The Aleutian volcanic awakening

Recently, Mjf and Tallis Rockwell commented in the café that several volcanoes in the Aleutians were undergoing unrest. After some thought, I realized that the Aleutians are within the United States territory, in Alaska, and that means they are very well monitored, and that this data is accessible. So I decided to dive into it. Turns out something really interesting is going down in this cold and remote volcanic arc. Four different volcanoes/volcanic complexes are experiencing growing seismic unrest, which started in late 2022 and has increased further into 2023.

First of all, I should say there is no need for much worry. All the volcanoes in this post are very remote, and the surrounding area is uninhabited. Only Aniakchak has a nearby population, which is Port Heiden, and for it to damage Port Heiden, it would have to go really, really big, which is highly unlikely. Even if they erupt, they shouldn’t do much damage. Additionally, they might not erupt. At this point, the activity is merely intrusive or gas/water driven. A drop of magma will not be much if you are trying to fill a whole bathtub. Recharges of magma and increases in hydrothermal activity, and degassing that might or might not be related to these recharges, are common phenomena in volcanoes. Volcanic systems have a rich internal activity that does not necessarily lead to eruptions.

In this post, I’m simply looking into a couple of volcanoes that are shaking. I will also look into their past activity, so that if they should erupt, we will know what to expect. So here we go into the four reactivating systems:

Katmai Cluster

In the Katmai cluster, we have three different areas of seismic activity under four different recent volcanoes. All of these three areas have been escalating in unrest. The volcanoes are contiguous. From east to west we have Katmai, Trident, Mageik, and Martin. Katmai and Trident have their own separate seismic clusters. While Martin and Mageik are immersed within the same seismic flurry, and this topographically makes sense, given that these two cones are part of the same mountain made up of a number of old eroded overlapping volcanoes.

I have downloaded seismic data from the IRIS Earthquake Browser for each of these swarms. The three of them have monthly numbers of earthquakes that have been on the rise throughout 2022-23. The Alaska Volcano Observatory upgraded Trident Aviation Colour Code to yellow on 29 September 2022.

Earthquake depths are variable. At Trident, a large flare of micro-earthquakes averaging 24 km deep happened on August 25, followed by events averaging 16 km deep over the following 2 days, and since then events have been mostly shallow, averaging 4 km deep. Occasional deep events still continue to happen, though. At Mageik-Martin, earthquakes are very shallow, 2 km, possibly triggered by rising volcanic fluids, water, gasses and such. While at Katmai, earthquakes of variable depths are happening at the same time, with a cluster at 30 km depth, another at 16-19 km, and a third cluster from 12 km to the surface.

I also checked earlier years. Earthquakes as small as M 0.1 are located in 1997. So, since 1997, there seems to have been a very sensitive network. I don’t think this sensivity has changed much since then. The three locations seem to have experienced the strongest seismic activity in the 1997-2023 period, within 2018-2023. Trident had the greatest number of earthquakes in 2022. While Katmai peaked in 2020, although 2022 is still the second highest year, and Martin-Mageik peaked in 2019, with 2022 being its third-highest year in terms of seismic activity.

The Katmai cluster seems to have produced a dominantly effusive style, with the 1912 eruption being a rare event. The youngest major prehistorical activity is probably the one that formed Mount Martin. Enormous viscous flows of andesite-dacite lava with 62-64% SiO2 erupted from the presently ice covered cone of Mount Martin. These flows reach individual thicknesses of 70 meters and extend 10 kilometres to the northwest, infilling a broad glacial valley. The volume has been estimated in 9 km3. It has been proposed they were formed in a singular years or decades long eruption episode, which I think is likely. It happened at some point in the early Holocene, before 6000 years ago, given that ash layers of that particular age rest on top of the flows. Mageik also had a young episode of effusive activity which formed a 5.5 km3 edifice made up of 60-64% SiO2 andesite-dacite lava flows, each with individual thicknesses of 50 meters or more, this may have happened within the Holocene, but the age is not exactly known. Minor explosive eruptions may have also taken place in these systems.

The next major eruption is famous, and is the largest explosive eruption in the 20th century. The 1912 Novarupta eruption. This event happened from a vent on the north flank of Trident volcano. It erupted 17 km3 of pyroclastic fall material, and 11 km3 of ignimbrite (pyroclastic flows). The eruption produced two different groups of magma. First, a magma that is very rare in the world and in the Aleutians, a high silica rhyolite, with 77-78 % SiO2 and 0-5 wt% phenocrystals. And a crystal-rich andesite-dacite series, with 59-68% SiO2 and 30-40 wt% phenocrystals. The main episode of the eruption started with the high silica rhyolite, but in the latter part some dacite and andesite started to be mixed in. Two smaller andesite-dacite eruption episodes followed. The final event was the renewed eruption of high silica rhyolite that filled up effusively the bottom of the 700-900 diameter crater of Novarupta, making a small circular dome. The top of Katmai volcano collapsed during the Novarupta eruption.

The 3 x 4½ km summit caldera of Mount Katmai, Alaska, that formed after the 1912 VEI 6 Novarupta eruption (Wiki)

The summit caldera of Mount Katmai, Alaska, that formed during the 1912 VEI 6 Novarupta eruption (Wikipedia)

A new eruption episode lasted from 1953 to 1974, birthing a volcano on the west side of Trident. Known as the West Trident volcano. The eruptions consisted mostly of ~50 meter thick lava flows with a volume 0.7 km3, and of andesite to dacite composition, similar to that of Mageik and Martin.


The next volcano is an isolated caldera system, 200 kilometres southwest of the Katmai Cluster. At Aniakchak, earthquakes as small as M 0 were first located in 2018. However, just over two months of activity in 2023, has produced more than three times the number of earthquakes recorded in four years during 2018-2021. Activity started to build up in late 2022 and rose rapidly in 2023. The Alaska Volcano Observatory upgraded the Aviation Colour Code of Aniakchak to yellow on 22 February, during its most restless recorded month since before 2018.

Number of earthquakes per month at Aniakchak during 2022-23.

Earthquake depths are mostly within 6 km of the surface, and make two clusters. One cluster is under the southern side of the caldera, near the most prominent vent, which is Vent Mountain. The other under the east caldera rim, nothing special is there. Deep earthquakes extend to over 20 km underground below the caldera and along a band extending south from it.

We know little of the Pleistocene activity of this volcano, but more about what it did during the Holocene. Aniakchak erupted an andesitic ignimbrite with 60% SiO2, the Aniakchak I eruption, sometime around 9500-7000 years ago. But the volume or extent of this ignimbrite is not known. As far as I can tell, it has only been described near the present-day caldera rim, so may have been small. The volcano produced another eruption around 7000 years ago, the Black Nose Pumice, which deposited a considerable thickness of rhyodacitic pumice with 69-67% SiO2. Volume has not been calculated, but I get the impression it may have been a VEI-5 event. Then, around 3400 years ago, the 10-km wide caldera formed during the massive Aniakchak II eruption, which produced a 50 km3 bulk volume of pyroclastic flows. It started with 70% SiO2 rhyodacites and ended with 58% SiO2 basaltic andesite. About half the volume is dacitic while the other half andesitic. The pyroclastic flows reached into the Bering Sea 30 km away from the caldera, where they are exposed in cliffs along the coastline. These three are the most important eruptions of this period, but there were likely other events too.

Aniakchak Caldera

Aniakchak Caldera. The prominent cone on the right side of the caldera is Vent Mountain. The largest crater, located on the left side of the caldera is Half Cone. Below and to the right of Half Cone is the 1931 crater.

Following the caldera formation, activity has constructed 17 vents, consisting of craters and domes of various sizes, arranged in a 6-km diameter ring within the caldera. Encircling a shallow magma chamber that is represented by a deflating area in the middle of this ring. At 2300 yr BP (years before present), a plinian eruption produced a dacite pumice fall to the north of Aniakchak. Following this event, eruptions may have been more quiet or more infrequent, until recently.

Around 380 yr BP a powerful eruption broke out from the northwest side of the caldera, violent explosions blew open a 2 kilometre wide crater, known as Half Cone, pyroclastic flows and surges swept across the caldera floor, showers of pumice extended for tens of kilometres, while fine ashfall may have reached as far as 330 km away, and a total volume of 0.75-1 km3 is estimated to have been erupted. The eruption ended with lava flows which effused from a 100-meter wide vent in the center of Half Cone and filled up the crater. The flows are relatively fluid and have a thickness of ~10 meters along the edges. Silica content ranges from 67 to 58 wt%, decreasing gradually during the explosive eruption, but the final lava flows were crystal-poor dacites with 65 wt% SiO2.

At some point before or after the major Half Cone eruption, Vent Mountain became active on the southern side of the caldera. Vent Mountain makes a elegant 500 meter tall cone with an 800 meter-wide crater at the top. The cone seems to be made up of layers of spatter and ash. At the end of its activity, lava ponded within the crater. It was possibly formed by repeated intense vulcanian explosions, adding layer upon layer of pyroclastic material. The last spatter deposits look more youthful than the Half Cone lavas (less ash on them). After the last activity of Vent Mountain, a 2 km long fissure opened across the southern foot of the cone. Vents all along the fissure emitted fluid looking crystal-poor dacites with 65% SiO2, which formed a series of lava flows. These flows seem to be less than a few meters thick on steep slopes near the vents, while the largest coulee thickens to merely 10-20 meters along its distal end in the flat caldera floor.

The last eruption came in 1931. It happened from the west side of the caldera. Explosions took place for 6 weeks. The climax was 10 days after the start of the eruption, when for several hours heavy ashfall produced total darkness within a radius of 100 km around the volcano. It erupted a volume of 0.9 km3 and formed a crater 700 meters wide. Lava ponded inside the crater at the end. As some other Aniakchak eruptions, it started with more silicic stuff and ended with a more mafic melt, shifting from 69 to 56% silica.

From left to right, you can see Vent Mountain, the 1931 Crater, and Half Cone. Fluid dacite lavas from the 2-kilometre long fissure cover the area behind Vent Mountain. From Google Earth.


Okmok is a rather quiet one. Seismically speaking. That’s why the relatively minor earthquake activity since late 2022 is significant. The oldest earthquakes with less than M1 are from 2007, but maybe a dense seismic network was in place before that year, only that small earthquakes did not happen. The most active years were 2008 and 2009, 2008 was the last eruption of Okmok. The third most active year is 2022, and 2023 is on its way to surpass it.

Number of earthquakes per month at Okmok in 2022-23.

The earthquakes are concentrated within 6 km of the surface under the southwest corner of the caldera, where Cone A is located, the most active historical vent. Deep earthquakes extend to 25 km depth to the south of the caldera. The deeper activity was dominant from 2009 to 2020. Instead, in 2021-2023 activity has been mostly shallow. Some minor shallow activity in 2007-2008 preceded the last eruption.

Despite Okmok being a near-continuously inflating volcano, unlike other Aleutian systems, its seismic activity seems very subdued. A GPS on the western side of the caldera has recorded about 1 meter of uplift since the 2008 eruption. Uplift will have been a bit higher towards the center of the caldera. Inflation accelerates and deccelerates repeatedly. But seismic activity has been below that of other volcanoes considered here.

The past eruption history of Okmok is very interesting. The volcano has undergone two recent caldera-forming ignimbrite eruptions. The first was at 12000 yr BP, and the second at 2000 yr BP. Both eruptions were very similar in size and style, and emitted at least 25 km3 of material each, but possibly as much as 50 km3. The 12000 yr event consisted of a minor initial rhyolite-andesite plinian eruption, which was then followed by a massive basaltic-andesite ignimbrite of 55 % SiO2. The 2000 BP event similarly started with rhyodacite, and then erupted the same basaltic-andesite ignimbrite. During the basaltic-andesite stage, pyroclastic flows covered the entire region within 10 km of the caldera rim in more than 10 meters of scoria, while dilute pyroclastic currents reached to almost 40 km distance, beyond the sea, into nearby Unalaska Island.

Okmok caldera.

Post-caldera activity is reminiscent of Aniakchak, but more mafic. More than 12 craters form a 7 km wide ring around the beating heart of Okmok. Some of the craters are larger, up to 1 km, formed in stronger explosive eruptions, while others are small, 200 meters wide, and formed in low-level strombolian activity. Cone A, at the southwest end of the caldera, erupted 9 times from 1938 to 1997. Eruptions were basaltic, 52% SiO2. Several channelized aa lava flows, around 5 meters thick each, have flowed to distances of 5 kilometres from Cone A. While strombolian eruptions must have taken place at the vent itself.

Okmok erupted again in 2008, but this time from the north side of the caldera. A more evolved basaltic andesite with 55% silica came out. At the onset, an ash plume rose to 15 km high, and the total size of the event is usually considered being a VEI-4. The eruption was completely unexpected. Despite the presence of 13 seismic stations on the island at the time of the event, no precursory activity was detected up until 3 hours before the eruption. The only precursor visible in the seismic catalogue seems to be 17 M 0.5-2.2 quakes that happened from July 2007 to April 2008. Which is so low that it couldn’t be considered alarming. Then, the eruption itself, which was in July 2008, came after 3 months of complete seismic silence. Only a few hours before the explosion did activity pick up in the area where the vent opened. But the immediate precursory sequence itself was very weak, with less than 20 earthquakes detected, none of them exceeding M 2.4. As a result, the advisory level was risen directly from green to red when the eruption was already underway.

Takawangha and Tanaga

As I write this article, a powerful swarm rages under Tanaga Island with large earthquakes of up to M 4.1, below the twin volcanoes of Takawangha and Tanaga.

Earthquakes of less than M 1 have been located near Takawangha volcano since at least 2008, however the number of earthquakes in 2022-2023 is nearly three times the total number of events located during all previous years. Activity increased dramatically at Takawangha volcano in November 2022 and has persisted up until now. Then, on 4 March, activity skyrocketed at nearby Tanaga, producing 350 earthquakes within merely six days, and has kept going.  AVO raised Takawangha to yellow on 18 November 2022 and did the same for Tanaga on 7 March 2023.

Number of earthquakes per month at Takawangha in 2022-23.

These two overlapping volcanoes consist mainly of crystal-rich basalts, basaltic andesites, and their alkalic equivalents. Takawangha is mostly eroded and little activity must have taken place in the Holocene. However, four youthful craters 200-700 meters in diameter crown the old edifice. They must have produced recent low-level explosive activity. Some lava flows have issued from the craters and reached down to near the sea. The fronts of these flows seem to be around 10 meters thick. These young lavas must be Holocene. Their composition is trachy-basaltic andesite with 54-56 % SiO2.

Tanaga is younger. Three prominent cones, with a combined volume estimated in ~32 km3, are aligned in a 7 kilometre E-W line. East Tanaga, Tanaga, and Sajaka. Tanaga towers nearly 2 kilometres above the ocean, and another 2 kilometres above the seafloor. It is partly built on the shoulders of Takawangha, though. There is very little erosion, and at least two of the three cones have erupted within the Holocene. The youngest lavas are at the westernmost cone, which is called Sajaka. Sajaka is made of basalt, with 47-51% SiO2. As a result of this mafic composition, it seems like an overgrown version of Stromboli. A small 200 meter crater at the top probably features frequent strombolian activity, and maybe occasional fountaining, fluid flows of lava, less than 10 meters thick, extend downslope on precipitous slopes, these at times will surely crumble into pyroclastic flows, and the cone itself might be at risk of avalanching. In fact, at some point, part of Sajaka seems to have collapsed into the ocean, leaving a horseshoe depression filled by a younger cone.

View of Sajaka, in the foreground, with Tanaga towering behind. From Alaska Volcano Observatory, by Roger Clifford.

Concluding remarks

So, four volcanic systems have become restless across the Aleutians. Coincidence? A lot of volcanoes are going in an out of volcanic unrest over time in this archipelago, so it being a coincidence is possible. Regardless, it is an exceptional situation to have so much unrest. As I mentioned at the start, there is not much reason for alarm, though. These four volcanoes are all remote, uninhabited. Other than flight disruption, a VEI-6 of Aniakchak is about the only thing that could do serious destruction, and that is obviously something very unlikely.

If the volcanoes do erupt, the scenario to expect should be similar to the earlier activity of the volcanoes. Four volcanoes/volcanic complexes undergoing unrest mean there are more chances of an eruption than in a more typical situation. But the activity at this point just shows an increased input of magma or hydrothermal/degassing activity, it may not lead to eruptions. Okmok and Sajaka/Tanaga have the best chances, I think. For big fireworks, Aniakchak. And if something happens in the Katmai Cluster, I think it will probably be a small ash emission, or a reactivation of Trident, but who knows. I will monitor the situation and write a post if there is a new major development.

Links to sources of information

IRIS earthquake browser

Eruption history of the Katmai Cluster:

Eruption history of Aniakchak:

Eruption history of Okmok:

Eruption history of Takawangha and Tanaga:

Click to access Geology-and-40Ar-39Ar-geochronology-of-the-medium-to-high-K-Tanaga-volcanic-cluster-western-Aleutians.pdf



72 thoughts on “The Aleutian volcanic awakening

  1. Thank you Hector

    Im Myself is doing a two part IO seriers

    • Thanks Jesper! I would love to read what you have to say about Io. Your enthusiasm will surely provide a unique perspective into Ionian volcanoes.

  2. Was not aware the lava was still so silicic at Aniakchak, the caldera forming eruption was silicic but I expected later stuff to be mostly basalt or basaltic andesite, being basically a reset for the volcano. Dacite spatter cones… at only a meter thick at the vent the flows are approaching basalt viscosity. I have seen andesite flows that erupted at low viscosity but dacite is new to me.

    I guess the lava must be pretty hot and with no crystals.

    • Some volcanoes seem to love erupting highly silicic lavas. You rarely ever see TVZ volcanoes erupting anything other than rhyolite, even after they caldera collapse. Aniakchak does erupt basaltic andesites, but it does seem to have a fixation with rhyodacites.

      I was also surprised to see how incredibly fluid the dacites of Aniakchak are. Near one of the vents the flows seem to only about a meter thick. Although they thicken rapidly downslope.

      There is an even more impressive case of silicic magmas behaving extremely fluid. Dabbahu volcano in Ethiopia has a series of fissure-fed pantelleritic rhyolite flows, with 72-73 % SiO2, that are very well preserved. Some of these rhyolites make flows that are only 5-3 meters thick, even at considerable distances from the vents, and on flat ground. I’m not aware of any other rhyolites that fluid, so it is probably a very rare thing. They average less than 2 wt% crystals, so they are basically crystal-free:

      • Bora Ale is supposed to be rhyolitoc too, but is constructed as a fluid lava cone that looks just like a basaltic stratovolcano, even some pahoehoe. But there is not much information I could find on it. It also seems to be rather old, pahoehoe from Erta Ale flows over its base, but then Erta Ale is a young shield so this might not mean much.

        Pantellerite is an alkaline rock, I know that adding sodium oxide to quartz sant is able to reduce the temperature that it melts from 1700 C to only a little over 700 C, so these flows might be fluid because of this factor or something similar. Soda lime glass is basically a form of rhyolite just very simple in composition compared to natural rocks

        • The Ethiopian rhyolites, like Corbetti, Aluto, and such, are higher in sodium than other rhyolites, even though they have the same silica contents. I think Corbetti was a high silica rhyolite, as high silica as Yellowstone or the TVZ, even though they seem to evolve from alkali basalts. But the sodium content is probably the highest among rhyolites.

          • Giant subduction zone calderas, like Toba, Cerro Galan, Yellowstone, instead, have rhyolites that are very high in potassium, but very low in sodium. They are also more crystalline usually.

        • I noticed too that the lava at Dabbahu is black, the ash is whitish but that can happen from vesiculation, but the flows are a dark colour even though very felsic. I know obsidian is black up close but most obsidian flows still have a light colour from a distance, not so here.

          I can only assume all of that couple percent of crystals are dark minerals

          It would be fascinating to see one of these fluid silicic lavas erupt. I guess we have those videos of Hekla in 1947, and fissure 17, for evolved but still fluid lavas, but those were at most higher silica andesites still not even dacite let alone rhyolite.

          • I would certainly love to see a video of Dabbahu rhyolites flowing. But even if it erupts, no one will record it, probably. The black colour might be because the lava is pure obsidian with no crystals. Unlike most rhyolites, which do have crystals. The rhyolites of Cordon Caulle, which are also crystal-free, are black. And going from memory, most of the Ethiopian rhyolites are black.

          • The Cordon Caulle rhyolites of the 1921 and 1960 eruptions are also remarkably fluid. Making flows with thicknesses of only 5 meters along the edges, and remarkably narrow channels at places. The flows of 2011 are noticeably more viscous though, with thicknesses of 15-20 meters. Cordon Caulle rhyolites are crystal free but not sodium rich..

          • Although it can’t be that simple. Chaiten erupted initially, in 2008, a crystal free rhyolite explosively, although a high silica one, unlike Cordon Caulle low silica rhyolites. The difference is enormous. The lava that came after the explosive eruption formed a flow 300 meters thick, and with a bright cream colour. Chaiten is one of those sodium depleted systems, though. I’m not sure if it could be the depletion in sodium that makes it so viscous, and impedes obsidian formation.

          • Yes, the lavas of those eruptions I would have thought to be basalt or andesite, not rhyolite.

            I was reading some stuff about lava on Hekla. Its lavas are nearly crystal free too, so I guess maybe we just dont have a lot of examples of truely crystal free lava that isnt basalt, it seems likely that all compositions of melt are relatively fluid liquids at standard eruption temperatures. The crystal free basalts of Hawaii are extremely fluid, flows are a few cm thick near a lot of vents.

          • On Chaiten, I did see an interesting idea that the flows are vesiculated welded pumice flows, the initial plinian stage stops at the surface as bombs stick to the vent walls and close off, but there is still the gas jet and explosive degassing underground. So it erupts crystal free but more like a foam, which is much more viscous than a free liquid. Same proposed for Puyehue 2011.

            But I saw this on, which is in French, so finding it might be hard…

          • Perhaps “normal” rhyolites, that have the sodium content that corresponds to their alkalinity, are more fluid. While those that are sodium depleted, even slightly, like the TVZ, Chaiten, Yellowstone, or Laguna del Maule, become far more viscous, and brightly colored. But this is speculative. There are some cases which may contradict the rule, I would have to check.

  3. So much unrest… Pick your volcano! Will the Aleutians awaken and give us another Novarupta? Will Grimsvotn or Katla finally end the ceaseless teasing and give us an amazing explosive eruption? Will Cerro Negro finally let loose after years of build up? Will Taal finally finish what it started 3 years ago? Interesting decade we have a head of us!

  4. I wonder if a slight adjustment in the subduction motion of the Pacific plate has a knock on effect on these volcanoes. It could have been an adjustment made years ago, but now coming to the fore by releasing that extra bit of magma. Maybe a subduction angle change of as little as 1°. Or maybe some extra cracks created by some of the strong previous earthquakes there has been.

    I don’t think we’ll see eruption as powerful as those of the past on the Aleutian islands simply because there’s far less of an ice cover as there was when they all started popping off 12,000 years ago. If anything big happens it will likely be around the Katmai/Trident region where multiple magma chambers could possibly merge, if one big one isn’t already feeding the multiple stratovolcanoes in that area.

  5. The two big prehistoric Okmok eruptions sound like there was a plinian throat-clearing phase ejecting old stale magma, followed by a massive pelean phase when the fresh stuff got to the surface. I suppose ring dykes were involved in these events and the lid basically fell in once enough of the support was removed. Doubt we have to worry about another one of that size for another 8000 or so years.

    I would think the two likeliest outcomes of this unrest, on a per-volcano basis, are: It turns over and goes back to sleep (intrusion ratchets it a bit closer to its next eruption, but that will take many more cumulative intrusions and decades to centuries), or it does a VEI 3-4, with an outside chance of getting a 5. 6 possible but very poor odds, 7 perhaps possible for some of them (though Okmok should need millennia more of recharging for that) but it’s probably more likely we get a substantial asteroid strike in the next couple of years than that we get a 7 in the Aleutians in that timeframe.

    Nothing in this region has produced an 8 that we know of and there’s no reason to expect that to change. Put the odds of that at zero.

  6. A great factual introduction in West-Alaskan and Aleutian volcanism!

    The map of AVO shows the geographical distribution of the volcanoes:
    It shows that on the northern side of the island chain there is a change of sea from Bering sea to Aleutian basin. The Bering sea is a shelf sea (like Baltic sea), while the Aleutian basin is a deep sea. From Akutan until the Alaska peninsula the northern side of the subduction zone is continental, while to the west it becomes oceanic. Katmai and Aniakchak are on the eastern, more continental part. Tanaga as Great Sitkin sit on the more oceanic western part.

    Great Sitkin is a historically very active volcano. Maybe it is an example of what to expect from Tanaga. This page tells what it did:

  7. I’m fully onboard the idea that the Katmai area may evolve into a rhyolitic supervolcano at some point in the geologic future. The whole area just looks like one large interconnected seismic / magmatic mess that will eventually conjoin and converge into a serious magma chamber.

    Or not, but it’s fun to speculate.

    • Fisher had 8 stratovolcanoes in a similar configuration prior to going big. It also had a history of increasingly silicic eruptions with an ever-shortening repose time.
      It seems that the magma from Novarupta was probably from a separate deeper chamber than the predominantly andesitic/dacitic Katmai. Also all of the eruptive products of Trident, Mageik/Martin and Katmai are very similar, only Griggs stands out as being a lot more potassic, which would indicate different partial melting and likely a different magma chamber.

      • Katmai has some rhyolite/rhyodacite lava flows that make up the last activity of the main stratovolcano. So I would say the rhyolite of Novarupta probably came from a shallow magma chamber under Katmai. The crystal rich andesites and dacites probably come from a deeper regional magma storage that fed the Holocene lavas of Mageik, Martin, and Trident.

        I think this andesitic/dacitic magma storage probably spans the 72 kilometre-long row of volcanoes from Martin to Kukak. A total of 9-10 volcanoes that are considered active by GVP. . The reason why I think this is that a lot of shallow seismicity occurs within 13 km of the surface along this whole chain. Probably driven by rising volcanic fluids, like water and such. But earthquakes deeper than 13 km are almost entirely concentrated in a pipe-like region under Trident that extends to 40 km depth. So I believe magma rises up under Trident, then pools in this broad andesite/dacite body at ~13 km depth. The magma is rich in mafic crystals, with an interstitial rhyolite melt between these crystals. This rhyolite melt, I think, is then channelized upwards into the shallow chamber under Katmai, or former chamber under Katmai.

        What happened on August 2022 may have been a pulse of magma rising up in the pipe region under Trident, which then activated multiple volcanoes of the chain. Since then, more magma has perhaps risen up, but now the connection might be more smooth and less seismically active.

  8. Semisopochnoi has joined the party too (barely)!
    About Katmai, is some sort of local rifting going on? Shallow earthquakes are in an 18-mile line from Martin to Katmai.

  9. Alaska is the most active of the US subduction zones, is it because crust is thinner and younger there? and its easier for the magma to reach the surface ? The belt have alot more fluid magmas than thicker continetal rim subduction belts have perhaps Beacuse there is less diffrentiation on the way up

    • Is it certain that Askja had an intrusion? That would in its own be an interesting geological event, even without the possibility to observe anything.

      Currently the Reykjanes peninsula has got more seismic activity again. Should we expect more rifting events to come there like in the “Krafla Fires” period?

      • In short yes, Reykjanes will probably be where at least half of the eruptions in Iceland will be in the 21st century. I mean, since 2020, there have been intrusions at Thorbjorn 3 times, then intrusion and eruption at Fagradalsfjall, another intrusio nat Fagradalsfjall, then another large intrusion near Thorbjorn again, another eruption at fagradalsfjall… There was also probably at least 5 intrusions at the submarine part of the peninsula, at Reykjanes, Eldey and another island further out to sea. And this is just in the last 3 years… 🙂

        Good news is that the eruptions on reykajes are nto especially big. The fast lava flood eruptions are mostly about 0.1-0.2 km3 in size, a lot like Krafla in the 80s, or Mauna Loa last year. Hengill might be able to go up to 1 km3, but it also skips most of the cycles. There are other eruptions going probably over 1 km3, but they are slow, like longer versions of the 2021 eruption.

        I would expect at least one more intrusion within this year, and if it is at Fagradalsfjall again there is a high chance it erupts, but probably more like in 2022 than 2021. It would be nice if there was new data on deformation, but I havent seen any. If one of the other volcanoes goes then there might be much more severe rifting and eruption is less certain at least in the immediate term, but best not invest in land in Grindavik I think…

        • Krafla had many rifting and intruding events without an eruption. It can be interesting to see the earth moving up and down or creating new faults. We are witnesses to the beginning of a centuries long series on the Reykjanes peninsula. Earth has time, but we humans not. We live too short to really watch volcanoes.

          Krafla has a real central volcano, while the systems (f.e. Krysuvik) on the peninsula usually consist of a rift with several possible eruption sides. According to Wikipedia the Krysuvik system has the potential for phreatic explosions/eruptions. If the Fagradalsfjall series spills over towards other Krysuvik places, it may hit explosive groundwater.

          • Krysuvik is also sort of a central volcano. It doesnt have a caldera or any mgma that isnt basalt, but there is a magma chamber or crustal storage of some sort in order to drive such intense eruptions. Same for Svartsengi.

            One of the things about Krafla too, its early intrusions were large and often didnt erupt, but eruptions beginning in 1980 became much larger while intrusions were a lot smaller. The rift filled for 5 years and occasionally leaked, then the magma had nowhere to go, except out. Only thing is, at fagradalsfjall in 2021, the initial intrusion did start a large eruption, but I think the same pattern might follow where eruptions get bigger over the coming decades, eruptions in the 2030s might be pretty spectacular.

          • OK, I had the thought that most systems there are primarily “rift volcanoes”.
            How can the magma chamber of Krysuvik survive such a long break after the last eruption 1340? Or has it already become some kind of gabbro?

          • Magma cools very slowly. The lava lake at Kilauea Iki took 30 years to go solid and us still over 500 C today in the center. The lavalake in Halemaumau now will take centuries, assuming such a time would ever happen without an eruption that long there.

            Those are not even underground. Also fluid magma would tend to settle out crystals.

            It is also possible at Krysuvik the chamber is deep, like at the base of the crust, same area the 2021 eruption started from. In which case possibly there were eruptions like that before the lava floods although there isnt any mention.

            The 3rd option is the magma chamber is a sill that grown new every time, like we have been seeing under Thorbjorn, and which eventually cracks open along the rift zone, allowing it to erupt all at once. I think this might be the most likely option actually. This doesnt bode well for Grindavik unfortunately, as this seems to be exactly what has been happening there since 2020. And in the immediate runup to the 2022 eruption there was a tectonic quake at Grindavik that broke the surface, so it might be pushed over the limit by the next intrusion.

            Or, it us all of the above or some combination. We will probably learn a great deal as Fagradalsfjall evolves in its cycle, at least that seems to be where the magma is going right now so is probably where the next eruption will be, again 🙂

          • Is the Fagradalsfjall-Eruption only the tip of the iceberg of a deeper magma accumulation? The magma chambers of Krysuvik or Reykjanestá need a hallway which feeds them with magma from the Moho. The Fagradalsfjall-Eruption was a direct eruption from the Moho to the atmosphere. It might be a small branch of a larger tree structure.

  10. So I stumbled across something on Google earth and had to share here to get general thoughts / feedback.

    I took a screenshot of what *looks* like a submerged caldera. I’m not going to say where or how large yet since I don’t want to bias people’s opinions. I have my doubts for various reasons, but to my naked eye, this looks like a large caldera with a resurgent dome in the center.

    If it weren’t for Gakkel Ridge being confirmed as an enormous oceanic caldera, I would have doubted this could be possible. Ironically, I remember seeing Gakkel ridge before it was discovered to actually be a caldera on Google earth and thinking “that looks like an enormous caldera”.

    But now with that in mind, I’m curious to get people’s thoughts from a visual perspective. I will share the size and location later.

    • What my mind sees in that image is a seamount surrounded by some kind of thick sedimentary deposit. That kind of deposit seems to cover most of the Mediterranean. Pressure from the convergence may have pushed it up around the seamount. It also seems to have risen up against the African continent. There is a small moat between this deposit and the African continental slope.

      • Sorry for revealing the location. I only read the last phrase after posting.

      • For what it’s worth, this is the Eratosthenes Seamount, just south of Cyprus. Still likely not a caldera, but there are some interesting things with it that can be found in the publication below.

        “The lithologies beneath the sedimentary succession drilled during leg 160 remain unknown (i.e., pre-Early Cretaceous). However, the evidence of the large elliptical magnetic anomaly, nearly 200 km across, which underlies the seamount (Woodside, 1977; Makris and Wang, 1994), suggests the occurrence of a magnetite-rich lithology at a depth of 2−4 km beneath the seafloor, presumably mafic or ultramafic rock (Ben-Avraham et al., 1976). Consequently, Ben-Avraham et al. (1976) suggested a genetic association between the Eratosthenes Seamount and the Troodos ophiolite. However, basalts beneath the seamount, if present, could also be associated with Lower Jurassic or Lower Cretaceous volcanics known from the southern Levant. Also, Upper Triassic and Upper Cretaceous basalts are present in southern Cyprus (see below). Seismic refraction data show that the crustal structure of the Eratosthenes Seamount is characterized by a seismic velocity layer of 6.0−6.3 km/s, at intermediate crustal depths (i.e., 15−20 km), and velocities in the lower crust are measured as 6.7 km/s (Makris et al., 1983; Makris and Wang, 1994).

        Combining the magnetic, gravity, and seismic refraction data, the most reasonable interpretation is that the seamount is underlain by thinned continental crust, about 25−27 km thick, associated with mafic or ultramafic bodies, possibly related to rifting.”

        • There is an associated elliptical magnetic anomaly. Impact crater is another possibility but I’m not sure on what they’d be looking for to identify underwater, would there still be shocked quartz etc.?

          • “should” be. Quartz shocking is a function of the impact force.

    • Would not surprise me if ocean ridges have more significant evolved volcanism than is usually expected. Well, the fast ridges are probably only basalt, but other slow ridges that are like Gakkel might be hiding some monsters. Even Iceland, which can be considered the most active segmebt of ridge, has got quite substantial silicic centers.

      Although, to be fair, most of the silicic volcanism in Iceland us associated with more plume related volcanoes, Reykjanes has only one center with something that isnt basalt. I guess though Iceland istechnically also just a bit of the Gakkel ridge that got a boost, so perhaps massive magma accumulation should be expected just on that alone.

      • I think felsic volcanism is practically non-existent in mid-ocean ridges. Much of the mid-ocean ridge system has been dredged and studied for chemistry. The mid-ocean ridges are almost entirely tholeiite basalt, although you can find a few very rare occurrences of andesite. But the first rhyolite in a spreading ridge was only found 5 years ago in the Gulf of California:

      • Pancake shaped shield volcanoes with small calderas are very common in fast-spreading mid-ocean ridges. But they don’t have rift zones, and I have only seen them on the flanks of the ridges, not along the axis. So they seem to be off-rift volcanoes. As far as I know, they are always made of tholeiite basalt.

        Rare calderas also happen along the rift axis in regions where the magmatic activity seems boosted. There is Boomerang Seamount (St. Paul Rise), Axial Seamount offshore Oregon, a caldera in a triple junction of the North Fiji Basin, at least 3 calderas in the North Lau Basin, and 2 calderas in the Galapagos Ridge. I think they are all basaltic too.

        The calderas in the Galapagos Ridge:

        • Unfortunately, only Axial Seamount is being studied. All of these calderas must be very active volcanoes.

        • Gakkel seems like a unique structure on the planet. Looks to me like it is located in the propagating tip of the ridge. But a shallow, remarkably linear graben does continue 400 km to the south of Gakkel. I wonder if this graben is formed by giant dikes propagating from Gakkel caldera. And if Gakkel is a caldera, which seems very likely, it is the largest caldera in the world. It is twice as big as Yellowstone and Toba in terms of area. It is a shame that practically nothing is known about it.

          • There are also several calderas centrally aligned in the African rift, which may be Continental-Continental but is the same principle. Awassa for instance has erupted pre-caldera and post-caldera alkaline basalts, but also peralkaline rhyolite.

            If I remember rightly silica is more concentrated in volcanoes on or near plate boundaries due to contamination.

    • I’ve seen it myself on Google maps. I was wondering if it might be a very large mud volcano?

      The structure looked more like mud had been sucked into the central conduit from all around, then spat out upwards. The mud volcano in Java which started erupting in 2006 is still going and according to the wiki is expected to continue erupting for another few decades.

      It’s hard to find the volume erupted so far. One number I saw is 30 million m3 per year, which if is steady would be about 500 million m3 so far. Approaching VEI 5 if you could use that scale for a big muddy mess!



    No, that’s not what you think. #Musk hasn’t been to Mars yet, and there are no elephants there, I think. Also, #Perseverance is just a rover on a nuclear diet, so, no… This must be some kind of rock. But #IANAG™(*)

    Processed, cropped MCZ_RIGHT, FL: 110mm
    Sol: 743, RMC: 37.0000, LMST: 11:37:48

    (*) I Am Not A Geologist™

    #Mars2020 #Solarocks #Space

    • If a mod could remove the first embedded image (raw image linked on NASA site) and just leave second processed embedded as I hadn’t intended to have both processed and the NASA server raw image embedded.

    • how did this unusual rock get here ? speculation on its history, anyone?

    • Hahhah dog poop

      Looks very much what the 85 old neighburs dog leaves every night on my dads backyard. That old man have set this into system as well, letting the black labrador leave a pile every night in the yard, when my father are deep asleep. Every morning dad throws the waste into the forest with a showel, yet its there next day again. The neighbur is too stubborn to be negotitations between, so we plan to make the ground unattractive for the dog, like pouring out salt or infrasonic sounds.

      Its little of a comedy here, a real world comedy film 🙂 and very much so, been ongoing for so long now that its pretty much only funny

  12. Shallow earthquakes along the northeast rift of Mauna Loa. Any relation to the rapid inflation?

    • HawaiiTracker podcast had them down as south flank quakes, related to spreading but not intrusion. Intrusions in Mauna Loa apparently make quakes on both flanks as it isnt buttressed the same way Kilauea is, at least not as high elevation.

      So not exactly a signal of a new eruption on the way but maybe related to the one just passed. But then at the time of that podcast the quake was under an hour old so very preliminary.

  13. So GeologyHub posted a video about the two plus decade long intrusion at Three Sisters mentioning how the USGS now believes the intrusion is basalt in composition due to depth (6km) and presumed temperature, along with the pattern of surface displacement. My question is; the way he explained this made it sound that the magmatic composition of the intrusion will be the composition in a potential eruption. But don’t these volcanoes that have erupted evolved magmas have pockets or chambers of older, staler magma with evolved compositions that could potentially be reactivated by a hot basalt intrusion? Totally different kind of system, but that’s what primed the last Yellowstone rhyolite supereruption; an intrusion of hot basalt.

    I gather every volcano is different but I get confused by this. Eruptions can trigger in different ways so why in this case would the assumed basalt intrusion need to breach the surface; couldn’t it technically hit a pocket of andesite or dacite and “overcharge” it into an eruptive state? Not that it will, just as a hypothetical.

    • Don’t take video so seriously, Geologyhub is good at making videos about news and reports but he has a bad rep for making exaggerated predictions, (Katla video and more). A basaltic intrusion doesn’t mean that an eruption is any closer. This just looks like the typical minor resurgence and unrest that commonly occur at evolved systems. The uplift rate is unimpressive and the lack of seismic instability doesn’t make me worried.

      In any volcano, that I look at for the potential of any sort of eruption in the near future, especially from long-dormant systems is progression. Look at Campi Flegrei, Tatun, and CCN, they’re not just producing uplift and minor quakes. Campi Flegrei’s deformation and seismicity have all accelerated over the past 10 years with sustained hydrothermal disruptions. The unrest has slowly but surely intensified more and more and this indicates that the system is growing more unstable. This is a small amount of progress in the sense that we haven’t seen any evidence of magma reaching the surface anytime soon. This is the reason why I am so invested in CCN. In the past 10 years, the volcano has produced 4 swarms and each swarm so far has brought the system closer to erupting. The first swarm was the result of an intrusion, the second swarm was the result of the magma chamber pressurizing, and the third swarm was the result of magma shallowing to 2-4 km below the surface. There’s been no confirmation of what the cause is but my bet is that a plug is weakening and volcanic vent is being formed.

      • I mean, at least he seems to be getting a little better in recent times. He did say that the swarm is of very little importance or concern by far and put that at a 3 on the “restlessness” scale. He did also somewhat “forecasted” that big eruption from Semeru, in spite of a lack of sources which lead us to kinda doubt him for a while (and that not mean that he “predicted” the eruption. It is just weird that he would do that).

        I am not saying that he is right on everything, but he’s getting better. I even wonder if he had heard of Volcanocafe here and checked up here time to time. I wonder if he’s one of us…

        I do agree, though, that we shouldn’t ignore the other “big ones” and that we shouldn’t exaggerate at all, though we should look at the scenarios in case the worst comes. I could imagine him making some big claim that he (or other media) says that Rainier is going to blow. People get terrified and go “doomsday” mode and, when the time comes, the eruption is minor and that it only produced a few lahars.

        Those same people who prepped will be disappointed, as they spent a lot on there (although not lost on resources) to the point they couldn’t, let’s say pay rent (although, they may turn self-sufficient thanks to the panic) and losing trust in the media, which is supposed to let us know what is truly happening. That tourism money will be wasted and the economy damaged. That is what I think the worst case scenario is in terms of human behavior, aside from the volcanic side.

        I could see, though, that if Geologyhub is half-right (if it is the case) and that the eruption is basaltic, that might boost the tourist industry a bit (if not accounting for the wildfires started) as some people are attracted to these sorts of things.

        But, again, we should not ignore the ones that could do some major damage and destruction. I do remember this one saying:

        Inaction (or ignorance) can be more dangerous than action. (And also that false facts can be very dangerous).

      • I find that he is good for a basic bit of info. If he talks about Hawaii for example then I find what he says is usually a bit basic and follows the HVO first page data. But then, that is likely the level of information that he goes to for every volcano, which is more than I know about most other volcanoes, so it is insightful and interesting. There is nothing in these videos that gives an inference of lying in the way those pseudoscience channels do, GeoHub is at least grounded in well founded data supported by the USGS etc. He is much more conservative in this regard than everyone on VolcanoCafe is about our favorite volcanoes 🙂 which is sensible as his sub count is likely more, or at least comparable to the view base of VolcanoCafe, but most YT videos are only 40% subscriber views, more or less.

        At Semeru I would expect his prediction is a slightly erroneous interpretation of lava flows building up on the flank as a growing dome or a flank bulge. Semeru lava is not really viscous enough to form proper domes but does build up on the steep flanks and periodically collapses all at once, it did it in December last year and a year to the day in 2021.

        Same thing for Just Icelandic, as he has been dragged around a bit in this topic before. To me hes just an interested fan of the subject like us, speculating a little bit and exploring Iceland giving some fantastic photography. His recent videos in particular are often directly referenced to the public data. Live above, I dont really know why there was any hate, his videos are far less speculative than most articles published here, even those by Carl who is very knowledgeable on the subject, but delves deep into speculation like us 🙂

        If you take them as absolute authority then yes there is a problem but if you watch it as a source of information I see nothing of a problem at all. I watch them both, if they say something I am interested in then I will look at the data myself.

        • I know I have been bashing on Just Icelandic, but I wouldn’t call it hate. The problem is that he gives people a lot of wrong ideas, which is apparent if you look at comments people who have seen his videos are dropping in social media. There’s a mix of good content and wild speculation and worst case scenarios, all mixed up so average viewers can’t really tell the difference. Even more problematic is that in the last video I saw, he was implying that scientists were hiding information and “sweeping it under the door”. That brings him one step away from Daily Mail land and one step closer to conspiracy land. Let’s hope that was just a one time thing and that he stays off the tin foil path. He’s got obvious talent for generating graphics and video. I just wish the facts were a bit more on target.

          • I mean, to be fair there are probably a lot of people who do the same thing and get the wrong ideas after reading articles here… It isnt really any different, and there are probably more people who read VolcanoCafe than watch Just Icelandic. I wouldnt be surprised if he has actually stumbled upon this place at some point 🙂

            I guess, those people who take his videos as absolute fact are not his fault, they will do that no matter what and probably arent willing to fact check anything they do. At least it seems like Just Icelandic does consult with the experts where possible.

          • VolcanoCafe is not perfect, but both main articles and comments try to be as rooted in science as possible. We also do wild speculations all the time, but we always try to be clear about it when we do. The dialogue also helps. Wild claims will always be disputed or challenged by someone else. The strength of VC lies in the community.

  14. Very interesting piece, Héctor. Commenting late as not available when you published it, but interesting enough to catch up.

Comments are closed.