Gravity eruptions by definition drain magma, if anything the opposite thing happened at grimsvotn in 1783.

            It is also very important for me to say that having higher melt fraction in the mantle under the dead zone is NOT the same as having a large crustal magma chamber there, nor is the idea of magma flowing under the crust. In fact melting under this area almost certainly occurs and would probably flow towards the central volcanoes, following the topography of the bottom of the crust. In ordinary cases the rifting event would be crustal, making an eruption that is roughly on the scale of holuhraun – 0.5-5 km3, but if the entire plate boundary ruptures along its full depth, as is uniquely possible at diverging plates, it would lead to far larger events where the already ‘soggy’ mantle underneath can participate without migrating to a central volcano first. Most of the magma still comes from a central volcano system with its own magma composition and identification but it isnt driven out of gravity and the weight of the central volcanoes caldera plug like a typical rift eruption would be. If the only thing driving the eruptions was the mass of the plug inside the central caldera, then laki would be smaller than holuhraun, and less intense.
            As can be seen from the relative comparison between kilauea and bardarbungas calderas that formed this decade, as well as the variable size of historical calderas at kilauea, the size of a flank eruption does not necessarily correlate to the caldera in a linear manor. The collapse made by holuhraun was 3 times the size of the one at kilauea and significantly larger than the volume of lava erupted on the surface, while at kilauea the eruption volume is larger than the caldera volume. I bring this up because in Alberts series on eldgja that was a key factor, but since then holuhraun has been studied better and kilauea had not erupted in leilani yet and now that it has it does throw some questions at the model.

          • I think that a deep voluminous more powerfull magma chamber under Grimsvotn would easily explain everything. A rifting event triggered by a shallow or deep intrusion would change the stress field probably causing a large influx of magma from this deep reservoir into whatever it found, shallow magma bodies, the deadzone. And then you get the Skaftar fires, I guess the main factors involved would be the pressure of the deep reservoir and the tensile stress accumulated in the deadzone.

          • That is more or less the same idea, and also pretty much what Carl’s idea is about the area too unless he has changed it.

            It is really just not plausible to me that laki was caused by exactly the same thing that caused holuhraun or leilani, because it was a completely different scale of eruption. Fissure 8 is 43 km from halemaumau, and holuhraun is about 60 km from bardarbunga (measured along the dikes). The most downrift vent of 1783 is 72 km from grimsvotn, and about 50 km from (but less alligned with) thordarhyna, and 700 meters lower in elevation than either of them. Fissure 8 is 900 meters lower than kilaueas summit, and holuhraun is over 1 km lower than bardarbungas summit.
            There is more than enough evidence now that grimsvotns caldera didnt form in 1783 too. large summit eruptions in basaltic volcanoes are the result of large amounts of magma moving into a system rather than being withdrawn. In kilauea it was previously considered the 1790 event was in a collapsing caldera, but that has since been shown to be very unlikely and was instead pretty much the opposite. Bardarbunga didnt have a summit eruption in 2014, and kilauea didnt have a summit eruption during the collapse last year. Grimsvotn though did have a pretty big and long lived eruption during and after 1783. This eruption was even one of its biggest in the last 1000 years even completely ignroring laki, being a large VEI 4 probably even a VEI 5 if the likely lava volume is added (it was not fast like 2011 though).
            Basically if all these eruptions were fed out of the same mechanism then they should be all roughly the same size. Two are in fact nearly the same size, but one is just 10 times bigger for some reason… It makes no sense unless laki was caused by a different mechanism that involves a more unusual set of circumstances.

        • Grimsvötn is one of the most mighty volcanoes on the planet when it comes to mantle magma supply ( alot is stolen by the rift ) Only Hawaii is higher output.
          Grimsvötn will become a pretty scary volcano in the future as Turtlebirdman suggest. It will evolve into a huge shield ”flood plateau” with numerous calderas inside

          Grimsvötn haves a very intresting future as the hotspot gets stronger and eastern spreading rift zone gets more established and dominant.

          Katlas future is intresting too it will get more magma and higher eruptive frequency as plume and spreading rift gets more focused ( the Iceland Plume supplys the most from depth ( ( ridge decompress melt ). Katla may become like Vatnajökull in the far future

          • I must comment about Veidivotn.

            The eruption was far bigger than you think, Turtlebirdman. It is just less spoken about,

            I have been there many times (in Veidivotn), and the fissure is much larger than Holuhran (also been there, during the eruption). Veidivotn size is more comparable to Laki and Eldja and Thjorahraun (which is just next to Veidivotn).

            Veidivotn was mostly explosive for just one reason. The efusive fissure opened under a large lake (what was back then possibly the largest lake in Iceland). Of course, this resulted in a powerful explosive eruption which created a ash layer which is the thickest of the past 1000 years in south Iceland.

            And the eruption was dual. The ash layer revealed that there was a first rhyolite eruption of pink ash from Torfajokull, followed by a large explosive balsatic eruption from Veidivotn. This was (even before 2014) interpreted as a dike from Bardarbunga that reached Torfajokull.

            Holuhraun just confirmed the theory that many dead zones eruptions are fed by central volcanoes from either side of the region (Katla, Bardarunga and Grimsvotn).

            Vatnaoldur, is an older eruption from another fissure also just next to Veidivotn, which was even more powerful.

            On the ground, Edlgja also feels larger than Laki. The lava field is very wide and partially buried by (the more recent) Laki lava field, but the fissure of Eldgja is far more impressive. In fact it’s an absolutely astonishing canyon, a minor version of the Grand Canyon in Iceland, except for the fact that this was a canyon opened by volcanic activity.

            I really recommend a visit to all these volcanic sites, to grasp a picture of their relative sizes!

          • ?w=700

            This is a photo I took at one of the several 1477 Veidivotn eruption sites. This one is actually inside Torfajokull caldera, where the rhyolite pink magma erupted. The ground is still fairly hot there, whilst in Laki or Edlgja is already cold.


            And this last picture is Edlgja graben

          • I wonder just it’s known how voluminous the Veidivotn eruption of 1477 was in terms of DRE and how much of this was actual lava flows?

          • The volume I was going by regarding 1477 was the fact most of it was tephra so the original lava volume is going to be somewhere around 1/3 of that. In this case it means the lava volume erupted is about 3 km3, which is more than holuhraun but not much more, and it is more than a lot of other eruptions in the area too.

            Some time ago I found a geological map of iceland that showed holocene lava in pink, and apart from thorsjahraun none of the eruptions in veidivotn sent lava beyond the fissure swarm, compared to laki and eldgja that individually cover a higher area than all of veidivotn minus thorsjahraun.

            Eldgja was bigger than 1783 but might have been less intense, unlike in 1783 where the explosive activity was on dry land inbetweenthe two glaciers during eldgja it was under myrdalsjokull which makes it uncertain what caused it whether it was phreatomagmatic or from high eruption rate. It also lasted longer, more than twice as long actually, but it is not twice the size of laki so necessarily it would have been less intense compared as a whole. The first two flows were flood lavas, especially the one closest to katla, but the distal eruption was probably quite a normal sized icelandic eruption that stayed on a scale that we have observed.
            Eldgja could have been partly caused by a crustal dike but the vents closest to katla probably fed directly out of its deep source and were on similar scales to 1783 while later vents were smaller and more gravity driven.

          • Can you give some sources for your information? You are stating some interesting facts but I don’t know where they came form so can’t judge them:

            Eldgja lasted twice as long as Laki

            There is more than enough evidence now that grimsvotns caldera didnt form in 1783

            Grimsvotn’s big and long lived eruption during and after 1783 was one of its biggest in the last 1000 years

            They may be correct but there must be uncertainties.

          • The part about eldgja lasting twice as long as laki is from you… 2 years vs 8 months.

            The other bit is because when you made the eldgja posts we had only observed holuhraun and it was still considered plausible that summit eruptions during large gravity driven flank eruptions are not a necessary part but can still occur. However kilauea last year had no magmatic summit eruption during the collapses either and it is in fact quite conclusive the opposite effect was happening. The overlook lake was overflowing and then disappeared in a week, then the caldera formed along with large scale deflation. If the overlook vent didnt exist before last years eruption it is likely there wouldnt have even been the ash explosions either, as HVO now believes those were made by magmatic gas from shallow magma under low pressure in the draining system (same with 1924). These events would fall under the definition of vulcanian eruptions as opposed to phreatic eruptions now. The 1790 event was considered similar up until about 10 years ago, but now is known to be a very different sort of eruption and not at all related to magma withdrawl, and was also not a caldera formation, the deep caldera that existed in 1790 was not formed in that year and more than likely existed for decades before that. Grimsvotns 1783 eruptions didnt happen in a collapsing caldera either because the sort of eruptions described are basically like all of its other historical eruptions that we know are not collapse events, and the composition of the 1783 magma and all the other magma up to now except some of 2011 is reflective of the volcano getting a large batch of magma at that time rather than losing any. Skaftar fires was fed by the same magma that feeds grimsvotn but it was not fed out of the central volcano by a shallow dike it was fed out of a deep massive surge that also recharged grimsvotn at the same time.

            I really cant say anything else because it is repetition, if you use the holuhraun model for 1783 you dont get an eruption as big in any way as what actually happened, there is no way to say it other than that. You cant make something a standard model where adding 10 km to the length of the dike makes the resulting eruption 10 times bigger than the other, the one that is 10 times bigger is going to require something that is about 10 times more powerful. It just doesnt work.

            As said below, the statistics for the 3 eruptions I have talked about.

            Leilani: 45 km dike from kilauea, height difference from kilauea, 700 meters. ~1 km3 volume (still under revision).

            Panaewa 600 AD: 40 km from mauna loa, 2.5 km lower than summit, 6 km3.

            Holuhraun: 60 km dike from bardarbunga, Height difference from bardarbunga, 900 meters.
            1.3 km3 volume.

            Veidivotn 1477: 80 km dike from bardarbunga, height difference 1200 meters,
            >4 km3 volume DRE.

            Frambuni lava field 1200-something, 50 km from bardarbunga, 500 meters lower, ~3 km3 volume.

            These all agree pretty well.

            Laki: 70 km dike from grimsvotn, height difference from grimsvotn, 800 meters.
            14 km3 volume…

            Eldgja: 80 km dike from katla, height difference of 700 meters.
            20 km3 volume…

            This just really doesnt work the way I see it… the volumes are way too high in comparison to the other eruptions. There is also the fact the lava fountains in 1783 are twice the height of the difference between the laki vents and grimsvotns summit is enough to give indication against the gravity theory.

          • Thanks. The length of the Eldgja eruption comes (I think) from the fact it is seen in two consecutive annual ice layers (but this is from some deep memory). If the eruption started in spring, it must have lasted over a year. If in summer, around a year. Both are longer than Laki. But eruptions taper off with time, and most of the volume comes in the first part. The larger volume of Eldgja may come from just a bit longer duration, but more likely it was more voluminous at the start. (‘more likely’ means that the other option is not excluded). BTW, Laki was not a dry eruption: each phase started with an eruption through a lake or wet ground, turning dry as as the water was disposed off. Just a detail for the discussions.

            We don’t know whether Grimsvotn had a caldera in 1783. You argue that it didn’t play a role. That is fine, but it is not evidence that there was or was not a caldera. What we can say is that if there was one, it would have been deeply ice covered in the colder climate of those days. The current caldera is too small to explain Laki. On the other hand, you can’t take out 15km3 of magma and not leave a hole. Where the hole is or was (deep, or shallow and large) remains unknown.

            Bardarbunga’s caldera has a volume of 50 km3. That volume has gone somewhere. It is more than enough to explain a decent fire eruption. The current caldera of Katla is smaller. But we do know that after Eldgja, the jokulhaups changed and came out in a different direction than before. This suggests there was a reworking of the structure of the summit. How deep a hole was left we can’t really know: volcanoes change quite quickly. By the way, you are no doubt aware of it but the height difference should be taken from the bottom of the caldera.

            The magma used in the eruption may well come from a deeper chamber (good chance, I think). But the head comes from the summit and any shallow chamber. The volume in that head should be comparable to the eruption (plus dike filling) volume, regardless of whether it is the same magma that erupts. You can inflate and pressurise your magma chamber prior to the eruption. That will push up the ground above and can make the caldera floor rise (a common occurrence in calderas). So the model still works.. You do need to confine the magma until the right time. That is easier with a deep chamber, I expect.

            What we need is a decent-sized eruption, to see how things work. Holuhraun was in fact quite large, and it was gravity fed (IMO). So was Leilani (with the possible additional action of an intermediate magma chamber). The large fires seem to have volumes of 5km3 and above. We are only a factor of 4-5 below that in our experiments. It is looking promising..

            ps. Isn’t it good that IMO stands for In My Opinion and Iceland Met Office.. It provides a sense of authority to my personal opinions.

          • I suspect that during the 1783 eruption the whole general area around vatnajokull deflated, with the resulting hole being the entire area sinking by some small amount. This would not require a caldera to form in a central volcano. As I said before it is very unlikely for a basaltic volcano to erupt at its summit when there is a large eruption draining out the system through a flank vent. This is the case of nearly every basaltic caldera event we have observed, the big eruptiobs that occur in basaltic calderas are a result of later rapid magma supply associated with decreased pressure. This is especially relevabt to kilauea which is why it will be interesting to watch in the coming decade but vatnajokull is very likely subject to this too particularly with the waning glacier and significant supply rate to grimsvotn in particular.

            My guess is that maybe in previous times bardarbunga did events like laki, where the majority of magma from the hotspot went through it, possibly until the start of the holocene, but after doing thorsja it has waned and now is in the 1-5 km3 range for fires eruptions of which we have just seen it do its latest. Grimsvotn might have been less active before the holocene, possibly a stratovolcano a bit like hekla but surrounded by ice, but then it did the big tephra eruptions in the early holocene and made its calderas, as well as a lava flow in the dead zone a few times from the mid holocene, but now it is pretty well established as the central hotspot volcano, it is icelands kilauea with bardarbunga being akin to mauna loa. Through this analogy, while mauna loa has an imposing figure and long history it is far outclassed by its small but hyperactive neighbor, same for bardarbunga and grimsvotn.
            As for katla it seems to be a very infrequent visitor to the dead zone, apart from eldgja it has only erupted there one other time in the holocene and it was a lot smaller than eldgja or laki and stayed close to the central volcano.

            In this idea, prior to 1783 the magma in the plume head all started flowing towards grimsvotn and to the rift zone, and the several hundred years since eldgja that was accumulated in stress gave way in a massive rift that opened into the mantle where all the magma that was going into grimsvitn was, and it probably gave way all between the laki vents and grimsvotn. The open rift rapidly filled with magma from wherever there was any, and we all know what happened afterwards.
            Not all of it was in one line, some probably erupted at thordarhyna which is not alligned with the 1783 fissure. The huge volume of magma erupted fresh and very hot, the high gas content and probably extremely high temperature (1300 C or more) allowing the lava to be sprayed over a km high, and even higher when it started and flow like a raging river. There was probably not a lake at the vents, there is a river but it is not big enough to do anything, a study relating to kilaueas 1790 eruption found only a deep lake would have any effect on an eruption, groundwater or shallow surface water would boil off rapidly, and kilaueas 1790 eruption was not as powerful as laki. The entire magma system of grimsvotn was recharged with magma, enough that grimsvotn erupted for over 2 years after the rest of the fissures stopped, which is a long time for an eruption in iceland that isnt a shield. It took until 2011 for magma that wasnt involved in 1783 to start erupting.

            This is how I think things happened.

  12. On the topic of magmatic input in large volcanic systems, there have to be more systems out there than just Atitlan that have not been venting in geologically recent times.

    The first that comes to my mind is Ata, which is the southern, less well-known sister of the Aira caldera. I’ve always thought this to be a much higher risk volcanic system than people realize. Then you also of course have Aso, which is the biggest of all the Japanese Kyushu systems, and hasn’t done much of anything significant for some 100k years despite having a still active system.

    On a side note, Aira itself has a vent (Sakurajima), but that vent is not adequately relieving pressure-build up at the rate of influx. The big Aira caldera has been notably inflating despite Sakurajima’s frequent eruptions.

    • Well, yeah, Axial seamount had a flood basalt thing going on for a while there a few years ago. It was coincidental with the “warm blob” off the west coast of the US but the specific heat calculations didn’t work. It may have had some influence in it’s formation, but I can’t prove it. I think Albert took a stab at it and said “no.” {Albert has a better physics noggin than I}

  13. ot,,, OK, Albert, i’ve been tootling around the site and found the flux goes xray readout… question: Does it always jump up like that or is it due to changes in the location of the satellites? Thanks, Best!motsfo

    • I don’t know! You can see it happened for both channels on GOES 15, and you can see that the signal became much less noisy afterwards. So the signal was just plotted on a different scale. It is a logarithmic scale, by the way, so every step is a factor of 10. Plot it on the next step and the noise appears ten times less. The background is set by an electronic offset so it is adjustable. The only thing it is after is the flares, and as I write this there has just been one. It would have looked ten times stronger had the scale not been changed.

      The sun is very quiet at the moment. There was some comment about a strong sunspot a few days ago. In fact that was only a B-class, so quite small. (It was actually a left-over from a month before when that spot was stronger.)

      GOES looks for early warnings on flares heading our way. This flare is minute, and going off to the side so no problem.

Leave a Reply