Big basalt blasts III. Over the world.

2 posts ago I started talking about the pinnacle of basaltic explosivity, 2 posts later there is no mention of anything bigger than the Tarawera 1886 eruption. The eruption of Tarawera was a relatively common scenario of a dyke intruding below a lake, sure, the one responsible was a giant volcanic system of the Taupo Rift, but you are still in all your right if you are thinking I was being overly exaggerated!

If that is the case, do not worry, we have just looked into 3 cases of the last 120 years when we can actually look much further back, centuries and millennia back. So where could we find those big basalt blasts? The 3 volcanoes so far studied are Fernandina, Miyakejima and Taal, each one of them is recurrent, and has performed depressurization eruptions at least twice.

We need first a way to drain the volcano laterally, large enough flank eruptions, or rifts which can swallow a vast amount of magma. Next what we need is water to run into the magma reservoir when pressure drops low enough. Miyakejima taught us that the danger lies in water reaching near the surface. No matter how much water is dancing around the reservoir if it can’t be released into steam at the surface the big basalt blast is inhibited. When released into steam, a shallow explosion can depressurize the deeper superheated water, that’s when it uncorks and all hell breaks loose. But water wasn’t initially at the surface in Miyakejima. As its caldera deepened, water started to run into it and it was then that conditions turned from dry to wet.

There are many volcanoes that can erupt in this way, in fact I will end with a list of all volcanoes I believe have potential to produce them, but some deserve special mention. One stands above them all, as the most dangerous and the best studied. I feel some people may have been expecting Masaya in this article, here it comes.


This year Nik Wallenda had the unique experience of walking above a lava lake, walking on a tightrope! And he wouldn’t have worn any harness hadn’t he been insisted on wearing it! I wonder if he was aware about the history of Masaya, it probably wouldn’t have mattered to him, for someone determined to defy the forces of gravity and volcanoes some past explosive eruptions are the smallest issue.

Masaya looks like a small shield volcano. A bubbling fiery cauldron inside a pit crater complex opens at the top of a gently sloping cone covered in lava flows, might give the wrong impression of being gentle. But if we look at its eruptive history, what we find are three gigantic phreatoplinian/plinian events. These are the San Antonio Tephra with a volume of 14 km3 6000 years ago, the Masaya Triple Layer with 3.4 km3 2100 years ago and the brutal Masaya Tuff, 10 km3, at 1900 years.

Now, I of course think these eruptions were triggered by depressurisation of the magma reservoir (a BBB), otherwise I would not be talking about them… But I should first give some evidence. To begin with, Masaya is located in a rift, the Managua Graben, with its enormous caldera close to the water table. The proposed model predicts this is the ideal location for them to happen. Second, the Masaya of the last few centuries is a gentle small-erupting volcano that wouldn’t seem able to produce such massive eruptions unless there was a really good and extraordinary reason, which the BBB model gives. Third, the eruptions show strong water involvement. As an example, people escaping the Masaya Triple Layer left deep footprints into wet soft ash. Last, the Masaya Tuff was an eruption very similar to that of Taal in 1911, therefore with the same style of mechanism?

Human footprints on the mud-like Triple Masaya Layer

Taal obliterated an area of about 150 km2 with a base surge in 1911 resulting in the death of 1300 people. So what if I told you that the Masaya Tuff was the same kind of eruption but 50 times bigger in volume? The entire deposit shows the characteristics of an exaggerated base surge (cross-bedding, accretionary lapilli, indurated…). It extended 20-30 km in all directions. The base of the tuff is abundant in plant molds, this is probably all the vegetation in the area being flattened to the ground. And how long did it take? It looks like one big explosion, which must may have been over is less than an hour. This is frightening when you consider that about 2 million people live in this area, including entire Managua, the capital of Nicaragua. The city would be blasted by a wave of mud, rocks and maybe acid, reducing it to mud-plastered ruins. Now you know what I was thinking when 2 posts ago I mentioned the pinnacle of basaltic explosivity. Is there a more destructive eruption that a basaltic volcano can come up with?

Managua has a serious volcanic problem, Masaya is just half of it. The other half is Apoyeque, a silicic volcano that can very much throw a VEI6 eruption near the edge of the city. Managua is dangerously located in the very center of a graben/rift. The rift zone of Masaya runs to its east and dives under Lake Managua, while the rift zone of Apoyeque follows the Nejapa Miraflores fault, a chain of explosion craters to the west of the city. No more building on rifts please!

The Masaya Tuff was followed by a second eruption phase, the Ticuantepe Lapilli. In terms of volume it was a small compared to what had come just earlier. But it was an energetic plinian eruption that sent a buoyant plume 25 km high into the atmosphere (contrasting with the horizontal flow of the Masaya Tuff), it threw dense blocks of old rock 30 cm across, 15 km away. Most of the volume are lapilli fragments of fresh lava. The eruption still has signs of water interaction, like layers of tuff, but the style is drastically different, much more “magmatic”. What caused this sudden change in style?

The 3 eruption cases that I have described in previous posts turned out to be once again the key to solve the enigma. Taal and Fernandina had more in common with the Masaya Tuff though at a much smaller scale. They are hydrothermal blasts. However the 2000 August 18th eruption of Miyakejima had more in common with the Ticuantepe lapilli. It was a more magmatic eruption, while still phreatomagmatic, there was a lot of fresh lava towards the end. The difference between them is that Taal and Fernandina blew long before or without a caldera collapse, while Miyakejima blew during a collapse event. And this suddenly made perfect sense. Let me explain.

If the Masaya Tuff was a hydrothermal blast then a jet of steam would blow open a path from the surface to the magma reservoir. What would happen if a collapse event took place as the jet was still active? Remember collapse events? They mean that the ground suddenly drops on top of the magma reservoir and pressurizes it, magma would be squeezed out into the steam jet, you can imagine what a mess this would be. Volcanic gasses being violently released, the magma meeting whatever water remains in the conduit, magma shooting upwards at amazing speeds, a plinian/phreatoplinian eruption is the reasonable result. This is an unproven model of course, but the only one as far as I know than can succesfully explain the formation of the caldera, the extraordinary explosivity of the often gentle basaltic magma and the sharp changes in eruption style at the same time.

The hydrothermal blast can happen before, at the same time, after a collapse event, or with no collapse whatsoever. There can be multiple blasts and collapse events too. This gives a wide variety of eruptions a BBBs can take. Here I have illustrated, one of them, the Masaya Tuff model:

The terrible past of the Managua Graben is written in rock. The fiery cauldron of Masaya is a reminder that this area is geologically alive, but there is no surviving memory among the people. Can we blame them? This region has dormant for so long, and it could be centuries or thousands of years before the next disaster is ready. Hopefully we could say the same for all volcanoes of the world, but some just don’t know how to stay put. Enter the hyperactive volcano who is an angry goddess and a giant volcanic laboratory at the same time:



A volcano of legend, abode of Pele, the terrible goddess of volcanoes, fire, lightning, and so on, feared by native hawaiians. Then as the western world arrived to Hawaii, Kilauea went for a whole century of highly touristic lava lake activity. No surprise it started to be considered as one the safest volcanoes of the world. But now it stands as the number 1 most dangerous volcano of the U.S. in the ranking of the USGS. Yes, it ranks above volcanoes such as Mount Saint Helens or Rainier. And believe me, there is good reason, and I have additional reason to believe it poses a threat in the near-present.

Back when Pele was feared, not like now that Aila’au is the one getting blamed when the volcano does something destructive, the explosivity of Kilauea was well known through oral history. The last story passed down and thus best recorded is that of Keonehelelei, means “the falling sands”.

Kilauea erupts in 1924. This was a dry collapse event, similar in this sense to 2018. USGS.

In 1790 a civil war was taking place on the island, after the death of the previous king, ambitious Kamehameha had tried to seize the rule, sparking a conflict. It was finally down to the cousins Kamehameha and Keoua. Learning that his homeland was under attack, Keoua decided to race there through the quickest route, which led through Kilauea, was his fate sealed at that moment? The victory of Kamehameha solidified? Or was the ambition of Kamehameha, who would later unify the islands, too strong to oppose from the beginning? Who knows.

While encamped at the summit an eruption started, explosions continued for 3 days, which they used in trying to appease Pele. Eventually Keoua decided to advance, they walked over a recent layer of wet accretionary lapilli in which their footprints became imprinted, much like at Masaya. I think about the people of Masaya, were they able to escape or would their fate be as sad as that of the people of Keoua? Just as they marched, the paroxysmal phase of the 1790 eruption started. The army had been divided into three groups, the first party felt as if they were suffocating and some “were burned to death by the sand”, this could mean either acid or heat. The last party was caught in the ash fall but was fine. When the darkness of the eruption cleared they resumed the advance to find everyone in the second party dead, they had been caught in a base or pyroclastic surge (hard to tell the difference and their definitions are not clearly set). The bodies were lying on the ground or sitting, some were touching noses, a gesture of affection. In Hawaii warriors travelled with their families.

There are many estimates on the number of casualties, but they are usually cited to be 400-800. Kilauea is a shield volcano, it is covered mostly in lava flows but it can do this. Fernandina is also a shield volcano… You see where I’m going. Don’t underestimate any shield/basalt volcano that has a caldera or lake at its summit, or better don’t underestimate any volcano at all.

The 1790 eruption fits the characteristics: It is mostly lithic, very violent, wet, there is also evidence for a rift eruption and a caldera collapse at the time, so I consider it to be an example of a big basalt blast.

William Ellis was the first westerner to visit Kilauea. In 1823 he described 3 calderas nested within each other, he named them Sunken Plain (1), Great Crater (2) and Gulf (3). (1) started to form in 1500 AD and later deepened further, the rim is still visible today. (2) or Great Crater is the caldera that must have existed before 1790 and was enlarged in that eruption. (3) formed the same year of the drawing, 1823, in the Keaiwa eruption. Note that part of the caldera floor was an enormous pool of lava at the time of the visit.

Some of you may be wondering about the fate of Keoua. He survived, but later lost the war. Even more of you are probably wondering about the new lake on top of Kilauea. Yes, it is bad news. Where the collapse of 2018 was dry, a new collapse in the caldera as it currently is will turn phreatoplinian in a way akin to 1790. There are many ways the lake can blow (dyke, open conduit, phreatic) but the most worrisome is probably if it is depressurization-triggered, the 1790 scenario. Kilauea collapses frequently, sometimes sizable collapses have taken place just 10 years apart from each other. With its south flank moving and its east rift inflating, another collapse in the next few decades is a very real possibility, or even a likely possibility. Some small towns are located dangerously close to the caldera of Kilauea. If someone from HVO asked me, I would tell them to prepare for the worst.

Before I list all the shield volcanoes not to be underestimated, there is one last place that deserves special mention, we keep cutting across the Pacific Ocean and reach the land of exotic volcanoes.


Central Vanuatu is home to 3 adjacent and unusual volcanoes, Ambrym, Ambae and Gaua.  They are a puzzle of mysteries and contradictions. They are basaltic but their surfaces are covered in thick ash deposits. They are in a geologic setting that should be highly compressive, where an ancient volcanic arc is colliding with the active one, but they have very efficient rift zones. The volume of their edifices and the supply they enjoy seems exaggerated for normal arc volcanoes, Ambae for example has a volume of 2500 km3, more than 4 times the largest volcano of the Cascade Arc! With their large calderas and 2500 to 3500 mm of annual rainfall, these 3 should be quite the hotspot for big basalt blasts, could this be the case?

Ambrym stands out for its high activity. It has been erupting almost every year of its short recorded history and in a myriad of ways, from submarine, to vulcanian, or subplinian, hawaiian and the impressive feat of sustaining 4 lava lakes at the same time. The island has been described as a giant tuff cone. There seems to be more ash than lava, and it is topped by a large 12 km wide caldera. Among the many eruptions that have shaped the island there is likely to be a contribution from some big basalt blasts to have accumulated such an impressive and widespread thickness of ash.

Lava lake within Marum. Photo from Claire Cousergue.

Does a large caldera like that of Ambrym mean potential for a larger explosive eruption? Intuitively it would seem so. Logically a broad caldera means that the hydrothermal system can be more voluminous and that there is a greater area of contact between water and the magma reservoir which may allow a higher heat flux if pressure falls. If the roof drops and magma is squeezed out there is more magma available for this too. It seems therefore that Ambrym may be particularly suited for large magnitude events, Gaua too. The tens of meters thickness of pyroclastic material at Gaua and the few hundreds at Ambrym do look suspicious, so far study is poor and there are no good dates available.

Ambrym poses a near-future threat, like Kilauea. The pit craters at the summit have kept deepening since 1943. It should be about time they hit the water table. A corner of the 12km caldera started rupturing during an intrusion in 2018, the next collapse might well rupture all around.

The twin summits of Ambrym, Marum to the left and Benbow to the right. The pit crater complexes at their summits didn’t exist back in 1943 they have formed in several unknown collapses and the last and greatest one in 2018, when a dyke 0.7 km3 in volume was intruded into its east rift zone. A future collapse could be wet and produce a BBB. Image by Stuart Rankin.

To the north of Ambrym we find Ambae, here they had to deal with damaging eruptions in 2017-2018, but these are really child’s play compared to what this volcano should be able to do. The potential to generate lahars has been remarked before, this is certainly a hazard but I am much more worried about a base surge. The summit lakes of Ambae are perched 1400 m above sea level, there is no obstacle to contain a base surge, the steep slopes would add even more speed to it and it would probably be highly acidic (Lake Voui is an acid lake). An eruption similar to the one of Taal in 1911 would be amplified by topography and lay waste to most of the island faster than people can realize what is going on.

Eruptions of Ambae in 1870 and 1914 resulted in casualties but nothing else is known about them, so it is possible that they may represent phreatoplinian events related to the formation of Lake Voui. No cone existed inside the caldera of Ambae before the eruptions this century, this suggests the caldera may have been pristine (from 1914?), freshly formed. So it is an interesting possibility for any researcher, and it could help figure out exactly what are the volcanic hazards this island faces.

Lake Voui, the exploding lake. Image by conradh.


Final thoughts

Once thought to be safe, basalt volcanoes can actually be quite deadly. Their greatest weapon? Probably the base surge. We have seen Taal deal a deadly devastating blow with a shockingly small volume, likewise Kilauea is to blame for the deadliest eruption in US history. Falling blocks, lahar, ashfall, tsunamis are among the other possible threats of big basalt blasts. I have evaluated the potential of many mafic/basaltic systems to produce these events and compiled the list below you may find volcanoes that you wouldn’t expect to see, but if you start to look at them with different eyes then I am relieved for this series has served its purpose.



At some volcanoes conditions to produce BBBs do not seem to be met, but it may not be completely impossible either. Unlikely: 

Sierra Negra, Darwin, Mauna Loa, San Carlos, Tolbachik, Bardarbunga.

The following could meet the conditions but I found this uncertain. Possible:

Nyamuragira, San Joaquin, Villarrica, Wolf, Ecuador, Pinta, San Salvador, Yasur, Hachijojima, Isla Tortuga, Fuji.

This group shows the adequate conditions or there is evidence for past phreatomagmatic eruptions or both. Probable:

Nyiragongo, Etna, Cerro Azul (Galapagos), Marchena, Genovesa, Piton de la Fournaise, Grimsvotn, Karthala, Niuafo’ou, Ambae, Aogashima, Izu-Oshima, Kilauea.

The following are probably capable and the size of their calderas suggests large magnitude events could happen. High VEI 5 to 6. Probable:

Masaya, Gaua, Ambrym, Taal.

The last ones have been confirmed to produce BBBs, as there is clear evidence for lateral draining as the trigger:

Miyakejima, Fernandina, Taal.


Relevant links

On the explosive eruptions of Masaya, a very complete thesis on all 3 events:




83 thoughts on “Big basalt blasts III. Over the world.

  1. Ticuantepe stage must been an impressive lava fountain! Plinian basaltic fountains are stuff of hell. An angry ligthing ridden yellow orange at base column to the tropospause.
    Perhaps an oversized version of Grimsvötn 2011 but without the ice and with flank drainage. It must been an impressive sight.. to see 😄🌋

    • The magmatic-like explosive eruptions of Masaya often have Pele’s tears, showing the lava was very fluid and hot. It would have been a powerful sight, an enormous incandescent fountain, as long as you were far enough not to be killed by the preceding base surge…

      • Though I am not sure if it would be visible with all the steam flying around after the initial stage. Would be hard to get a glimpse on it.

    • If it lasted long enough to be visible at night (so at least a day, quite probable) then you would see the glow 🙂

      Basaltic plinian eruptions and lava fountains seem to trend directly, bypassing the peleean and vulcanian phase, probably all the diagrams need to include that. These eruptions are as Jesper says pretty much right out of hell, bright glowing fountain and black ash. Etna in 2015 and 2016 is probably the most visual example, its basically picture perfect even if on the relatively smaller side.

      I cant remember the source and believe me I have looked, but I remember reading about a series of eruptions at Kilauea, part of the Uwekahuna Tephra member, that were not dissimilar to the Ticuantepe eruptions but had associated lava flows of (apparently) such considerable scale as to be considered flood basalts. Possibly the source has been paywalled since I read it…

      • The K3 Upper Kulanaokuaiki eruption of around 900-1000 AD had 2 phases like the Ticuantepe Lapilli, but smaller, and 2 phases that resemble hydrothermal blasts. There are few facts known about older explosive eruptions apart from their existence.

      • The Sakursunarvatn eruptions from Grimsvötn with Thoelitic magmas produced 150 km3 of materials in many large events. These events destroyed the large shallow pleistocene chambers. A quite heafty glacial flood must have been accompanied these events. Today it woud shut down the airspace for months if it happened again.

        Each event involved more than 30 km3 of basalt in just a few days basicaly 2011 just 40 times larger each time. A very crazy sight with quite heavy ashfall in Faroe Islands and an angry ligthing ridden spectacle in Iceland. Perhaps with an intense orange glow at base. Ashfall from Sakursunarvatn tephras are common in western europe.

        No known rift lava flows where erupted under Sakursunarvatn basalt plinians, appears to be entirely summit eruptive events.

        • No rift lavas, but that doesnt mean no rift eruption. Sakursunarvatn is not in Iceland and theres no tephra of that eruption in Iceland apparently, sounds very very sketchy to be so certain about its eruptive mechanism. If it actually was that volume to have none of it exposed it probably all erupted onto ice that flowed to the sea and dumped it offshore, in that case its source could be a rift, no way any eruption is melting all that ice and there looks to be a lot of subglacial ridges near the laki cones so who knows.

          • From Icelandic Volcanoes:

            “Geochemical analyses [of the Saksunarvatn Ash or G10ka series tephra (9900-10,400 years old) ] confirm its origin in the Grímsvötn volcanic system. These large events may have been the result of a lower mantle pressure due to changes in isostacy at the end of the last glaciation.”


          • It is not what volcano the eruption came from that is in debate, it is that the eruption was an ignimbrite caldera eruption, those only happen in volcanoes with viscous magma, basalt is too liquid to erupt that way and Grimsvotn has basalt nearly identical to Kilauea, and more importantly to Bardarbunga, both of which are not known for big ignimbrite eruptions but for large rifting effusive eruptions especially the latter, as im sure you know. Add a lot of ice (which is the only way I can think of why the tephra isnt actually found in Iceland) and you make those rifting eruptions explosive, like 2011. 150 km3 of tephra probably is under 50 km3 of actual magma, which is about 3x the volume of 1783. If the magma generation was double todays value 10,000 years ago then 3 skaftar fires type eruptions in 200 years is not impossible, it doesnt require some sort of VEI 7 scale eruption that seems to have been inferred in some of the old articles on VC.

          • I found the knowledge on the Saksunarvatn Ash to be very small while researching for the articles, which was disappointing because these could be the largest basalt explosive eruptions of the Holocene, or at least as far as we know.

            But Grimsvotn did do a big basalt blast in 1783-85 during the Skaftar Fires, this means the volume and VEI of the Skaftar Fires may actually have been underestimated. The tephra volume has been calculated from the explosive eruptions of the rift, while the summit explosions have been ignored, but they are there. It is also the longest lived observed (and reported) BBB eruption.

          • Grimsvotn should be in the confirmed to produce BBBs part of the list, now that I think of it.

    • There is often faulting at that location, it should gradually calm off if that is all that is happening.

    • Ka’oiki fault zone, tectonic faulting related to the flank of Mauna Loa sliding slowly. This place did a 6.3 in 1983 as a prelude to Mauna Loa erupting in 1984, Pu’u O’o had already opened by that time so no reaction was noted from Kilauea but that is no longer the case, doing the same thing now I suspect would lead to a significant eruption.

      It is most certainly unfoutunate that Europeans and their ‘recorded history’ turned up when Kilauea was in continuous gentle eruption. They didnt even actually do that, Cook landed in 1778, before the eruption of 1790, and there were Europeans on the island in that year… Now we have to resort to geological records. Fissure 8 is just another cone in that part of the rift and actually a bit smaller than some of its neighbors, evidently there have been a lot of comparable eruptions. Theres even another Pu’u O’o, next to Makaopuhi crater that HVO are certain to be aware of but seem dismissive of its size. That shield I dont know how old it is or what it is called (might be Kanenuiohamo mentioned in some of the early Mauna Ulu reports) but certainly shows at least one other case of an open east rift, maybe one of those cones in Puna is what ended it. It is like trying to do a whole puzzle with only half the pieces, just a slightly better guess. Hector you probably have a much better understanding of this than I do 🙂

      • A puzzle indeed, I think I got that puzzle more or less. Kane Nui O Hamo is around 1000 AD judging from paleomagnetic evidence, a century up or down. My guess is that it came just before the Upper Kulanaokuaiki eruptions and caldera formation, the K3 BBB? marks the end of that cycle, akin to 1790, and the K4 subplinian eruption was the start of caldera filing, akin to the Golden Pumice. It is my reconstruction of the puzzle.

        • Do any flows or cones in the LERZ line up with the end of the Kenenuiohamo eruption, an ancient fissure 8? I wonder if the K3 eruption was during the collapse of that eruption, or if it was an eruption following on from that after a lake had been created, equivalent to a caldera collapse happening right now… One of those things is decidedly more dangerous.

          I know of that paper on paleomagnetism for the lower east rift which you linked to, are there other papers on other parts of the rift? I much want to know of the area that is buried under Pu’u O’o, it was very active in the 18th century but little else is known, I suspect there is no data on it though.

          I found a paper on the tephra of Hawaiian eruptions, and the reticulite is in most cases about 1% of the eruption in total, 1959 is a good example being that in a few hundred years the only big left might be the distal reticulite but we know it made a massive lava lake in the crater, and just look at the enormous lava field around Pu’u O’o which is what happens if you do that outside a crater. There are some deposits that are about 10 million m3 of reticulite which is maybe 1 km3 of liquid lava, that is erupted as a lava fountain in a single event… I dont know if theres much study on that stuff but one of those now would totally fill the 2018 caldera which is quite incredible. Certainly 1959 seems far from the peak of these eruptions. I think the biggest reticulite deposit is the first layer of Keanakako’i tephra, goign as far as Volcano, that is the one with over 10 million m3 volume, I do wonder if it is the source of the glow seen from Oahu in the Pele-Hi’iaka legend, you cant directly see Kilauea because the rest of the island is in the way from Oahu but a massive glowing lava fountain you probably could, or at least the glow in the atmosphere.

          • I have the lava flow f6d6 from “geologic map of the lower east rift zone of Kilauea Volcano” as similar in age to Kanenuiohamo based on paleomagnetism and C14 dating, but as you can see it is just a small kipuka. The LERZ was almost completely resurfaced in the 1500-1800 interval, so anything older than that is just a kipuka here and there.

            K3 erupted in a deep caldera because there had been surtseyan eruptions some time before. The legend of Kamapuaa talks of a water lake before a battle of Pele and Kamapuaa and a shower of stones over the flank of Kilauea, I think it refers to the K3, the only thing in the geologic record that can be called a shower of stones…

            I should put the information that I compiled somewhere for download so that it is available,

          • Human habitation postdates the K3 eruption, though I also noted the similarity when I read the story to the HVO paper on the K3 eruption. The association with Pele is also indicative of a date in the last 500 years being the origin of the Kamapuaa story, she is associated seemingly universally in all the stories I have read outside some modern ones with being short tempered and agressive, fitting with the often explosive summit eruptions and fast flowing rift eruptions of the past 500 years. Aila’au is the ‘forest eater’ who seems to be true neutral, for lack of a better work, his actions are destructive but gentle which doesnt actually seem to fit that well to 2018 but its not really my place to question. It has occurred to me that the literal namesake eruption of Aila’au is associated with Pele in the story though…

  2. Very informative article! Some of the volcanoes listed as possible locations for future BBBs are disconcerting, such as Etna. Having a BBB there would be bad news!

    • I considered the 122 BCE plinian eruption of Etna a possible candidate of a BBB because it has signs of phreatomagmatism and is thought to be related to the last caldera formation of Etna. That one is reported to have collapsed houses on Catania. And there are several other plinian events in the Holocene.

      • Katla already has regular eruptions about that size, VEI 4-5, it is still recovering from Eldgja so probably no risk. Askja more so, but it erupted like this in 1875 so probably not a lot of risk there. That could change though.

        Hekla is not capable of erupting like this, it has no shallow magma chamber abd lacks the supply rate and volume to induce a rift, probably will take a long time to reach that point yet.

    • Askja was initially planned to be in the series, it was driven into plinian eruption in 1875 by a large rifting event with 2 small fissure eruptions beforehand. But the plinian eruption was of rhyolite lava, I didn’t include Askja and Katla on the list because they contain silicic magmas.

      The way silicic explosive eruptions can be triggered by lateral draining is something I may write about in the future.

      • According to GVP, Katla has had large basaltic tephra producing eruptions since Eldgja.

          • For simplicity of the articles I wanted to leave the bimodal volcanoes for some other time (volcanoes with a significant amount of both mafic and silicic magmas). Otherwise I should have talked about all Azores and East African calderas plus other locations.

            The thing is that if either Askja or Katla drained laterally to the point of collapsing or near-collapsing what would come out would be the rhyolite at the top of the reservoirs which is very explosive on its own, unlikely basalt, so the mechanism would probably work somewhat different and have more to do with fragmentation of the magma.

            That said hydrothermal blasts do probably exist in silicic volcanoes and are probably could be responsible for eruptions that produce large ignimbrites but small distal ashfall.

          • Think the silica magmas for Katla and Askja come from crustal melt. It’s rifting which is driving volcanism there so predominantly basalts.

      • Katla haves a very large magma mass influx

        ”using the lower limit of the observed CO2 flux of 12 kt/day the required minimum magma mass flux (assuming 1.1 wt % CO2) is 0.18 km3 per year. If the lower limit of 0.1 wt % is assumed, the volume of magma increases by an order of magnitude (1.8 km3 per year). Further ground deformation studies at Katla are recommended to shed light on the volumes suggested by the gas measurements.”

        • That just means it is degassing the CO2 out of that much magma each year, CO2 comes out of solution very deep, there is probably very little of that magma actually going into the active system of the volcano or it would erupt more often. If it was SO2 that is a better indicator of real eruptible magma flux because it is showing magma getting close enough to erupt. There are a lot of volcanoes which a supply rate of crazy levels is reported, but the only place I know of with long history of actually erupting at sustained rates in this ball park is Hawaii, other places where high supply is presumed seem to be chimneys to degas massive areas rather than actually erupting huge volumes of magma or any of that magma even getting near the surface at all, it sounds similar but its really a very big difference. It makes sense anyway, Hawaii is a quarter million km3 pile of basalt that is under a million years old, if basically any other volcano on earth supposedly had higher supply why is it not a colossal structure too

          Katla erupted pretty big in 1918, high VEI 4 basaltic plinian eruption. Thats probably realistic for its next eruption, maybe go a bit higher to VEI 5 for it taking a long nap.

  3. nice swarm just west of Kilauea crater, just 3km deep. not an usual area for these.

  4. If anyone is curious here, Grimsvötn haves the shallowest magma chamber in Iceland. The uppermost resovair sits only 1600 meters below the caldera floor. And probaly is even more shallow now with pooled magma since 2011.
    Grimsvötn boasts one of the worlds most powerful geothermal systems, the heat been rising. It should be possible to drill into Grimsvötns magma chamber knowing it have bare outcrops in the caldera rim.

    • I wouldn’t advise drilling into a magma chamber small or shallow or otherwise mind!

      • There is project of drilling into Krafla’s reservoir:

        It was accidentally done in 2009, so I guess they are now opening up to the possibility of doing it regularly. It should be done in more places I think, it is the best way of learning how volcanoes work, being able to see what is down there in reality and not by indirect methods.

        • Does this not destabilise the pressure in the chamber? As well as creating an access point for rain/water? I imagine as soon as you drill in past a certain point you get a flash of superheated steam.

          Just seems like a risk not worth taking unless of course it’s an isolated and relatively low risk volcano away from civilisation.

          • There have been already 3 accidental drillings into felsic magma chambers, at Krafla, Kilauea (PGV), and Menengai. It seems magma goes up a few meters into the drill hole and freezes due to contact with colder water. So it seems to be something relatively safe but you never know 100%, a magma chamber should have enough pressure to push through the entire drill hole into the surface but so far it appears it can’t overcome the cooling from the water in the conditions of a drilling, there are many unknown factors though, what if the magma is basaltic? or what if the volcano overpressurized?

            Krafla is away from populations so I’m glad that such a project is being taken there, it could mean a new clean source of energy plus being able to measure a magma reservoir directly from the first time.

          • No worries, I think. A volcano conduit may be 5 to 50 meters diameter. The drill well is a lot narrower. The cooling of magma is much faster in a narrower conduit. And water (circulating in the drill well) cools the magma a lot more efficiently than the surrounding rock does (rock is a pretty good insulator, circulating water is not). I think the magma stands little chance.

          • Krafla had a drill well erupt lava in 1979, and probably that would have happened in 2018 at PGV if the dike went under it.

          • Seems as though there is a potential positive despite the risk. Just watched this video where they drilled down into magma and were able to convert the steam into far more energy meaning they only needed 2 geothermal wells instead of 20+

  5. Fabulous final article, Hector! Thank you! I’m never going to look at a basalt volcano in the same way again. As you said: don’t underestimate any volcano at all.
    I am impressed by your careful arguments and evidence. You have put so much work into this, and I hope this will be taken aboard by volcanologists.

  6. A wonderful series Hector. Thank you very much for such thought provoking articles.

  7. Nice run. Might have slowed down Pahala. Little movement, little pressure to be relieved?

    2020-10-24 23:20:46 1.9 2.5
    2020-10-24 22:14:48 1.9 2.9
    2020-10-24 21:31:35 2.3 1.9
    2020-10-24 21:31:34 2.3 3
    2020-10-24 21:30:34 2 2.9
    2020-10-24 21:09:35 1.7 2.5
    abbreviated – admin

    2020-10-23 09:05:02 2.6 3
    2020-10-23 08:48:37 2.4 3.3
    2020-10-23 08:06:26 1.8 2.3


    • There was just a 3.9 at Pahala, actually more to the east of Pahala, 30 km deep. That swarm is probably going to go on for many years more.

      The part of the swarm closest to Kilauea is quite dense with quakes compared to being pretty randomly distributed elsewhere, it looks like a bow shock almost, the majority of quakes happening at the leading edge as it advances. The swarm is showing a deep intrusion, probably a big sill forming that is heading towards the most open hole to the surface which is Kilauea, it is increasingly clear this is the case, the big spike in August of last year, it hasnt quieted down too much since then, all the deep tremors, it is obvious magma was on the move back then and now it is clear as daylight where it is headed.

      I wonder if LERZ activity at Kilauea is common to occur in pairs within a few years of each other, the first event drains out a long term accumulation after decades of steady activity, while the second is a consequence of excessive decompression melting overfeeding the system and causing a second event. Eruptions in 1955 and 1960, intrusion in 1838 and eruption in 1840, eruption or eruptions in about 1780 and intrusion or submarine eruption in 1790.
      Also envokes the scenario of the article…

      • The NE boundary and densest part of the swarm is possibly a transition from brittle to ductile mantle, the surge of volcano-tectonic earthquakes already reached that line in the first months of this year, so the intrusion is probably advancing somewhere beyond the boundary, silently, in a hot ductile mantle.

      • There is no indication that the edge is advancing. The shape shows that the stress in this area is governed by Mauna Loa. The only indication for movement has been to Lo’ihi, and even that may just be tectonic. The eye sees patterns where non exist.

        • The quakes closer to Loihi are deeper at 40-50 km, maybe another magmatic feature but a separate one. There are also frequent quakes at the same 30-35 km depth as the Pahala swarm between it and Kilauea and generally scattered close to Kilauea, but not around the other volcanoes, though amongst the shallower quakes it is difficult to see them. If you set the HVO quake map to show depth over age it is easier to see.

          It would be very helpful if a 3D render of the swarm exists but apart from one in 2018 which was not public I have never seen such a render.

          • It is best if you see for yourselves by animating the earthquakes of the following link:


            But my interpretation is this: The Pahala area has been swarming for as long as monitoring exists but there is clearly a surge in 2019-2020 that propagates like a wave 30 km horizontally and 10 km up in the general direction of Kilauea.

            It starts in the 1st half of 2019 with offshore quakes that were spectacular bursts of deep tremor, up to 9 tremors in a day. Then there is an onshore swarm of ground cracking earthquakes that moves to the NE until February 2020 when the peak of the swarm is reached and propagation seems to end, I think the intrusion moves into an invisible ductile area at this point, and the leading edge may have been travelling silently since then.

          • Very interrsting, I dont know how well the 3D render on that site calculates depth, but if it is accurate then the swarm is very flat, it is a sill that has been created with the quakes on its exterior. The age progression is interesting too, how it begins diffuse then with a sharp local increase and high activity sustained since.

            If you hypothesize the intrusion became aseismic after reaching ductile mantle closer to Kilauea, possibly that magma is now intruding into the volcano and is what is causing the GPS signals at the summit to record upward movement but not the tiltmeter as the shallow system hasnt been fed yet.. 7 months is probably sufficient time to give this some plausibility.

          • I played around with that site, this is the quakes that have taken place deeper than 25 km since September 1 2018, the aftermath of the 2018 eruption. It is quite clear and undeniable that Kilaueas deep system is seeing a lot more action, theres a notable cluster just right of the caldera at between 25 and 33 km deep, while at Mauna Loa theres a handful of deep quakes only, and all very deep over 33 km.


            This is the same area and statistics from the time that Pu’u O’o was being created in the mid 1980s.


  8. And harking back to a post that I feel dovetails quite well with this series: Carl’s Watervolcano article.

    The really interesting bit is that Héctor has opened up quite a few more systems that are potentially some of the more dangerous systems around the world. The thought that comes to mind for me, is that when a volcano looses magma down a dike swarm, it is likely that the main vent seal has not been broken and the magma chamber is effectively only coupled with the dike swarm where it is loosing magma. Magma, specifically basalt, has a density of around 3100 kg/m³. When flowing, that’s an appreciable amount of momentum. In my opinion, the reverse of a hydraulic hammer effect could occur and impart a vacuum of sorts on the “lid” of the main vent. This radical pressure drop + the mass of the edifice could easily be enough to invoke structural failure of the lid. This could be what happened with Katmai when the 1912 Novarupta event caused about 1000 feet of Katmai to collapse down into the chamber leaving a caldera about 3.2 by 4.8 km… which has now become a summit lake.

    A very good article Héctor, definitely food for thought.. and timely, Halloween is coming and this is quite spooky, 😀

    • {The reverse water hammer came to mind since I recall an event at a fire that I was on when we had a pumper pulling a vacuum on the closest hydrant, pumping water 1100 feet up a slope to the equipment working the fire. We managed to collapse a few hot water heaters on the nearby water circuit. Yes, the county had to pay for them.)

  9. Here I sit on a basalt substrate as deep as, as hard as and shaping the geology as far as the eye can see. Surrounded I am by seven prominent volcanic cones. The lava flows extend east and then many hundreds of kms to the south and south west; where even bigger cones dominate the landscape, many with lakes in the calderas which have collapsed. The biggest known as Tower Hill. All are said to be ‘at rest’ but one never knows. Hereabouts, it was under this basalt that Ordovician rivers were buried and from their beds came fabulous riches in gold. The basalt itself and the water table were significant obstacles, but no match for gold fever. Have our extensive lava plains here in Downunder Victoria been overlooked?

    • Yes, its overlooked. It isnt as unknown though anymore but still overlooked. There is 400 mapped vents in the area and so at that it erupts every 10,000 years, but the number of vents is likely much higher. There has been at least two confirmed holocene eruptions, at Mt Gambier and Red Rock Craters, and theres a lot in the ambiguous ‘20,000-5000’ year old range, so it is probable that eruptions happen there somewhat often, and it has been 5500 years since the last one so its not exactly out of the question for another one to appear pretty soon.
      Blue lake was studied and found to have been driven by a very rapid ascent though, it is something like a few days from earliest signs to first eruption, so basically it is pretty unlikely you would know more than a week in advance if a volcano would appear, except that there wont be a volcano in the next few days, anything past then is uncertain.

      Theres also holocene volcanoes in Queensland at Atherton and McBride volcanic fields, including Kinrara which was as large and similar in eruption style to Pu’u O’o but in reverse, 7000 years ago.

      There is also one technicality that is entirely overlooked, there actually have been historical eruptions within Australia proper. Papua New Guinea was a territory of Australia until for 60 years up to 1975, so the major eruptions at Rabaul in 1937 and Mt Lamington in 1952 were technically on Australian soil, and unlike Heard island are also on the same continent 🙂

    • Those water-filled craters, maars, are dangerous. They usually form in 1 eruption, but that eruption can destroy everything and everyone in a radius of several kilometers due to surges.

      Luckily for your area eruptions are rare there, and if one happens you will know very well before it starts. Shaking will be very intense, scary and go on for days. These eruptions have to open conduits all thw way from the lower crust or mantle which is a lot of cracking.

      • Hector there have been studies done on a number of the volcanoes of all types, the eruptions are a day to 3 days advance at most mantle to surface, they are incredibly fast. Eruptions now are relatively infrequent but the area was very active back at the beginning of the Pleistocene, tholeiitic flood lavas are common, showing pretty serious melt generation, now it is alkali olivine basalt, the field is past its peak but the heat is still there. This part of of the crust is nowhere near as thick as the rest of the continent, only 30 km or less similar values to a rift valley of which it indeed represents the north edge of one that existed in the Mesozoic.


    Best video of Masayas lava lake.
    Spectacular, beautyful, scary and dramatic video recording from Masayas lava lake. Among subduction zone volcanoes this coud be the hottest and most fluid of all subduction zone basalts. Masaya is also Thoelitic in its chemistry, suggesting vigorous melt production. It was a few years ago the current lava lake formed. Masaya is really really fluid ( not far behind Hawaii sometimes ) 1160 C been measured in the lava lake. The lake is really convecting violently in this video.
    The viscosity is extremely low for being a subduction zone basalt, Masaya rises quickly from depth, with little or no diffrentiation. I loves this video.. the lava lake is a giant whirrpool

    • Spectacular video for soure!
      Thanks alot finding it.
      Way back in the early 1500 s, the conquestadors thought its liquid gold, that was roiling in the pit.

      • There is a tale (possibly apocryphal) that one determined ‘gold’ hunter was a Spanish Friar -so much for his vows of poverty! Climbed into the pit with a bucket and a long pole. Amazingly, he survived, only lightly toasted, but empty handed – the bucket melted.

        • I remember reading that story. It comes from an account by Pascual de Andagoya, around 1515. He was accompanying Pedrarias de Avila, appointed governor who was certainly looking for gold. Whether the friar was looking for gold is not stated:

          “A friar, they say, entered as far as the ledge half way down the mouth, and thence he looked down and saw a certain thing like metal, of the colour of fire, and he let down a link of an iron chain by a rope, but when he drew it up he found nothing”

    • Chad if Masaya overflowed woud it be pahoehoe or Aa?
      This lava is certainly very fluid but may lack hawaiis shiney smoothness when it comes to the thin crust, suggesting viscosity just a bit above Hawaii. But overall this is a very fluid subduction magma.
      Its also very gas rich. Its certainly more runny than Pacaya and Etna and maybe a bit more runny than Ambrym

      • It would be pahoehoe because that is what formed when it has overflowed in the past. The crater fill is also pahoehoe, the flows turn to a’a after flowing far enough but that is the same in Hawaii too. Masaya seems to have been very active since it blew up last and probably the older cone formed above the Ticuantepe vent, or at least the modern volcano is probably directly related to that. Theres a lot of young lava flows, it evidently used to overflow frequently, maybe the rift is more active now than it was 1000 years ago and prevents large eruptions, which is maybe not a good thing…

      • Perhaps the lava thats churning in Grimsvötn woud look very similar to this video if it was exposed to the air.
        But Grimsvötn shallow magma is probaly 40 C hotter than Masaya is


        Chad watch this video
        In this video of Masaya its easy to see that the viscosity is a bit higher than Hawaii for example.
        Masayas magmas are more clumpy in apparence and does not form a smooth paintlike glowing crust in the night – shots.
        I also knows that this lava is extremely gas rich. But the viscosity is not like the lowest of basalts. ( Masaya still haves very low viscosity )

        • Officially the lava lake at Masaya is about 1100 C but not directly measured, 2011 eruption of Grimsvotn has variable temperature between 1100 and 1150 C so it is negligibly hotter. Hottest confirmed measurement of lava I can find is 1240 C for the old Halemaumau lava lake before 2018, though temperatures of 1400 C have been recorded in recently active vents and tubes on Kilauea where hydrogen in the plume (created by reaction of water) combusts, this was noted even in the early 20th century.

          • Possibly the turbulent nature of the Masaya lake, and also at Ambrym, causes the transition to a’a.

    • I happened to be there just as the lava lake was emerging on Dec 26-28 2015. I was there for a wedding where the couple got married in the parking lot next to the crater. The night shots from the wedding showed the orange glow just emerging from the crater floor so we went back each night we were there to see the beginning of the lava lake. Pretty cool experience

    • During the recent swarm Kilaueas caldera expanded slightly and there have been a lot of quakes around it, UWEV-CRIM cross caldera GPS shows a slight upward movement which suggests there was actually some sort of magmatic element there. There is no such change at Pu’u O’o or at Mauna Loa, actually the cross caldera GPS at Mauna Loa has taken a dive (its a really small scale of a few cm but still not something typical), dont know what to make of this.

      There was also a rare deep quake at 14.6 km depth half way between Mauna Ulu and Kilauea Iki, there is magma on the move down there too, and the GPS units at the summit are still showing deep set inflation, 5 cm in 3 months at AHUP, BYRL, CRIM UWEV and PUHI stations, basically all the stations that work, the entire summit of Kilauea has risen about 5 cm since the start of August but it is all in the deeper storage so it shows nothing on the tiltmeter. It is looking quite promising now 🙂

      • Rates of seismicity are high, 100-150 earthquakes/day, lately, even though there is no long period swarm and no shallow inflation which would be the situations that usually result in such levels. I do think that the deep inflation is creating this activity, it is stressing faults over a broad area. There are 2 possibilities that I see, 1 is that the shallow system has completed refilling so that inflation has switched to the deep levels that have been at low pressures until now, 2 is that a deep magma surge maybe the same responsible for the swarming at Pahala is at the summit, already, making its way through the deeper part of the storage.

        1 would mean that there won’t be much change, 2 that rapid inflation is likely to affect the shallow plumbing soon.

        • I think it was determined that the deeper system didnt get involved in the 2018 eruption other than being a hydraulic connection between the shallow magma under Halemaumau and the conduit to the east rift, so it probably has not changed in pressure much at all.

          I did notice that the cross caldera GPS at Kilauea actually also reads about 5 cm, so it is seemingly a radial expansion deep down, I think the mogi model would work to try and get a volume.

Comments are closed.