Magical Mount Fuji

The story begins with Japan’s iconic volcano. Mount Fuji has become the poster child of volcanism – it is what every volcano strives to be. This is the most climbed mountain in the world. The spring-time view of Mount Fuji framed by cherry blossom is the very emblem of Japan. The reflection of Fuji is printed on the backside of the 1000-yen bank note. Wedding ceremonies are held at the temple at its summit. The summit is Japan’s highest point. How did such a perfect volcano form?

Mount Fuji is where the Pacific goes to die. The western half of Japan is continental by origin, at one time part of the Asia. The eastern half is younger, with volcanic arcs adding new land and expanding the nation towards the sea. Out in the ocean is the subduction zone where the oceanic plate sinks into a trench and dives below the land. Behind Japan, the Sea of Japan has opened. It has left Japan as an island fighting trench warfare with the ocean, while protecting the continent behind it. The conflict involves earthquakes and volcanoes, whilst the multitude of hot springs are signs of the heat of the battle. This post is about a four-way battle for the heart of Japan, centred around its very own Lonely Mountain with the Five Lakes.

Mount Fuji

As the crow flies, it is 100 km from the centre of Tokyo to Mount Fuji. Driving to the slopes will take you 2 hours (do be aware that the access roads can be closed to private cars during the busy season), or you can take public transport. There are four main trails up the mountain. Each trail is divided into ten stations. The road or bus will take you as far as the 5th station, and the main climb is from there to the top. The summit, at 3776 meter, is the 10th station. The climb is on well-trodden paths and is of only moderate difficulty. For ease of access, there are separate paths for the ascent and the descent. The altitude is considerable, though, and you will find yourself short of oxygen.

The trails can be busy: a quarter of a million people climb it each year – this number has been counted at the 8th station. But the climbing season is short , and all those people are packed into the short period from July to early September. It is very much a community crowd activity, part of Japan’s social fabric. (However, until 150 years ago women were forbidden from climbing the mountain.) During the peak of the holidays, mid August, expect walking traffic jams. If you feel tempted to go for the quiet time outside of the climbing season, don’t. There will be snow and high winds, and the mountain huts (and the trails themselves) are closed. Only the climbing season is snow free. The climb up takes anywhere from 5 to 8 hours, also depending on which trail you take. The walk around the summit crater to the actual summit takes an hour. The return is 4 to 6 hours. To be on the summit at sunrise (4:30 am) requires booking a sleeping place in a hut. In fact, this is recommended for any trip to the summit. Remember to carry money (nothing is free (or cheap), including the emergency oxygen). Wooden walking sticks can be purchased at the 5th station: they make good memorabilia afterwards.

One of the famous 36 views of Mount Fuji, by Hokusai

Fuji’s fame comes both from its location in Japan’s heartland and from its near-perfect shape. Hokusai’s 36 views of Mount Fuji are an excellent example of how Japan’s art is dominated by the beauty of this volcano. But beauty requires maintenance. Mount St Helens used to have the reputation of being the Fuji of America. How different it looks now, after the destructive eruption. But soon, probably within centuries, it will rebuild its cone, resurface its flanks, and re-enter the volcano beauty contest, the ultimate Miss World. Ragged volcanoes are a sign of erosion, when erosion wins out and flank collapses leave holes that do longer get filled. Age builds character but not beauty: a symmetric mountain is a sign of youthful activity. Fuji clearly must be a regular eruptor, an applicant of frequent volcanic make-up. Indeed, it has had two major eruptions in written history.

There are hidden surprises in this familiar volcano. Fuji is actually a conglomerate of five volcanoes, three of which can be recognized in the landscape. On the northern slope, 7 km from Fuji, there are outcroppings of an older volcano, called Komitake. Underneath Komitake lies a still older volcano, buried by Komitake, called Sen-Komitake (meaning ‘before Komitake’). The extinct Ashitake volcano, 20 km southeast of Fuji, forms a deeply eroded but recognizable hump in the landscape. (This is a distant member of the Fuji family and perhaps is not directly connected.) Ashitake (ojiisan) is the oldest, and was active around 400,000 years ago. Sen-Komitake started erupting around 270,000 years ago. It ended with a series of explosive eruptions 160,000 years ago. After this self destruction, the new Komitake grew until is ceased erupting about 100,000 years ago. Now another volcano took over, Ko-Fuji. This new volcano had an explosive flank collapse about 18,000 years ago. Following that collapse, modern Fuji (called Shin-Fuji) arose just west of the old summit and eventually buried its predecessor.

The next surprise is that this perfectly symmetric volcano actually has a severe case of acne. The flanks are covered with 100 parasitic cones and craters. Beauty rarely survives a look with a magnifying glass.

The final surprise lies in the magma. Eruptions in and around Mount Fuji have mainly produced tholeitic basalt. Basalt is not common in Japan, and there is clearly something different about Fuji. Ashitake was andesitic, a common lava type of Japan. But each of the later volcanoes of Fuji started with effusive basaltic eruptions, over time evolving towards explosive andesite/dacite eruptions after which the reset button was pressed and a new volcano took over with basaltic output.

The current volcano, Shin-Fuji, started with voluminous basaltic eruptions between 17,000 and 8,000 years ago. After a quiet interlude with less activity, a mix of large explosive (pyroclastic) and effusive eruptions took place between 5,500 and 3,500 years ago, both from the summit and the flanks. By the end of this phase, the mountain had reached its current height. 2,900 years ago there was a major flank collapse with a land slide volume of 1 km3. It is not known whether this was volcanic or was triggered by an earthquake. Now a series of summit eruptions followed, the last one of which was 2,200 years ago. Since that time, Fuji has been less active, with two major eruptions and 8 small eruptions, all on the flanks.

The two major eruptions of the past 2000 years were a large effusive eruption in 864 and a plinian explosion in December 1707. The 1707 eruption was triggered by an earthquake, not just any earthquake, but this was the Great Hoei earthquake. It had an estimated magnitude of M8.6 and came only a few years after the ‘minor’ M8.2 Genroku earthquake on 31 December 1703. These huge earthquakes are another aspect of Japan, and they can apparently affect Fuji’s magma chambers.

The 1707 eruption came 7 weeks after the major earthquake. Volcanic rumblings had started already in February 1704, suggesting the first trigger was the 1703 Genroku earthquake. Mount Fuji is thought to have two main magma chambers, a dacitic one at 8 km depth and a basaltic chamber at 20 km. Models suggest that the deeper chamber was affected by the quakes, and that after the first event magma may have risen from there to the shallower chamber. This pressurised the upper chamber and after the second jolt it exploded. Between 16 December 1797 and 1 January 1708, there were three main explosions. Together they amounted to a VEI-5 eruption of about 1.6 km3 of ash and tephra. The ejecta covered a wide area which included modern Tokyo, severely affecting the local agriculture. The eruption left three considerable dents in Fuji’s flank, at altitudes of 2100 to 3000 m. The largest of the craters is 1 km across. The eruption produced no lava, but magma accumulation under the surface created a notable bump which nowadays is called Mount Hoei. (Hoei is the name of the era from 1704 to 1711. Genroku is the name of the preceding era.)

An atmospheric non-eruption at Mount Fuji

The other major historical eruption started in June 864. This was a mainly effusive eruption which produced about 1.4 km3 of lava. It erupted at lower altitude on the northwestern flank, from a 6-km long fissure. The eruption lasted two years. The lava buried an existing lake, and the interaction with the lake water formed some pillow lava and a phreatic explosion. The area is now deeply forested and is in part inaccessible due to the terrain and the trees!

The smaller eruptions were also flank events, in some cases on both the northern and southern flank simultaneously. Eruptions in the years 800 and 1083 were mainly explosive. Eruptions in 937 and 1033 produced significant lava flows, with lengths of 17 km and 7 km.

The average eruption rate of Shin-Fuji over its existence amounts to 5 km3 of magma per 1000 years. It may already be in a declining phase. The older eruptions were mainly effusive whilst explosive eruptions are a more recent phenomenon. The rate is very much higher than that of other volcanoes of the Japan arc. There is something about Fuji.

A repeat of the 1707 Hoei eruption has the potential to cause major disruption. If the ash were to be blown towards Tokyo, after two weeks a few centimeters of ash would have fallen in the city area. If it rained, this would be enough to make roads impassable for all but four-wheel drive vehicles, stop trains from running and close the airports. Power outages are likely even at 0.3cm of ash and rain and mobile phone masts may also cease to work. Food supplies would be disrupted. This is a worst case scenario, and a two-week event would give time to keep on top of the ash. We have seen the impact of ash in the La Palma and Etna eruptions, but never in a city on the scale of Tokyo.

At the moment there are no indications of any imminent eruption. But Fuji is active. There is a steady rate of low-frequency earthquakes, located a few km northeast of the summit, at a depth of 15 km. This is suggested to be a dike that is receiving new magma, plausibly from the main magma chamber at 20 km. This is probably normal activity for Mount Fuji and does not presage an eruption. Shallower magma intrusions are likely to cause notable earthquake activity, as was the case in 1704. Any resumption of eruptions is therefore likely to give us years of warning. The quarter of a million annual visitors are safe from the magma. Just be cautious with the weather.

There is something about Fuji. The picture shows three volcanoes in the region: Fuki, Izu-Oshima and Miyakejima. This figure is taken from Aoki et al, 2019, Earth-Science Reviews, 194, 264. The first two volcanoes are elliptical in shape while the third one is round. The ellipticity indicates that Fuji and Izu-Oshima are affected by the local stress in the rocks. They are children of plate tectonics. Mount Fuji is also unusual in showing evidence that its eruptions can be triggered by earthquakes. To really understand the mountain, we need to look at its faults. And Japan has an abundance of those.

Faults of Japan

Japan is riddled with over 2000 active faults. The two most significant in-land faults are the Itoigawa-Shizuoka line, which runs north-south across the country just west of Tokyo and the Median Tectonic line which follows the axis of Japan from Tokyo southwards. The latter is a strike-slip fault. The destructive 1995 Kobe earthquake happened on a side shoot of this fault. Mount Fuji is located around 50 km from the intersection of two main faults, although it is not associated directly with either.

The most important fault lines of Japan. Itoigawa-Shizuoka tectonic line; Tanakura tectonic line; Median tectonic line. Source: Bor-ming Jahn, Masako Usuki, Tadashi Usuki and Sun-Lin Chung. American Journal of Science February 2014, 314, 704-750

The faults are related to the battling plates. Did I mention the four-way battle for the heart of Japan? That was the simple version. The picture below gives the actual situation. Japan is sitting on two separate plates. The northern half of Japan belongs to the Okhotsk plate, a mini-plate which also contains Kamchatka and the Kuril Islands. At one time the Okhotsk plate was thought to be part of the North American plate. Now it is seen as a remnant of the extinct Kula plate, one of the plates of the Pacific ocean. This remnant carried an ancient oceanic flood basalt and this helped it to avoid subduction. The southwestern half of the Japan belongs to the Amur plate, a break-away part of the Eurasian plate which runs from Lake Baikal to Japan. On the eastern side of Japan is the old Pacific plate, and to the south is the younger Philippine Sea Plate. Finally, between the Philippine plate and the Yangtze plate (part of Eurasia) is a microplate called the Okinawa plate. It owns the tip of Okinawa, the southernmost of the main islands of Japan. So Japan lies on three plates, and borders two more. No wonder it is so earthquake prone. (In fact one sliver of Japan, the Izu peninsula, sits on the Philippine plate, so that makes Japan worth four plates. Of all the continents, only Eurasia has more.) It really is the Battle of the Five Plates. And it is centred on the Lonely Mountain. Mount Fuji.

(The precise location of the boundary between the Okhotsk and Amur plates remains under discussion. GPS measurements indicate that the two move independently and should be considered as separate plates but it is not easy to trace the precise line of separation.)

I have left out one more detail. Honshu, Japan’s main island, originally started out as a double. It split off from Eurasia as two separate fragments, one on the Okhotsk plate and one on the Amur plate. Only later did the two meet and join forces. The suture of the two runs near the Itoigawa-Shizuoka tectonic line. The rocks on either side are very different, with granite half a billion year old on the southwest side and young sediment on the north side.

In between the Itoigawa-Shizokua and the Takankura tectonic lines lies the Fossa Magna: the central valley stretching from the Pacific to the Sea of Japan. This is a wide area which includes the Kanto plain of Tokyo. It contains up to 6 km depth of sediment from the sea which separated the two halves of Japan. The Fossa Magna is known as the rift valley of Japan.

Fossa magna: the rift valley of Japan

Most of the major earthquakes of Japan come from the two main subduction zones, which form two deep trenches along the east (Pacific) coast. The Japan Trench is where the Pacific plate dives down underneath the Okhotsk plate. This trench follows the Japanese coast (at some distance) until Tokyo: south of Tokyo the trench bends further away from the land. From here on, the Nankai Trough takes over: it comes from the subduction of the Philippine plate, and it follows the coast rather closer to land than the Pacific plate did. There is a short boundary between the Pacific and Philippine plate which is called the Sagami Trough: it runs east-west, southeast of Tokyo.

Source: wikipedia


Japan is famous for its earthquakes. 20% of the world’s largest earthquakes (magnitude 6 or higher) occur in the country. The 2011 Tohoku earthquake is well remembered: it left 20,000 dead and 360 billion dollars of damage. Surviving the earthquake itself was only half the story: the devastation came from the tsunami afterwards. The Great Kanto earthquake of 1 September 1923 (next year this will be a century ago) with magnitude M7.9 destroyed more than half of all buildings in Tokyo and caused the death of 105,000 people, mostly from the fires after the quake. The M6.9 Kobe earthquake of 1995 hit in an unexpected place, killed 6,000 people and caused immense damage. The Great Hoei and Genroku earthquakes of 1703 and 1707 were already mentioned. In 1792, a comparably small earthquake (small for Japan, which suffers an M7 on average once a year and an M8 once a decade) with magnitude M6.4 caused a flank collapse of Mount Unzen, an active volcano. The landslide swept up a tsunami that killed 15,000 people. In 1891, an M8 struck Gifu, in-land. Nothing was left standing in the city. And there are many more historical examples of major earthquakes, often accompanied by towering tsunamis.

All of the plate boundaries can produce major earthquakes. The 2011 Tohoku earthquake ruptured much of the Japan Trench. (What actually fails in such a megathrust earthquake is the contact layer between the subducting plate and the overlying plate. The two lock, the subducting plate pulls the overlying plate with it and in the process pushes it forward. When the lock fails, the overlying plate shoots back in a large quake.) The Japan Trench tends to fail in single, huge events, perhaps 500-1000 years apart, as it did in 2011. But in between there are plenty of smaller (M7 or M8) events as well. The 1707 megathrust earthquake ruptured 700 km of the Nankai Trough south of Edo (near Tokyo). The M8 earthquakes in 1944 and 1946 too were on this fault. The Nankai Trough often fails in doublets, producing two major earthquakes every 100-150 years. The 1703 Genroku earthquake ruptured a large section of the Sagami Trough. The 1923 Great Kanto earthquake also was on the Sagami Trough, but closer to Tokyo.

And the other side of Japan is not safe either. An M7.6 earthquake in 1833, possibly originating on the boundary between the Amur and Okhotsk plates, hit the northwest coast of Japan with a large tsunami. The M7.7 Hokkaido earthquake in 1993 also came from this region. And as mentioned, there are also large earthquakes on the other faults, especially the Median Tectonic Line. Japan has shaky foundations.

Plates of Tokyo

Let’s have a closer look at the Tokyo and Mount Fuji region. The plate boundaries are complex here, and the precise location is not always known, especially on land. The dashed lines drawn below indicate such uncertain locations. The arrows show the approximate directions of the plate motions. Tokyo and the Tokyo Bay are on the Okhotsk Plate. But the Philippine Sea Plate is not far away, and underneath the Pacific Plate is heading for Tokyo.

The Philippine and Pacific plate are both subducting. You may imagine some competition here, about who can subduct better. The Pacific plate is older and colder and therefore is the densest. The Philippine plate is younger and warmer and is not as dense. Thus, the order of priority is set: the Pacific plate subducts under the Philippine plate which subducts under the Amur plate. There is a layer cake under southern Japan, where the Amur, Philippine and Pacific plate are stacked. The situation further north is easier as the Pacific plate subducts directly under the Okhotsk plate, meaning there is one less layer to worry about.

If we focus a bit more on the Tokyo, area, we find more complications. There is a large bend in the Philippine plate, creating an extension into the area of the Izu peninsula, between Suruga and Sagami Bay. This is the result of a volcanic arc, a line of volcanic islands stretching into the Philippine Sea called the Izu-Bonin arc. As the plates moved, this arc collided with Japan. Volcanic islands do not subduct willingly: instead they tend to join the land they collide with. The arc also brought a larger volcanic block with it which became the Izu peninsula. This is how Japan grows.

The Izu peninsula, land that was plastered unto Honshu, contains several eroded (extinct) volcanoes. Hot springs also tell of a volcanic history. It is a mountainous land with a spectacular coast line. If climbing Mount Fuji was a step too far, a visit to the Izu peninsula may help you regain your sense of wonder and your breath. And the volcanoes here are much lower, thanks to age and erosion. (The area is also full of golf courses, if that is a sport you enjoy and can afford. The old calderas should make for an easy hole-in-one. Not to be attempted during an eruption.)

The Izu peninsula

Remember the layer cake? It is slightly more complex than I said. Underneath Okhotsk you should only find the Pacific Plate layer. But the Okhotsk Plate is itself moving southwestward – not fast, but measurable. So it is overrunning places where the Philippine plate used to be, with the Pacific plate underneath it. Because of that overshoot, the Philippine/Pacific layer cake does extend underneath the Okhotsk plate. In addition, as recent as 3 million years ago the Philippine Sea Plate was moving north, before it changed to northwest, so this added to the complex layering. Below the surface, the Philippine Plate extends further than you might expect.

The double layering has been detected underneath the metropolitan area of Tokyo. The subducting Pacific plate here lies at a depth of 80-90 km. But there is also a transition zone at 20 km depth, and this is thought to be the top of the Philippine Sea Plate. It becomes even shallower further south, and lies at around 10 km deep underneath the entrance to Tokyo Bay. Follow this 10 km contour, and you’ll find that both the location of the 1923 and 1703 earthquakes are on this line. These were shallow megathrust earthquakes close to Tokyo. The location and depth explains the damage they did.

It provides a warning. An exact repeat of the 1923 earthquake may be a century or more away, as the quake has resolved the stress here in the only way it could. But further east the boundary may have remained locked for 300 years. The Tohoku 2011 earthquake which ruptured the boundary between the Pacific and Okhotsk plates left the Philippine plate untouched and therefore also did not resolve the stress here. A large earthquake in Tokyo in the next 30 years is far from certain, but the danger should not be neglected. The probability of an earthquake causing ground shaking larger than 0.9g in the Tokyo metropolitan area within the next 30 years has been estimated at around 30%.

Fuji and the Arc

Close to the Izu peninsula stands another volcano, one that is often ignored. Mount Hakone is famous for its large and scenic crater lake, but is perhaps best known for its beautiful views of Mount Fuji. The caldera is 10 km wide (the lake is located along part of the caldera wall). It overlaps with a second one, almost as large. Twice, a large stratovolcano blew up here; the last time this happened was 50,000 years ago. Since that time, smaller cones have grown inside the caldera. The last eruption of one of these cones was 3000 years ago. There are traces of a few smaller, phreatic explosions. That was dramatically confirmed in 2015 when there was such a phreatic eruption. Luckily, this eruption gave sufficient warning and the site had been closed to tourism.

Fuji, from Hakone

In its time, Hakone produced lava ranging from rhyolitic to basaltic, the entire suite of possibilities. The mountain may not have been much smaller than Fuji is now, produced a similar (but not identical) range of lava, and it is only 30 km away. It seems to be a member of the Fuji family. However, it is the black sheep of the family as it twice explosively self-destructed. The other family members have had large eruptions and flank collapses, but they have avoided VEI-7 type explosions.

Hakone lies along the line of the accreting Izu volcanic arc and seems to belong to it. It also lies close to Mount Fuji. Does this indicate that there could be a relation between the volcanism of Fuji and that of the approaching arc? The behaviour of Mount Fuji is a result of many things: this is an extraordinary complex region. But Fuji is clearly affected by the accretion of the Izu arc. First, the ellipticity of the mountain is along the direction of the Izu arc, and is the same as that of Izu-Oshima, the largest and closest island of the Izu chain of volcanoes, located in Sagami Bay. Second, the lava composition points at a relation to the Izu-Bonin arc. The arc shows a mix of tholeitic basalts and rhyolite, very different from the andesite of northern Japan. Mount Hakone shows a similar range. Fuji shares the basalt but it is not documented as having produced rhyolite: instead it provides us with occasional dacite and andesite. You could see Mount Fuji as a mix of Izu-Bonin arc and more traditional Japanese volcanoes while the intermediate Mount Hakone is more similar to the Izu-Bonin volcanism.

The source of the Izu-Bonin magmas is the melt of the subducting Pacific plate, with added melt from the Philippine plate crust. For Mount Fuji, it has been suggested that most of the magma also originates from the Pacific plate, but moves through the overlying layer of the Philippine plate without this adding much to the melt. This forms the magma chamber at 20 km. From there the magma moves up to the 8-km reservoir where it evolves to the dacite/andesite composition – if given enough time.

A look at the maps above shows that it is not so clear which plate Mount Fuji is on! It does not help that the volcano family has obliterated all surface rocks. At depth Fuji belongs to the Philippine plate but all three plates put in bids for this real estate at the surface. (They may have missed their chance. Some 20 years ago the government returned the title to the summit and surrounding area to the original owners. It is now private land.) Mount Fuji is located not far from the triple point between these plates. However, triple points are not normally the location of major volcanoes unless they are caused by a hot spot – which Fuji is not.

The precise location with respect to the triple point is probably not too important. The main aspect is that it is at the end of the accreting volcanic island arc and that is mixing behaviour from the arc and from other Japan volcanoes.

The volcano of Izu-Oshima. Pretty, but no Fuji

The high magma production rate, 100 times that of other Japanese volcanoes on Honshu, is not so easily explained. Izu-Oshima, roughly 100 km away, is the nearest frequently erupting volcano on the Izu-Bonin arc. It erupts every 30-40 years, and has a subsidiary volcano just below sea level which also has left deposits on the neighbouring islands. But it’s eruption volume is dwarfed by Fuji. In between lies Hakone but this erupts rather little, apart from the occasional VEI-7. Otherwise the region in between Izu-Oshima and Fuji lacks active volcanism. Either Mount Fuji is collecting magma from a large area, or it has a separate cause of melt.

The melt may be affected by the Fossa Magna in which it is located. This has the appearance of a rift valley, where the low stress allows for easy magma melt and accumulation. Imagine a volcanic arc entering a rift zone! And perhaps from this mix of cultures, Mount Fuji grew and became the unexpected icon of Japan. Japan is not known for encouraging cultural melt, but volcanoes make their own rules.

One can speculate further. Three million years ago, the Philippine plate changed direction from north to northwest. Before that time it would not have affected the region of Mount Fuji. Perhaps the plate has only recently reached this region, pushing down the Pacific plate in the process. Could this have triggered the melt and initiated the volcanism, 400,000 years ago?

There is much we don’t know about Mount Fuji. On the outside: it is the most beautiful, recognizable and touristic volcano on Earth. On the inside, it remains a bit of a mystery. Eruptions are rare enough that we are likely never to see one in our life time. But they are not that rare, with perhaps a 10% chance in the next 30 years, and when it does erupt there may be impacts as far as Tokyo.

But that is not the main message. Mount Fuji is a tectonic sign post where five of the most important faults of Japan come together. Only the Japan Trench keeps its distance. Fuji is not just a poster child for volcanoes. It stands as a reminder of the other powers of nature to which few nations are as exposed as Japan. The reverence for Mount Fuji shows a deep respect for powers we cannot control.

Albert, 1 September 2022


  • Climbing Mount Fuji:
  • Yosuki Aoki et al., Recent progress of geophysical and geological studies of Mt. Fuji Volcano, Japan. Earth-Science Reviews, Volume 194, July 2019, Pages 264-282
  • Yanagisawa, H. and Goto, K. Source model of the 1703 Genroku Kanto earthquake tsunami based on historical documents and numerical simulations: modeling of an offshore fault along the Sagami Trough. Earth Planets Space 69, 136 (2017)
  • Christine Chesley et al., The 1707 Mw8.7 Hoei earthquake triggered the largest historical eruption of Mt. Fuji. Geophysical Research Letters, 39, L24309 (2012)
  • Earthquake source fault beneath Tokyo. Hiroshi Sato et al, Science, 309, 463 (2005)
  • Kazutake Mannen et al., Chronology of the 2015 eruption of Hakone volcano, Japan, Earth, Planets and Space, 70, 68 (2018)
  • Ioan McIntosh et al, Past eruptions of a newly discovered active, shallow, silicic submarine volcano near Tokyo Bay, Japan. Geology (2022)
  • Fossa Magna
  • 330 thoughts on “Magical Mount Fuji

    1. The eruption continues in Kilaueas summit caldera Halema’uma’u: lava is flowing into it constantly non stop: Kilaūea is rather like an unstoppable Tap of liquid rock.

      As the lava flows into the lava dam the crust on top rises constantly.

      Anyone knows When the caldera will be completely filled up and the lava flows will flow into Kau Desert?

      • Overflow into the wider caldera. Overflow into the Kau desert in 2028, at current rates.

        BUT, reach 1000 m/1 km elevation way sooner (by 2025), and that is when the 2018 lake proved too much to hold. 1823,1832,1840,1868 lake drainouts also probably happened at a similar elevation. I think to fill the caldera and overflow there would need to be a lava shield. As it is now, and also back in the first half of the 19th century, it is a massive lake of liquid, over 1 km2 in volume by the point of overflow, so rather different. Best guess is there will be a flank eruption within 3 years, or at least an intrusion, and that the ERZ will wake up and take the pressure off.

      • Hawaii is really an insane unstoppable Tap of liquid rock

        I guess its almost 1000 C above normal astenosphere temperatures 140 km down .. so you gets non stop decompression melting as its in a enviroment thats way above its melting point .. and more mantle materials are constantly comming in from below

        Hawaii is a real thermal channel from the core

        • There’s been conversation on the likelihood of a failure in the lava lake within 5 years and the risk profile dependent on direction of outflow. Where’s the USGS and Emergency Service on this for the communities near and far? Are they talking?

          • Currently they are not considering it at all. This might be fair as 2018 was a traumatic event and it will probably be at least a year or two before the lake drain is actually likely, a bit early for a warning.

            In saying that though HVO does sometimes seem a little forgiving on Kilauea, presenting it as only being dangerous if their hypothesized model for summit explosive activity is present, a model they actually got to test in 2020. Their efforts recently are much more focussed on Mauna Loa being dangerous, which is not unfounded but seems to be swapping one for the other when they should treat them both with equal attention. This also is a bit strange considering USGS as a whole considers Kilauea the most dangerous volcano in their catalogue and worthy of particular attention…

            It also seems they model all ERZ eruptions being the same. But they arent, 1840 was a long dike that started up near the Chain of Craters area, and moved very fast 25 km in less than a day carrying fresh summit magma all the way down with nothing in the way, it was a lot like most Mauna Loa eruptions. The 3 eruptions nearby in the past 100 years all began on shorter dikes that themselves began locally, so there was magma sitting in the rift that had to be pushed out first, in all cases. So 1840 was nothing like 2018 in its origin apart from being a high lava stand and a lot of pressure. It also seems like from reports it was not preceded by anything at all on the rift in the immediate time, only by vigorous summit activity. There was an intrusion 2 years earlier but nothing after. So was quiet then the floodgates of hell opened, a lot like at Nyiragongo…

        • Its more like a shield vent than a lava lake .. or a lava lake feeding a ”shield”

          Its a rising rootless lava pond

          If this was outside Halema’uma’u it woud be a shield vent like Kuapaniaha a bit

          A shield vent trapped in a caldera becomes a rising rootless lava lake

          • Not sure that is actually true though, because the filling from 1868 to the 1920s was actual shield building, the conduit keeping a small lake but the caldera was filled with solid lava.

            Its also not rootless anymore, the vent is well and truly submerged over 100 meters and closer to the middle than the edge now. One day when there is a flank event and the lake withdraws there will be left a great open conduit at the west vent location, much wider than the vent originally was, and probably violently degassing until it can overflow again.

            • A large, open conduit…like the photo in Jaggar’s book…

            • Probably more of a vigorously convecting lake within a narrow conduit, it isnt old enough to be like what existed in th 1920s but one day that will happen. It might be some several 10s of meters wide but on the lower end of that scale, vs the over 100 meters that was the case of the 2018 lake and the 1920s lake.

              If the lake drains out and has to refill it might even go into a lava geyser stage, like happened in 1959. The vent submerged itself before the conduit matured so that didnt happen originally but maybe a future drain could allow it. I would doubt all of the tephra around the caldera is exclusively from caldera formation events, some is probably from normal fountaining.

          • That huge mass there will never get solid as long as magma gets feed into the rising lava dam .. it will collapse by its own weight

            Ellis days where also like this but much larger scale

      • Looks like it. Lots of heat anyway. Has it had a lava lake before? I thought it was mainly strombolian.

      • This is what it looked like on May 2. Horseshoe-shaped lava flow and ocen entry.
        Picture borrowed at VD:

      • Its an emerging lava flow
        Or pancake dome

        As Hector say its too viscous for a lava lake

      • Not a lava lake at all

        Thats just the active vent where the lava is extruded

        Anak Krakatau is too viscous to have lava lakes in traditional sense

    2. Happened to read some about Angelina Jolie and saw that her uncle is a geologist, Barry Voight, last in Penn State, who is specialized on flank collapses and seems to have predicted the Mount St. Helens eruption which caught my interest. He’s worked on Soufrière Hills, Mount St. Helens, Merapi and volcanoes in Kamtchatka and South America.
      Not possible for me to get an access to these papers, but maybe interesting for others esp. Carl or Albert:

      For Mount St. Helens he seems to have been engaged by the USGS where he worked for a while.

    3. Little bit of noise, 10 miles sse of Pahala. There was also a 3.2 SE of Loihi.
      2022-09-13 14:21:10

    4. So, Albert I got to reading all of Nyaragongo by now (brillant, Jesper) and about Taupo (also brillant).

      Then I compared it to Lake Ilopango, much smaller (about a tenth of Taupo’s size, resulting from a VEI 6 that has probably contributed to the annihilation of the Maya (as known from the Maya series, also brillant, on VC).

      Similarities: A bit from the coast where the Pacific Ocean subducts with deep trenches adjacent, mountanous area.
      Difference: No geothermal fields.

      So I decided that apples have to be compared with apples and not pears and Taupo is playing in the league of Yellowstone, Long Valley Caldera or Campi Flegrei. So super-eruptions and geothermal fields seem to be coupled.
      Nothing new to any reader here.

      My question to whoever wants to dare an answer, including Hector, Jesper, Chad i.e:

      What was there first?: The super-eruption or the geothermal springs, so water that added to the monstrosity of those historical eruptions, the same normally wonderful water that we saw drive the second eruption of HTHH?
      Good q or plain stupid? I also dare plain stupid as it is questions you ask that get you somewhere.

      • And – forgotten – same league: Lake Amatitlán, Guatemala. Hot springs. The beast in the beauty.

      • Yellowstone has had hydrothermal eruptions of enormous scale, probably equivalent in power to a VEI 6 in some cases. This had nothing to do with new magma either, just decompression of the hydrothermal system. Mary Bay crater is the biggest at about 2 km wide, formed after the last ice age glaciers retreated, all of the other ones are younger and from when the lake receded or was otherwise changed, one was associated with a tsunami from an otherwise unrelated landslide apparently, when the waves went over the area it decompressed in the trough of the wave and blew up.

        In a way it is the far extreme end of a marr crater. Some maars are created entirely by magma, some are phreatomagmatic, and some are phreatic with only minor new magma. In this case there is no new magma at all or even any evidence of a new intrusion, yet we get a massive eruption that would probably be among the biggest in recent time in terms of sheer power, its just a big explosion enough to blow a 1 km hole in the ground…

        I dont know if hydrothermal systems like this are exclusive to large calderas though, it would seem likely that any volcano that is in a place with a lot of groundwater would be like this. Might also add a huge heat source. Yellowstone is just an extreme example, both of a volcano and heat source and the amount of water around. It might also be that the silicate precipitation at most of the vents tends to fill in small cracks so that pressure can build faster, so it might be specifically rhyolite calderas that have these characteristics. It is actually a little bit like how magma in the crust will fill in cracks and tends to accumulate in large volumes over time, only that decompression of water is much faster and more violent.

      • Hi Denaliwatch. I think there is room for debate on whether water played an important role on the explosivity of Hunga Tonga. It is true that Hunga Tonga happened in underwater and that could be seen as evidence of there being a connection. But it is also true that underwater eruptions are very frequent, Tonga has had several of those in the past several years, but none of them has been even remotely close to the explosivity of Hunga Tonga.

        I don’t know why Yellowstone, Taupo, or the Altiplano-Puna happen. But I think a key question to solve is how giant magma chambers come into existence, and why they eventually collapse. One problem is the briefness of human lifetimes. We have never been able to track the evolution of a caldera system or cluster of caldera systems, which is something that lasts ~millions of years.

        Another problem is that we cannot see magma intrusions, they are hidden from our view, and still have no good method to detect those that already exist. We can see new ones form. Existing intrusions/magma chambers not so much. The day we manage to picture magma plumbings with precision, it will be a revolution in volcanology.

        • I think the question should be read in relation to Carl’s explanation behind the explosiveness of HTHH, namely that prior to the eruption there was already an abundance of supercritical water in the ground and that the eruption set off a chain of events where the loss of pressure caused it all to flash into steam.

          • Indeed. Supercritical water is strange stuff. Until Albert mentioned it I assumed it was a bit like the CO2 system, but it really isn’t. There are actually two phases at reasonable pressures, but it takes energy to convert the ‘gas’ to the ‘liquid’, which is what a negative latent heat of evaporation means. So a large amount of very hot ‘liquid’ water under extreme pressure will hold ginormous energy which can be liberated just by reducing the pressure. When it transforms to steam it GENERATES energy and gets hotter, unlike at lower pressure where hot water ‘evaporating’ into steam cools the syste,

      • This isn’t a response to your actual question Denali but within the past couple weeks I finished reading a recent paper on Illopango that postulates a rebuilding magma chamber with a very high gas / volatile content that could be nearing a state of destabilization. It talked about the faults running through and around the caldera, how it functions as a trapdoor with the hinge on the north shore, and that it poses more or less a current risk to San Salvidor. It was very interesting and a good, although relatively complex, read. I’ll definitely look for it and try to post it.

      • I don’t really think that having geothermal springs is related to whether a volcano will be in the league of yellowstone, taupoa, etc. There are many volcanoes that are perfectly normal sized with geothermal springs.

        Geothermal springs are just a byproduct of water being in an area where there is a large magma body.

        But to answer the question of “what makes a large-caldera producing volcano or volcanic region?”, there are a few common traits that exist across most of the current large and active caldera systems (“supervolcanoes”) as well as extinct supervolcanoes.

        In short, you need to be able to create an enormous amount of magma without that magma immediately erupting first to alleviate building pressure. Typically that requires more viscous magma such as rhyolite. This is possible because you can accumulate tons of rhyolite below a threshold of being able to erupt due to being partially crystallized (which is owed to having too high pressure or not enough heat). This allows you to accumulate lots of the stuff without seeing the magma exerting pressure to push through to the surface. But the catch is that the rhyolite or viscous magma can be reactivated either through heating or a reduction in pressure (or commonly both), which can then cause a transition which makes the enormous volume of magma suddenly eruptible again, and likely to exert pressure upon degassing or releasing volatiles. It’s been studied quite a bit that you tend to see a process of magma mixing and reactivation prior to most of these huge caldera eruptions.

        So how do you get enough rhyolite in the first place? That tends to occur in two ways, one through slow fractionation of less viscous magma. This is what happens more commonly, and why we see volcanoes in Iceland that can erupt rhyolite. But this doesn’t on it’s own seem to create the enormous caldera zones we associate with large VEI-7 and VEI-8 eruptions. The other way to get lots of rhyolite is to melt tons of crustal rock. This is what is likely the main culprit in most “supervolcanoes”. Notice that despite having tons of volcanoes in the ocean, there are no known supervolcanoes in the ocean? That’s because oceanic crust does not melt to form the same type of magma that continental crust forms.

        The Taupo volcanic zone for instance melts prolific amounts of the New Zealand continental crust aka “country rock” via large and very hot injections of basalt coupled with spreading and extension of the crust. The combination of lots of heat and depressurization creates the environment needed to melt lots of bedrock, and subsequently reactivate that bedrock via enormous injections of fresh and very hot basalt.

        If you go through and find a map of most of these large calderas, you’ll notice that of the systems located in volcanic arcs (currently all except yellowstone), they’re all situated slightly further away from the primary volcanic front, in what is often seen as part of the back-arc system. Why is this? Well, it’s likely due to the fact that back arc areas experience the extension and spreading that can reduce pressure in the overlying crust. So this begs the question of why Yellowstone exists… Well, it turns out that you don’t *need* the extension and decompressions component, but if you’re lacking the that, you will instead need an incredibly prolific heat source instead, which Yellowstone has due to the hotspot beneath continental crust.

        So the summary here – Ingredients to build a supervolcano:

        1. Create prolific amounts of magma, preferably by melting the bedrock.
        a. In order to do this, you either need a truly prolific heat source, or significant decompression in the crust (or preferably both).
        b. Once you have the components, you just need enough time to accumulate an enormous magma chamber of semi-crystallized magma.
        2. Enough of this melted crust needs to stay crystallized for a while so that it does not force an eruption before getting large enough to result in something truly enormous.
        3. You then need fresh injections of heat or more rapid reductions in pressure to cause the crystallized rhyolite to reawaken and mix with the hot and fresh basalt.
        4. From there, you just need to break the surface in a significant enough manner, and you can start a runaway feedback loop where the eruption triggers further depressurization below (due to opening a conduit above), forcing more magma to activate and erupt until you empty out a very large amount of the eruptible magma and trigger caldera collapse.

        Keep in mind however, that most of the time, these volcanic systems erupt without enough eruptible rhyolite. Or in certain instances, the volatiles have all already escaped, which just results in slow dome building or large rhyolitic lava flows.

        • FWIW, I may be slightly off in some ways – this is largely my own observation after reading about many of these volcanoes past and present.

          You can comb over other systems around the world to verify. See some examples below.

          1. Taupo volcanic zone: in a significant zone of crustal extension.
          2. Kagoshima graben (home of many large japanese calderas): in a zone of crustal extension
          3. Toba: in a small zone of significant crustal extension
          4. Altiplano (andes mts) supervolcanic region: in a zone of very thick crust as well as crustal extension. Note: the thick crust here probably allows larger than normal magma chambers to grow before they erupt.
          5. Atitlan: In a zone where there is significant crustal thinning due to tectonic components.

          Then if you go back in history, you’ll notice that regions and time periods of intense supervolcanic activity such as the Mid-Tertiary Ignimbrite Flareup ( were a product of the crust seeing lots of extension during that time period (when the basin and range region formed in the USA).

          Similarly, the largest silicic large igneous provinces all form under similar conditions, often during a period where significant crust rifting occurs. Some of the largest eruptions ever found in the distant past are associated with the breakup of south america and africa, africa and arabia, or australia with a now-sunken region (whitsunday volcanic province).

          All that being said, I may be off on some things since I’m not a volcanologist by any means.

          • Thank you to all of you, Chad, Hector, Tomas Andersson, Farmeroz, Ryan and esp. the very long two answers of cbus, nearly an article.

            The theory with the crustal extension is extremely logical and interesting and leeds to the question, whether New Zealand i.e. or also the US could develop a rift, whether we see precursors of later continental break-ups and also maybe whether there is an inbuilt mechanism for that in case continents collide.

            This would mean
            1) that CAMP i.e. before it started working as LIP was maybe also a volcanic area like Yellowstone (pure theory)
            2) that places like Yellowstone or Taupo could mean a potential for a LIP when needed for continental breal-up (theory) in the future.

            Btw, I am of the opinion that talking about a new Pangaea in the future covers up the fact that we might be living on one as all oceans that separate America/Asia, Asia/Australia&Indonesia, Africa/Europe basically have to be called “Shallow Oceans” esp. between America and Siberia. It just doesn’t look like that ball but more like an animal with legs.

            Anyway, very kind of all of you. Fascinating stuff.

            • The only continent that is definitely separated from the rest (by a spreading ridge) is Antarctica.
              These balls (Rodinia and Pangaea) might be a bit too simplistic.

              I am writing with another device. My old dear MSI (game computer) robust and sixteen solid years old broke down. A new one would cost me at least 1,5. I will try to have it repaired. They will stare at me at MSI and think: That’s crazy. But I loved it, and I have hundreds of articles stored there. Next time: Double storing. My old MSI was like a rock, has seen Uncharted and Assassins Creed, all parts, has an ink stain from ink coming out that stopped and was just a lovable piece of history. History is lovable whether it be personal history or volcanoes and plates or political and scientific history.
              That’s why most of us also love the late Queen. She has covered since her Royal working started at 21 a solid 75 years of history.

        • Great response and informative overview!

          Was just reading how Yellowstone triggered 600,00 0 years ago after two very large injections of hot basaltic magma that primed the chamber for a super-blast.

          People constantly ask “what if Yellowstone erupts suddenly,” but I’d wager the chances of us missing two large basaltic injections taking place over decades into its magma chamber to be effectively nil.

        • Hi cbus20122. What you are saying regarding magma mushes is more or less in line with the popular view within the scientific community. A view I’m familiar with and in the past I was convinced of. I’ve also read about the idea of continental crust melting to feed flare-ups and silicic LIPs and at some point it may have also convinced me.

          I will have to oppose these ideas though, after seeing plenty of evidence against them. First I have checked the isotopic composition of Long Valley and Yellowstone, because for a moment after reading your I did consider the idea of continental crust melting to produce rhyolites being correct and had to check. I downloaded lead isotopic data from GEOROC and plotted it:

          Volcanoes have different isotopic compositions. Even in Hawaii, Mauna Loa, Kilauea, Hualalai and Loihi have different lead isotopic chemistry that distinguishes each volcano from the rest. If basalt came from the mantle, and rhyolite from the crust, then the isotopic composition should be very different. But as you can see in the graph above the basalts and rhyolites from each volcano plot within the same area. Yellowstone is more spread out in its isotopic chemistry, while Long Valley makes a very tight cluster, and this can also be seen in both the basalts and the rhyolites. This strongly argues for the rhyolite to have evolved from the basalt in both these systems.

          Further evidence for this is that volcanoes have different alkalinity, making crystallization series that are separate from each other. Say a basanite, a highly alkaline basaltic magma, will evolve into a phonolite, a highly alkaline felsic magma. I don’t think there is any volcano in the world which erupts basanites and rhyolites at the same time. This argues for felsic magmas originating from mafic magmas of a certain alkalinity.

          Regarding the idea of caldera systems containing mush reservoirs, I also see some evidence against this. Magma is known to be eruptible up to crystal contents of 40-50 %. Many stratovolcanoes, for example Stromboli or Merapi, usually erupt magmas with such high crystal contents. Caldera systems instead tend to erupt magma with relatively low crystal contents, under 20-25 %. For example the following article has tables that sum up of the composition of Taupo and Okataina rhyolites from all recent eruptions:

          Of all Taupo eruptions those that have the highest crystal contents have 10-20 % crystals. Okataina has never erupted more than ~25 % crystals, with only one exception, the Earthquake Flat eruption. Earthquake Flat, however, is arguably not an Okataina proper eruption since it did happen outside the caldera and usual area of eruptions.

          This caldera crystal poor pattern holds for most eruptions I’ve checked. The only exceptions as far as I’ve seen are some very small dacitic calderas, like Pinatubo or Quilotoa. So although I cannot say with certainty that the mush idea is wrong, the lack of erupted lavas with more than 25 % crystal in caldera systems worldwide does seem to argue against the theory.

          Added to this the reading I’ve done on ancient mafic magma chambers, layered intrusions, and cooling lava lakes appears to indicate that they gradually crystallized from the bottom up, with a melt rich region always being present, rather than being mushes.

          So I think it is pretty certain that rhyolite comes from basalts, in at least many large calderas. Maybe I’m wrong with the mushes not being mushes, but I don’t think so, evidence considered.

          • I forgot to mention, in Yellowstone and Long Valley all units are rhyolitic except those marked as being basaltic in the legend.

            • Thanks Hector, great to hear opposing views that come with some pretty solid evidence.

              I wonder, do you think the similar composition of the rhyolite and basalts in each of the volcanoes you mention a result of simple mixing and intermingling of the two? Also, I’m not suggesting there is no fractionation in these volcanoes, but the model I’m operating under assumes that a significant or even majority amount of the rhyolite involved in large silica provinces comes from crustal inclusion as opposed to fractionation.

              Most of this I got from reading about the TVZ.

              All that said, I can 100% see a different model that makes just as much sense based on what you’re suggesting, which fits why we see so many of these volcanoes in regions where there is spreading and decompression. The reason there being that the spreading and decompression allows for weakening of the roof of what were previously deep and very large magma chambers. Without this spreading, I could see that in many instances, the magma would never be able to erupt since it’s just too deep with too strong of an overlying roof.

              Regardless, when it comes to anything that is close to a large caldera system, with very rare exception (such as yellowstone), you will find elements of decompression or spreading on continental land mass.

            • I decided to do some new reading on this, and found this passage rather relevant: This is taken from

              “The zircon chemistries give insight into the degree of differentiation of the remelted igneous rocks, and the high-U zircon subpopulations indicate highly fractionated igneous rock representing a component of the source region undergoing remelting. Additional outcomes of these studies are that these antecrystic zircon-bearing rhyolites:

              (1) represent Zr-undersaturated magmas, where little to no new zircon crystallized prior to eruption;

              (2) may contain other inherited crystal populations (e.g., feldspar, apatite);

              (3) have most likely been generated and emplaced rapidly, based on zircon dissolution modeling (Bryan et al., 2008), which is a finding from studies of other rhyolitic magmatic systems (e.g., Charlier et al., 2005); and

              (4) show A-type geochemical signatures (Ramos Rosique, 2013).

              These age data thus indicate that while, at the first-order, silicic large igneous provinces, like the mafic large igneous provinces, record new crustal additions from the mantle through basaltic underplating and intrusion, and potentially substantial igneous crustal thickening (Fig. 5), with time, much of the silicic igneous activity instead reflects significant crustal remelting and recycling.”

            • Thanks cbus. Yellowstone may have a rift element, being located in the extensional basin and range, with many grabens around. The vents of Yellowstone are tentatively distributed along two rift zone like swarms, in a pattern reminiscent of Okataina. The one place I don’t see any evidence for extension is the Altiplano-Puna, and maybe Hokkaido.

              The basalts should not have underwent much mixing with rhyolite, or that would have sent them into the basaltic andesite field, or to an even more felsic composition. So the basalts must be representative of the primitive isotopic composition. The rhyolites could be a mingle between between basalt and a melted crust composition. However, if there was an important crustal component in them, that should produce rhyolites with a different isotopic chemistry to the basalts. In Long Valley the match between the basalts and the rhyolites is perfect, and corresponds to a very small area, so any crustal component in the rhyolites must be small.

              Regarding Yellowstone, the rhyolites marked in blue, strike a little more to the right in the graph. However these are the youngest, from the past 200,000 years, the Central Plateau Member, the Pitchstone Plateau and other voluminous rhyolite flows. These are younger than any of the Yellowstone basalts. So if isotopic compositions have shifted slightly, after the last caldera collapse, it would not be expected for this new composition to show in the basalt. The Lava Creek and Huckleberry caldera forming eruptions, match relatively well with the basalts. So I don’t think there is much crustal material in the rhyolite from these 2 eruptions. I should also mention that Yellowstone has a very weird isotopic composition, most lavas and rocks would have lead isotopic compositions near the right side of the graph. The way that both basalts and rhyolites display the same degree of weirdness also argues for a common chemistry and origin.

              This of course doesn’t mean other calderas of the world may not have significant amounts of rhyolite from a crustal origin, although I haven’t seen any evidence of that either.

            • Final comment here, I think the takeaway from several things I’ve read is that there is crustal assimilation required for a lot of silicic lip’s or large caldera systems. But this doesn’t have to disagree with what you’ve pointed out Hector.

              This crustal assimilation would make sense with the magma chemistry being so similar during instances where the remelted crust is actually recycled volcanic material. I’m pretty sure this is what is happening in the TVZ, where the greywhacke crust is old emplaced and highly hydrated volcanic rock.

              But who knows haha.

            • That passage is interesting. They provide evidence of zircons coming from re-melted material. There is something contradictory though:

              “Recognition of zircon inheritance and the magnitude of inheritance is difficult because of often subtle age differences amongst the dated populations and because individual zircon grain ages overlap with the general duration of Sierra Madre Occidental igneous activity (i.e., 38–18 Ma; Bryan et al., 2008)”

              “The zircon population ages are consistently older than the corresponding 40Ar/39Ar age, and this leads to the conclusion that the majority, if not all, of the zircons present in these silicic magmas are inherited and antecrystic (Bryan et al., 2008). The ages of the antecrystic zircons indicate that they have been derived from mostly solidified plutonic rocks formed during earlier phases of silicic magmatism.”

              The zircons ages are from the Sierra Madre period of activity. If there was crustal melting why aren’t there Cretaceous or Jurassic aged zircons? At most half the volume of rhyolite could be re-melted material. If a number of plutons were emplaced during the flare-up, then melted again and erupted, the first half must have come from the mantle and the second half from the crust.

              That does not solve the problem of the lack of zircons older than the Sierra Madre Occidental activity though. Certainly an older pluton could have been re-melted too if the SLIP ones were. The Sierra Madre Occidental had seen volcanic arc activity for a long time before the recent SLIP, the so called Nazas arc which is of Jurassic age was active in the same location as the Cenozoic SLIP. And I imagine zircons are found in non-plutonic rocks too.

            • I remember reading a FB post by USGS about rhyolite at Yellowstone, and how it us actually very hard to get basalt and rhyolite to mix. It is why basalt doesnt erupt often within the active caldera but will do so through the older calderas in the Snake River Plain. Rhyolite is a bit like oil on water, except the difference in viscosity is probably a lot bigger. I expect some mixing does happen but it is not easy.

          • Considering that the reaction to my original question was overwhelming I suggest that you both, you and Cbus (and also Albert) write a series about super volcanoes not focussing on the risk (we have a lot of this stored in VC) but about mechanisms and chemistry with a lot of debate. That could be quite interesting. I would love it.

            • That is a great idea. I was actually going to write some articles about resurgent domes and piston uplift of calderas but ran into a conundrum and abandoned it. I will try my best to write a series on calderas at some point, even tough I think I’m still missing on something important.

        • Posting image from the following link:

          Like most volcanic arcs the Sierra Madre Occidental has seen volcanism for hundreds of million of years. If re-melted continental crust plays a role in flare-ups, old Mesozoic-aged zircons should show up in young calderas, in places like the Andes, or western North America. I have not read many articles that study the age of zircons so I don’t know if this evidence has been found or not.

          • Anyway, interesting question, lots of mysteries. My previous headache on this topic was how magma chambers are constructed. I considered doming/piston uplift, coalescence of individual sills, and thermal erosion. Couldn’t come to a clear conclusion.

            • Now that I think of it you need about twice the volume of basalt to generate a volume of rhyolite through fractional crystallization. And that also means that assuming fractional crystallization is the way that the majority of rhyolite forms, then ~half the volume of a pluton should be crystallized basalt cumulates (gabbro). Leaving aside layered intrusions, I don’t think plutons have that much gabbro, most are mainly granitic.

              So where would the basalt go? Perhaps the fractional crystallization takes place deeper than the pluton, or maybe basalt simply convects and mixes slowly with the rhyolite while most goes back down.

            • Basalt is heavy. When is cools, it will tend to sink back down

      • Naples is a good example. Vesuvius, run of the mill volcano, and phlegrean fields, remnants of a much bigger eruption a long time ago. The latter has hot springs, the former doesn’t. To get hot springs you need a water circulation to several km depth, so a broken-up rock base and a large extent. Maybe that is what comes from those suer caldera collapses. Hot springs and geysers are an indication of stability. When you get them, it indicates there is little or no risk of any eruption.

        • Didn’t see this. Thank you very much. So, if there are hot springs being a sign of stability I conclude they developped after the super-eruptions. Maybe.
          What about Tarawera and the Pink Terraces? Is this a mean question?

    5. I consider the inflow of reactions like setting of an avalanche. Lots of interest in the mechanisms of the super-volcanoes. Do about it, I can’t. I can add to it or ask questions.Your Taupo piece got us there, Albert, good thing.

    6. Geochemistry and Petrology Group @ Uni. Iceland
      We are very proud to present our new paper in @Nature
      on the #Fagradalsfjall eruption!!! ✨📣✨📣✨
      Have you been wondering what’s going on deep beneath that beautiful volcano? Read the article for free:

      And the research brief:

        • It is the Czech version for Charles and, I believe, also the Dutch version. In Polish it is Karol.

      • Thanks for a very fascinating article.

        Fagradalsfjall seemingly did the opposite of the Timanfaya eruption of Lanzarote. The 1730 Lanzarote eruption rapidly decreased in potassium contents and incompatible elements during the first several weeks of eruptions. Fagradalsfjall in contrast instead increased the potassium and incompatible element contents in its first weeks and reached a level of alkalinity, higher than any other Reykjanes eruption sampled. It is an interesting question if the higher alkalinity helped drive the intense fountaining, given that the fountains started when potassium contents had risen to the highest levels. What a unique, weird eruption Fagradalsfjall was.

    7. There is some interesting deformation signals at Kilauea, the lamehas stopped rising and the cross caldera extension that has been ongoing since about a year ago has abruptly stopped. There might also be a signal of contraction on the ERZ too but hard to tell.

      Also of note is that there is a small cluster of quakes near the Kaoiki faults, a place that usually quakes in response to deeper magma movement at Kilauea. Its not a swarm to suggest abything major but this is a notable change in an otherwise long term trend.

      It looks like maybe the connection between the deeper magma system and the shallow part that feeds Halemaumau could have been partly obstructed. The GPS stations individually all are showing upwards movement to varying degrees, some being quite noticeable, but must also be moving closer together relative to each other which would suggest a loss of pressure inbetween, which is where Halemaumau is. I dont know if the eruption will actually stop like it did last year but things might be pretty quiet for a few months. Of course then there will be a few months of lava erupted in a day which should be quite a show 🙂
      It is also far from impossible that another part of Kilauea will erupt instead, Keanakako’i and Kilauea Iki come to mind as alternative summit sites.

        • The first of the Nature links, the one titled “Deformation and seismicity decline before the 2021 Fagradalsfjall eruption”, is a real gold mine for anyone interested in how to read and interpret the various signals.

          There’s a really detailed description of how earthquake focal mechanisms, magnitudes and relocations are determined and how they differ from the routine analysis by the SIL network. Then there’s a very nice description of how GPS data, satellite interferograms and earthquake locations and focal mechanisms are combined to create a geodetic model that is iteratively refined in order to map the dyke orientation, depth and opening, along with other deformation sources such as the main Reykjanes fault and the larger triggered quakes.

          Worth to note is that in order to avoid triggered quakes and aftershocks along their fault planes, they used relocations of quakes in the magnitude range 0.5<M<1.4 for mapping the location of the dyke. Without the refined relocation methods and velocity models used here, the hypocenters of these quakes are not very well constrained. It is mentioned, for instance, that the routine analysis tend to place the hypocenters roughly 1km too deep. Something to keep in mind when we amateurs try to do the same thing by plotting data straight from the SIL catalogue.

    8. Deep tremor shakes the island of Hawaii. The tremor lasts 35 minutes, and has some sudden spasms which HVO has located:

      2022-09-15 09:07:40

      2022-09-15 09:06:24

      2022-09-15 09:04:22

      2022-09-15 09:00:33

      The rapid sequence of three large spasms in 3-4 minutes with M 3.3, M 2.4 and M 2.8, is remarkable, and can be seen clearly in seismograms. It reminds of the big tremors that were frequent in early 2019 and early 2021.

      The tremor comes from the same location as seemingly almost all deep tremors in Hawaii, a ~40 km deep location offshore Pahala. There were three other weaker tremors in the past 48 hours. One on the 13, and two on the 14. So we are looking at a tremor flare. Is it over or will we see more tremors? There were two periods of enhanced tremor activity in early 2019 and in early 2021, both of which were followed by increased fracturing earthquake activity under Pahala.

      • I wonder if these tremor events are recording the creation of new sills, or other intrusions, and that those inflate for a while until then another one will form nearby. The first was the original swarm under Pahala itself, then a second possibly larger one that was to the east in the direction of Kilauea, and then now a most recent focus that is southwest of both of the other swarms, sort of in the direction of Kame’ehuakanaloa. Why there would be intrusions here at all is a question but evidently there is something going on.

        Question is still whether this is actually an independant proto-volcano or if the interpretation of it being a feeder to another volcano is correct. Both Kilauea and Kame’ehuakanaloa show some evidence of being at least plausibly connected, there are frequent quakes the same depth as the swarm that occur under the summits of both volcanoes as well as inbetween their summits and Pahala swarm, although nowhere near as densely as at the swarm itself. This is actually presently quite easy to see on the 1 week monitoring page for Kilauea right now, which shows the quakes of that timeframe a bit more accurately than the live map. Mauna Loa doesnt look like it has anything to do with the Pahala swarm at all though, only influencing the stress field, as the swarm looks like it is tilted up away from the direction of Mauna Loa in the depth chart.

      • I decided to take another look at Pahala, this time I noticed some new things. The following graph shows magnitude of earthquakes versus time. Red are tremor spasms, sudden spikes in tremor events which are, sometimes, located by HVO. Tremor spasms/tremor quakes are always in the depth range of 43-38 km deep and in the same area offshore to the south of Pahala, so I can isolate them by spatial location. Green are fracturing earthquakes under Pahala at depths of 35-37 km. This depth seems to be the depth were waves of Pahala earthquake activity usually start, before moving to shallower levels.

        Pulses of tremors seem followed by an increase of 43-38 km deep earthquakes. The first time this happened was is 2015. Tremor activity had been building up slowly since 2013 out of a period of relative absence of tremors of at least absence of tremor flares that goes back to before 2000 (I don’t really know when was the last time there was this much tremor activity in Pahala, certainly the 2000-2013 period was of low tremor activity). Tremors became frequent and energetic from January to June 2015. This was followed by the first deep fracturing/volcano-tectonic swarm. In October 2015 a swarm happened under Pahala, at ~35 km depth lasted only a few days but was among the most intense Pahala has seen. View of this swarm:

        Unrest under Pahala continued for the following years both seen in tremors and fracturing earthquakes. Then came the next big flare in tremor activity from roughly February to August 2019, with itself came in multiple shorter pulses I talked about in one of my articles some time ago. As the tremor activity died down a swarm of volcano-tectonic earthquakes started under Pahala in August, at depths of ~35 km. This swarm seem to roughly propagate northward as a wave, before months later moving to shallower depths of ~30 km. View of the first few months of the swarm that started in August, and the arrow shows earthquake propagation:

        Tremor activity picked up again from July 2020 to May 2021. During the later part of this tremor flare, deep fracturing/volcano tectonic activity emerged under Pahala again. It is not clear where and when this swarming started exactly but moved gradually towards shallower depths some months later.

        So while I don’t know what is going on, there is a trend for these pulses to start offshore Pahala at 40 km depth in the form of tremors, propagate 10 km up, and 20 km north or northeast. Because there is vertical propagation I think they are more likely to be dikes than sills, that is if they are intrusions. I also think most of the earthquakes happen in pre-existing faults that get triggered. Many of the earthquakes are happening in the southern section of the Mantle Fault Zone, a sub-horizontal reverse fault that runs towards Kilauea at 30 km depth. Other deeper faults are also getting triggered, like was probably the case for the M 6.2 of October 2021 offshore Naalehu.

        • Your line of progression runs parallel to the southern rift of Mauna Loa. (it also follows the edge of the national park but that is probably accidental.) The depth suggest it follows the stress field of the island. The weight of the island and especially Mauna Loa push outward and own, and the easiest way for magma to travel is perpendicular to this force. That would give approximately your direction. Further in-land you would expect it to turn more to the east, if this is correct. It makes it a deep flank fault

        • Little more associated with this quake. Picture is not as impressive, not posting.
          2022-09-15 16:11:32

      • I hope HVO releases their new data on Pahala soon, they did a proper survey mid this year, with a dense array of seismometers deployed that they retrieved just recently. It is probably a lot of factors combining in order to create this swarm, but given the depth and presence of tremors at least the most likely and plausible option is that this is an intrusion of some sort. A rather big one it would seem… There could be other options but I cant think of what those would be in this situation, its not water that deep down.

        I have been for a long time proposing this to be a future major volcanic event, most likely at Kilauea although Kama’ehuakanaloa is also a possibility but in that situation it will be visually boring 🙂 Magma is on the move, where it will surface we can only wait.

      • This would be the perfect location for a good VEI 6 eruption; well monitored and in an remote region! But the swarm is weak and declining so don’t get your hopes up. 🙁

        • I would be surprised if it came so soon after the Novarupta 1906 eruption (also a vei 6)

          • Note – I say this from the viewpoint that Novarupta likely stemmed from a deeper joint magma chamber that connects Katmai with Trident (and other nearby peaks).

            • Interestingly I think if not the article I linked has speculation that the area could be something of a very large interconnected volcanic field could have large collapse events in the future (larger than Katmai). I think the mechanics of the Novarupta eruption as a distal flank vent supported that, along with the fact that Trident has four tightly clustered cones. Perhaps that general area goes big, big big, in the future?

            • For some reason my comment came out as barely comprehensible gobblidigook. Let me try this again:

              Interestingly, I think either the linked article or the comments had speculation that the area could be something of a very large interconnected volcanic field that could produce large collapse events in the future (larger than Katmai). I think the mechanics of the Novarupta eruption as a distal flank vent supported that, along with the fact that Trident has (now) four tightly clustered cones along with other nearby peaks. Perhaps that general area goes big, like Aniakchak or Okmok big, in the future?

            • Might not be that simple though, because Trident erupted andesite in the 1950s and so would seem is not based around a rhyolite chamber. That doesnt mean the area cant go caldera in a massive way but it could be a while before that does happen.

              I guess it all revolves around where most of the 1912 eruption came from. Maybe Novarupta was a more minor component and Katmai is where the main eruption did happen, it is where the caldera formed after all and viscous magma like rhyolite is probably not so easy to transport laterally in an intrusion at high rate.

              Maybe Novarupta is an entirely new volcano that was formed from some old rhyolite remobilised by the same intrusion that caused Katmai to erupt, but was not a lateral vent of such. Would explain why Trident is not rhyolitic despite being between them.

        • Yup, nothing that impressive at the moment. But it was a good excuse to brush up on Novarupta / Katmai and read a bit more about the area.

    9. More noise. There are several quakes in the Pahala area for this event.
      2022-09-15 09:07:40
      2022-09-15 09:06:24
      2022-09-15 09:04:22
      2022-09-15 09:00:33
      2022-09-15 08:40:20

    10. Hello, I’ve been a silent reader at VC for years.
      But now I would like to say thank you for your great contributions and also the comments and help with questions!

      There is also news about the Eifel volcano, see:

      How deeply does the Eifel volcanism sleep?
      The Large-N measurement campaign

      Here is the URL of the GFZ contribution:


      You are very welcome! First time comments are always kept back by the system for approval by an admin (sadly this is necessary). Future comments should appear without delay – admin

      • Thanks for the link. The VulkanEifel is an area that still needs to be closely monitored, because a VEI5/VEI6 eruption there will wreak havoc in Germany and also in Belgium and in The Netherlands. I live in the northeastern part of The Netherlands and I can literally say that I have an active volcano in my neighbour’s backyard, that volcano is Laacher See.

        • There was a series of articles on the volcanism of Eiffel a few months ago, os quite a good read. Chances of a VEI 4+ are low though, Laacher See was pretty rare event, most are mafic pyroclastic cones. Eiffel also seems to like to sleep, millennia between eruptions. Its a bit like the two active volcanoes in Australia, historucal time is not long enough to see an eruption, but they are very much there just working to a different timescale than we do.

          Probably in regards to a volcanic hazard posed to northern Europe, would be mostly from Iceland, and only during really major eruptions in (un)favorable weather. So maybe about once a century, with variability.

          • Certainly the chance of an eruption in the Eiffel is low. But it is not zero, and its location in the industrial centre of Europe means that a medium eruption could have significant impacts. A small eruption would be ok.

      • A very interesting development indeed, and a welcome one in my opinion. It’s not about the statistics of how often a volcano forms if the Eifel, nor is the nature of a new vent. It’s a fairly densely populated area located in the heart of Europe, so the least thing that could and should be done is to monitor it properly. I’d rather see 10.000 boring years with a proper view being kept of what is going on in the Earth’s crust, than an equal amount of time of coming up with excuses and not having more than a vague clue on magmatic behavior.

        I’ll be on the lookout for research results in the coming years. Thanks for bringing it up, Gerhard!

        • I was happy to share it and I’m also eagerly awaiting the first results
          That’s how I feel too, it was a tragedy until now what happened in relation to the surveillance in Germany and also happens in other areas. See Hegau: Hohentwiel, Staufen…. all “extinct” volcanoes and that in the vicinity of the Hohenzollerngraben… as well as in the direction of the Czech Republic in the Vogtland, here is a link to a report on swarm earthquakes from 2018

          another area…
          But that’s at least a start. 😉

    11. Aaaa having Lucid dreams all the time of walking IO s surface… as Albert can figure out: IO is my favorite place now among geological surfaces in the solar system, the the 70 kilometers wide and 3 km high lava falls of Pillan Patera pit, the 300 km wide lava lake of Loki Patera, the giant plinian eruptions of Twasthar, the ultra plinian eruptions of Surt and lava flows the size of UK

      freaking cannot get better… hopes Juno Spacecraft gets close up views of this

      • Will the gas drag from the Red Giant Sun in the future cause Jupiters moons to spiral into Jupiter itself?
        Jupiter will be well inside the planetary nebula… while thin in density… it will be ALOT thicker than todays cosmic enviroment around Jupiter

        woud suck if Jupiter system as well gets trashed by the dying sun…

      • InSAR, a satellite-based method to make maps of ground deformation, could be a way of determining when an sheet intrusion-fed eruption will happen if data was obtained with greater frequency. But the problem is that for a satellite to take data of the same location it takes something like 12 days. So in the end it is a lack of the technology necessary to view intrusions.

        Earthquakes do not really show the shape of the intrusion well enough, earthquakes mostly happen on faults that get triggered by the intrusion, so it takes experience in the volcano to know what they show. Piton de la Fournaise, for example, when it makes intrusions earthquakes are mostly limited to the ring fault of its summit caldera. There are some faults in the east flank that have been triggered by intrusions towards the east in the past. However the intrusions themselves cut aseismically through the rock, and thus do not make much seismicity outside those two fault systems.

        The September intrusion was very small as shown by InSAR. I checked the interferograms in the Mounts Project page for Piton. You can see the concentric fringe pattern inside the caldera of Piton. At most it could have culminated in a small summit eruption, but many small summit intrusions do not erupt. In contrast the larger intrusions that reach down the flanks almost always erupt. But you can only know these things once the InSAR data comes out. And interpretation is always difficult anyway, we are just starting to learn how intrusions behave.

        • Wrong one, this is the interferogram that really covers the “crisis”:

          There is no difference with the one I posted before though, which was from before the seismic crisis. There is again an inflation signal at the summit, but maybe that is just the summit pressurizing, no clear evidence of an intrusion. That said the intrusion-eruption event could happen soon if Piton de la Fournaise keeps inflating at this rate.

          • Been a relatively quiet year so far, since 2015 there has been more than one eruption every year and there were 5 in 2020. Kilauea is the most productive shield volcano (well, volcano in general) but Piton de la Fournaise is probably the one with the most frequent eruptions. It might actually be the record holder world wide, most eruptions there are a whole new complete intrusion every time, a lot of stratovolcanoes erupt often too but usually have open conduits so it doesnt really count.

            Only thing I can think of that might be comparable are rifting events like the Krafla Fires or what is happening now at Fagradalsfjall but those are temporary, maybe a few decades at most and driven by tectonics not magma overpressure. Piton de la Fournaise has been as active as it is now for at least all of the last century and in much the same style, only 2007 being a big deviation. The counter is most eruptions there are rather small, around an order of magnitude less than in Hawaii or Iceland, except again for 2007.

            Probably will see an eruption some point in the next month.

    Leave a Reply

    Your email address will not be published. Required fields are marked *