Introduction (Albert)

Iwo Jima is famous. The battle between the US and Japanese forces are well remembered – by both sides. Before the war, about 1000 people lived here. After the war, none. There is a military base only. The beaches are filled with rusting hulks of ships, sunk in the long battle.

And that is funny, because those ships weren’t sunk on the beach. They were in water. And that is the second reason why Iwo Jima is famous. The land here is rising, and has been doing so for at least 700 years. Since the time if Captain Cook (who visited posthumously – he had already did when the expedition came past here), the island has risen by 40 meters. It has grown in size as well: reefs that were mapped by Cook’s expedition are now part of the island, including one dubbed ‘Stonehenge’. VC designated it as the most dangerous one in the new decadal volcano list. It was a choice which surprised people, as the volcanic nature of the island was not well known. It made the news. This was before Hunga Tonga showed that a VEI-6 sea level eruption could cause a tsunami across the Pacific. VC got there first.

The growth of the island is readily seen from images. After the war, the island was a waste land, but images show where the coast were. Here is an aerial view from 1945

Here is a satellite view from 2018. The island has greened, but has also acquired extended sandy beaches. The sides are staircase like, showing the effect of wave erosion during the uplift. The small island to the north, part of the caldera rim and outside the frame of the 1945 view, is now almost attached to the main island.

There has been strong uplift since 2020, although the island hasn’t grown much further in that time, judging from images. The phreatic explosions that were occurring have becoming magmatic. A new island formed off-shore recently. It has eroded away already.

So what is going on? Is this normal caldera activity, ups and downs as seen in Campi Flegri (to which Iwo Jima is sometimes compared)? Or is magma accumulating? Will the current burst of activity temper down again, or will it continue to grow? Could another caldera eruption occur, and what would the warning signs be?

Tallis and Hector have combined to tell us their views. Enjoy! Before reading their words below, you may want to read the earlier posts on Iwo Jima: Henrik’s now famous eruption story, top of the VC new decadal list is a must-read. Later we had a post on the pre-VC history of the island.

The New Decade Volcano Program No. 1 – Ioto, Japan

Iwo Jima in 45 eruptions

Do remember that the island is now called Ioto. Iwo Jima is history. But what about the future? Keep reading.

Tallis Rockwell’s view (Crisis on Ioto):

Considered to be the most dangerous volcano in the world on this blog, Ioto is quite a temperamental volcano, rising at an average rate of 15-25 cm/yr for more than 500 years. I’ve got no interest in delivering a refresher at this moment, if you want a general understanding of known information on this system please read Albert’s articles on this system. Simply put this volcano lies on the Izu-Ogasawara arc. It has chamber of unknown depth, likely deeper than 2 km but shallower than 10 km. A shallow 4-5 km wide sill at 0.8-2 km depth and through that sill there is a cone-sheet intrusion as the magma flows around the Motoyama block. The volume of these structures are currently unknown. It should be known that despite the chamber supplying the sill and cone-sheet intrusion, it is still a significant supplier of deformation at the
volcano.

During the past decade and a half, Iwo-jima has seen over 10 meters of uplift, numerous phreatic eruptions and since 2022, it’s first magmatic eruptions in over 500 years. Up until recently uplift had seemingly slowed and the frequency of eruptions slowed. It wasn’t until the ending of August that things escalated dramatically. The most dramatic deformation pulse at this volcano so far is taking place, with critical fractures having taken place along the Asodai Fault and purported ring-fault. These cracks are hot, constantly emitting steam. All eyes should be locked on Iwo-Jima right now as it is likely we’ve reached a tipping point. In order to understand this proposition we need to understand some basics about how this unrest has progressed.

The average rate of uplift has been undergoing a centuries-long acceleration phase. The coast had risen 40 m since 1779 but over 17 of those meters took place in less than 70 years after 1945. Clear evidence of acceleration. This acceleration trend had recent expanded an entire of order of magnitude, rising well over 12 meters in 12 years. The past period of unrest culminated in the a series of magmatic eruptions off the coast of Okinahama starting in August 2022
continuing on into February 2025. The magma from these eruptions came from shallow depths likely not exceeding 1.2 km. The first eruption in 2022 consisted of degassed magma but the 2023 eruption’s material was more volatile, probably due to some maturing of the vent. These eruptions weren’t particularly intense or voluminous. With all of total volume of all the eruptions from 2022 to early 2025 not exceeding 1 million m3 but it did indicate one thing, the Cone-sheet
intrusion had finally expanded enough for a magmatic eruption after more than 500 years.

The current deformation pulse is likely related to the sill and cone-sheet intrusion and is exceedingly high, around 3-4 cm/day. Motoyama is experiencing a deflation, likely sourced from the sill surrounded by a large ring of uplift. Like the last deformation cycle, The cone-sheet intrusion seems to be the dominant producer of the uplift. And while we don’t know it’s volume as of late, we can guess with the given details. Just looking at these insar maps, we find that the uplift from the cone-sheet intrusion covers most of the island, around 2/3s of it with some more uplift taking place underwater, as such it would be reasonable to assume that the cone sheet takes up an area of over 25 km2. It’s height is probably around 0.8-1.6 km depending on the angle of magma ascent. This leaves us with a volume of 20-40 km3 of magma. Not accounting for magma in the sill.

The last uplift cycle(2012-2023) caused the cone-sheet intrusion to pressurize by 10 MPA, not good considering the magma is stored at a depth 25+22 MPA. It’s safe to assume that the Cone-sheet intrusion is now overpressurized. Iota has likely been cycling through different deformation cycles. This is likely reflected in the clear long-term acceleration
phase. The broader early cycles were likely the result of the chamber becoming receiving the bulk amount of magma, leading to it’s overpressurization, it produced a sill as result, leading to a spike in uplift speed and decrease in scope, the sill then broke and produced a cone-sheet intrusion that would be the cause of the current observed uplift.

We’ve got an overpressurized chamber producing an overpressurized sill that has now produced an overpressurized cone-sheet intrusion. The craziest fact still is that there is no ; large-scale subsidence of the magma chamber has been noted in decades of topographical surveys, on the contrary. We’ve found that the chamber is still uplifting! During the last deformation cycle, the NE rim of the caldera rose! That fact means that all the magma that has left the chamber was replenished and then some!

If the current deformation pulse is related to the magma chamber then I believe that we’re seeing past “failed” uplift.These insane speeds make magma supply an unlikely cause if it related to the chamber. Despite delivering some of it’s supply to the sill and cone-sheet intrusion, the chamber has still been building pressure and strain at more rapid pace than most volcanoes on the planet. It is almost definitely overpressurized. Hector made a quite timely article concerning vertical-t CLVD quakes and as such I’d point you in that direction. Simply put, sometimes a ring fault has to potential get “stuck” and accumulates strain and energy soon the ring fault slips and rapidly rises producing strong earthquakes. This pulse could be fueled by a similar mechanism, just slower, instead of being released in a rapid slip event, the strain could be released in a slow failure leaving us with prolonged extreme uplift periods. This way, magma supply, although still definitely extreme, won’t reach truly impossible levels. The evidence for this idea could be supported by some of
the GPS data. You will notice that during the past few years despite uplift slowing, its associated horizontal deformations didn’t pause. This could imply that the magma supply rate remained unchanged.

The issue with this idea is that not everyone will agree that Ioto has a ring-fault, in his own article, Hector disagreed with the existence of a ring-fault but if you ask others they’ll say yes. The only other options beyond preposterous magma supply or slow-slip ring-fault uplifts would be the plug starting to fail or rupture of the magma chamber. If it is from a failing plug, we’ll see continuous intensifying eruptions rather soon. If it is a sign of an incoming rupture from the magma chamber, we will probably about to see a lot more magma enter the shallow system. As you’d guess, none of these propositions are good for this volcano’s future.

Even forgoing the magma chamber the cone-sheet intrusion has enough magma to produce a high-end VEI 6 or a low-end VEI 7. Remember all the chamber has received at least as much as it lost so the magma chamber is much more capable than whatever the shallow system can do. A low-end VEI 7 being the worst-case scenario for this volcano
could be a solid underestimation.

Current eruption and Caldera Trigger

All of the possible causes for this intense uplift suggest that either this chamber is primed and overpressurized or it will be very shortly. But we need more than an overpressurized system for a caldera-forming eruption, we need an adequate conduit for magma to flow through. That is the only ingredient missing and that may also change soon. The current eruption is much more intense than the past eruption producing pyroclastic surges and lava fountaining, still it’s not
very strong in comparison to other volcanoes. It is a Phreatomagmatic eruption. What we can expect from this eruption depends on where the magma is sourced. If it is from the cone-sheet intrusion then the chance for an imminent escalation is lower but if it is from the magma chamber, then the chances would go up. The conduit for this eruption likely developed from the extreme stress the uplift is putting on the crust regardless of cause. Assuming that the eruption and all future eruptions for the foreseeable future are fed by the sill and associated cone-sheet intrusion then we would ultimately need to wait for violent enough eruptions to compromise the plug. A strong VEI 4-5 or series of intense VEI 3-4s could do this. The shallow system likely has enough magma for this, and while the quality of the magma may not be the best, there is no shortage of water to make for violent reactions so this will always be a solid possibility. This is actually preferential as it would give ample time for study and warning. A process like this could take months or even years. Even if this scenario is true, it is still possible that another conduit directly connected to the chamber will develop, risking a faster-build up. If the magma is sourced from the magma chamber, then we already have direct conduit to the primed system and this could quickly escalate as it expands and matures. It might take months or even weeks leaving less time for warning and study. We don’t know where the magma is sourced so we can’t jump to any conclusions.

Another risk is a major earthquake at the Asodai fault, as I’ve stated before numerous cracks have formed along this feature so it’s also under stress. The failure of this fault could also lead to a significant escalation. Caldera-forming eruptions follow a very predictable trajectory, Phreatic phase<Phreatomagmatic phase(We’re here)< Non-Plinian Magmatic phase (occasionally with effusive properties, but this phase does not always occur)<Plinian phase<Caldera-collapse.

Ioto, Iwo-Jima, Sulphur Island, whatever you want to call it, is undergoing massive changes over a vast area. This volcano has the potential to crush every other known historical eruption in terms of intensity. Not only could it produce an eruption as intense as Krakatoa or HTHH through interactions with sea-water while being an entire order of magnitude larger, it could even be more intense as there could be significant amounts of magma mixing between a slew of mafic, intermediate, and felsic magmas! It was always king, for centuries, this volcano has been building itself up and despite so many finer details being lost to us, it’s still terrifying. A caldera-forming eruption should no longer be treated as a distant hypothetical but as a likely future, possibly sooner than anyone might expect.

This is my first time talking about Ioto/Iwo-Jima, and I must say this volcano is fascinating, with many unique aspects to its activity that lie beyond what I could have initially imagined. This article happens as Ioto produces its most intense historical eruption recorded. A new vent opened on Sept 1, and has resumed erupting since Sept 14 until today. The ongoing eruption and large-scale deformation associated with it deserve immediate attention.

Since I haven’t tackled the volcano yet, I will start with a brief geological overview and a commentary on the structure of the volcano. The submarine flanks don’t seem to have the large fields of concentric waves (created in repeated caldera-forming eruptions) that other volcanoes like Kita Ioto or Kaitoku to the north have, so Ioto may be relatively new to caldera-style volcanism, or it has an older pyroclastic edifice that is buried under an effusive one. The north flank of the volcano is mostly covered in lava from the effusive stage, which is probably correlative to trachyandesite flows exposed near the west rim of the caldera (Kamaiwa coast area). Only in the north flank valleys, there seems to be clastic material and locally turbidity ripples, probably from PDCs that were channeled down the ravines. The south flank is instead entirely covered in clastic material and turbidity ripples that fill the valleys; likely from the caldera-forming eruption that must have been focused on the southern side of the caldera, between Motoyama and Suribachiyama, possibly.

The caldera of Ioto is circular or chestnut-shaped (if Suribachiyama forms one end of it), of about 10 by 11 km in size. A thick pyroclastic layer forms the island of Ioto, which, although it hasn’t been interpreted as such, I think represents the caldera-forming eruption that ponded inside of it, now brought to the surface by resurgence. This deposit (extensively described in Nagai & Kobayashi, 2015) comprises three layers: the Hinodehama Ignimbrite, the Motoyama lava/Kongoiwa pyroclastic deposit, and the Motoyama pyroclastic deposit. The Hinodehama Ign. is a pyroclastic layer ~7-8 meters thick that is extremely welded and lava-like (has columnar jointing even), and contains abundant carbonized wood within its lower section. The Motoyama lava/Kongoiwa pyroclastic deposit is a layer about 15 meters thick that has been variably interpreted as an intrusion or lava flow, but I think is part of the ignimbrite sequence; it’s made of sheets of lava (or lava-like ignimbrite?) in places and in others breccias with large lava blocks as big as 20 m, bits of this layer are also mixed into the lower part of the Motoyama pyroclastics above which I think supports an explosive origin. Then comes the Motoyama pyroclastics, which is a pumice tuff (unwelded PDC deposit) 60 meters thick, making the bulk of the eruption. The composition of the three layers is very similar and mostly overlapping, and consists of trachyte with 62 wt% SiO2 and 1 wt% MgO. The presence of MgO indicates the lava is not fully evolved (sort of the alkaline equivalent of a dacite caldera).

The 80-meter-thick pyroclastic sequence has been interpreted as a subaqueous caldera-fill postdating the caldera. However, the lower section, the Hinodehama Ignimbrite, has to be subaerial; the abundant carbonized vegetation likely came from a forest on an island that was wiped out by the currents, and a lava-like ignimbrite shouldn’t be capable of forming underwater (and I doubt it can occur in eruptions that aren’t caldera-forming). Given that the pyroclastic sequence must have started above water, but we know the current Ioto island only became subaerial around 800-500 years ago (age of the highest marine terrace), then the deposit must have formed during the collapse, taking out the island that existed before the caldera (which we know existed from parts of the rim that are at sea level) and transitioning to an underwater deposit as water flooded the collapsing crater. So the Hinodehama ignimbrite must have formed subaerially, wiping out the vegetation of the island and settling into a strongly welded layer, then the Motoyama lava/Kongoiwa pyroclastic deposit must be a hybrid, in places emplaced as a subaerial deposit, in others, already underwater, spatter exploding in contact with the water formed breccias, and lastly, the bulk of the eruption happened subaqueously when the vents were drowned and the eruption turned phreatomagmatic, producing a pumice tuff. It’s also remarkably thick; just inside the caldera, the pyroclastic layer has a volume of 4-5 km3, with likely most of the actual volume being distributed over tens of kilometers down the flanks of the volcano, and mostly over the south flank.

Carbonized woods from the bottom of the Hinodehama Ignimbrite have yielded two identical ages of about 750-800 BC, although, given the uncertainty in the age calibration, it could be anywhere from the year 600 to 800 BC. I think this can be solidly placed as the age of the Ioto caldera-forming eruption. It also matches a large volcanic sulphate spike in ice core records (the ice core signal is from around 700-750 BC); this spike is about 2/3 of the size of Tambora’s 1815 eruption as estimated in terms of the global mean sulphate aerosol depth. If this is indeed the case, then it underscores the potential for many mystery volcanic sulphate events to have come from shallow submarine calderas, particularly in the Tonga-Kermadec and Izu-Bonin arcs.

The floor of the caldera is uplifted into a massive resurgent dome. Around 800-500 years ago, the caldera floor was only starting to emerge above the water, when the first paleo-shoreline was formed, and since then it has risen about 100 m (an average of 13-20 cm per year). Assuming a ~500 m deep caldera, the average uplift since the caldera formation 2700 years ago has been 25 cm per year. Under this same assumption, the total volume of the resurgent dome is ~34 cubic kilometers, which would yield a magma supply of 0.013 km3/yr or nearly half a cubic meter per second, since the last collapse, which is a remarkable rate, comparable to about half the long-term growth of Reunion Island, or to the long-term growth of the entire Mariana Arc to the south. Most remarkably, Ioto is not alone in its high supply, among its immediate neighbours; Kita-Ioto is one of the top CLVD earthquake producers in the world (rapid piston-like elevation of the caldera floor), Nishinoshima is one of the few most productive stratovolcanoes of this 21st century, and Fukutoku-Okanoba had a VEI-4 eruption in 2021.

During the past decades, the volcano has continued to produce vigorous activity and a rather unusual one. I don’t know of any volcanoes in the world that behave the way Ioto does in two aspects: the shape of deformation, and the frequent tiny phreatic (recently upgraded to magmatic) eruptions along a ring area near the sides of the caldera. The first is that the continuous uplift of the volcano is stronger towards the edges of the caldera than near the center. For example, J604 GPS which is closest to the center of Motoyama has uplifted 7.5 m since July 2015, while X086 which is closer to the edge of Motoyama (further away from the center of the resurgent dome) has uplifted 10 m for the same period, while outside the caldera there’s almost no uplift with J605 GPS that is probably around the edge having only 1 m of uplift during this period. Not only is uplift larger near the edges, but the GPS stations also get pushed inwards, J604 and X086 move towards the center of Motoyama and towards each other, while, strangely enough, J605 (Suribachiyama) moves away from it and from the other GPS stations. Studies have also found that the lowest uplift at present is near the center of Motoyama and in the very tip of Suribachiyama (jumps suddenly on the NE side of Suribachiyama), while the highest is on a ring around Motoyama.

The complex deformation pattern is very hard to explain. The only answer I’ve reached is that the magma chamber is shaped like a bowl, and the sides of the bowl are inflating faster than the bottom of the bowl. This is the same as the cone sheet idea that some models advocate, but I think possibly spanning the bulk of the shallow storage, and I also prefer to use the term “bowl” since it’s more intuitive and less confusing, given the term cone-sheet is often used for thin petal-like intrusions that occur in swarms in igneous petrology which doesn’t seem the same we have here.

The pattern of deformation wasn’t always like this, since Motoyama is the tallest, so for most of the post-caldera period, the center of Motoyama must have been the center of uplift, but this has changed as the magma chamber has developed into a bowl shape, and the edges of the bowl now concentrate deformation. Or at least that is the way it seems to me. There is also more to the deformation. Lengthening between J605 and J604 must reflect growth near the edges of the bowl, while uplift of J604 likely has a contribution from both, but thus has a greater relative contribution from inflation near the bottom of it (which the J605-J604 distance does not have at all), and when both variables are plotted together, an interesting pattern emerges. Inflation of the edges of the bowl (widening between Suribachiyama and Motoyama) is steadier over time; it increases and decreases with the overall inflation of the volcano, but not as much. It seems that magma flow into the bottom of the chamber fluctuates more, but the edges get a steadier supply.

Additionally, there are deflationary events from time to time. These events are associated with deflation centered over the central/northern part of Motoyama (unlike inflation, which centers near the edges), have earthquake swarms, and usually include eruptions (phreatic, except for the current one). During deflation events, the deformation does NOT reverse that of inflation periods, because the deflation is actually strongest over the center of the resurgent dome, and the distance between J604 and J605 increases as during inflation. I believe these events are associated with intrusions above the edges of the magma bowl structure that sometimes result in eruptions and extract magma from the center of the bowl chamber. The current eruption is one of these events and has not only resulted in deflation of Motoyama and distancing between Suribachiyama and Motoyama, but also a (partial?) ring of inflation and fractures around Motoyama. Previous deflation events happened on 2 May 2012, 12 Sept 2018, 10 Oct 2019, 24 Nov 2021, and all but the 2019 event resulted in tiny eruption events (presumably phreatic). The magmatic eruption of October 2023 was instead not associated with any earthquake swarms or deformation and was probably just a batch of magma quietly seeping through, which was chemically similar to that of the caldera-forming eruption from 2.7 ka.

And it’s these eruptions that make up the second main trait of Ioto. Frequent phreatic eruptions, tens of them have taken place over the last decades all across the island, though concentrated in a ring inside the caldera, around the center of Motoyama. This is very weird. Other shallow silicic magma chambers tend to build pressure and culminate in spectacular eruptions; for example, Cordon Caulle, Chaiten, Rabaul, or Havre will, when opening a vent, produce enormous plinian rhyolite eruptions or vast lava flows. But Ioto erupts tiny. Even when trachyte finally reached the surface in 2023, the ensuing eruption was minuscule, so tiny as to not even make a dent in the inflation trend of the volcano. This minor distributed activity all around the resurgent dome is the other strange characteristic of Ioto.

In my head, to my bewilderment, Ioto is not a volcano building pressure for a massive explosion, nor is it a volcano preparing the ground for a ring fault that suddenly unzips into a ring dike and empties the volcano. It’s a magma chamber growing slowly, but consistently, UP, throwing small intrusions through the sides that build an expanding bowl of magma, until the chamber eventually reaches the surface. My perception is that the volcano is not going to need excessive pressure or a massive intrusion to collapse, but that eventually the very magma chamber is going to come up to the surface, through the edges of the caldera structure, and destroy itself. It’s like a train heading towards a cliff with no brakes. It will caldera-collapse, I just don’t know how much track lies ahead. I find it possible that this eruption could be the final drop that overflows the vase, but also that years or decades are left. Though I doubt it could be much more than that, given the extraordinary pattern of deformation and how the situation has been rapidly evolving these past few years.

To end my take on this volcano, I show a map below with the changes during the current crisis, based on InSAR images and GPS data.

By the way, thanks, Jim, for the title suggestion!

Sources:

https://www.jstage.jst.go.jp/article/jgeography1889/94/6/94_6_464/_pdf/-char/en

https://www.jstage.jst.go.jp/article/geosocabst/2024/0/2024_228/_article/-char/ja/

https://www.jstage.jst.go.jp/article/geosocabst/2024/0/2024_229/_article/-char/ja/

Isana Kobune, Youichiro Takada (2024). Uplift of Iwo-jima island during 2007-2023 detected by InSAR and its
physical interpretation: Effect of thermal stress, Japan Geoscience Union Meeting 2024. https://confit.atlas.jp/guide/event-img/jpgu2024/SVC26-13/public/pdf?type=in

Dambly, M. L. T., Samrock, F., Grayver, A., Eysteinsson, H., & Saar, M. O. (2023). Geophysical imaging of the active magmatic intrusion and geothermal reservoir formation beneath the Corbetti prospect, Main Ethiopian Rift. Geophysical Journal International, 236(3), 1764–1781. https://doi.org/10.1093/gji/ggad493

Nagai, M., & Kobayashi, T. (2015). Volcanic History of Ogasawara IOTO (Iwo-jima), Izu-Bonin ARC, Japan. Journal of Geography (Chigaku Zasshi), 124(1), 65–99. https://doi.org/10.5026/jgeography.124.65

Sigl, M., Toohey, M., McConnell, J. R., Cole-Dai, J., & Severi, M. (2022). Volcanic stratospheric sulfur injections and aerosol optical depth during the Holocene (past 11 500 years) from a bipolar ice-core array. Earth System Science Data, 14(7), 3167–3196. https://doi.org/10.5194/essd-14-3167-2022

https://www1.kaiho.mlit.go.jp/kaiikiDB/kaiyo22-2.htm (From the Japan Coast Guard)

https://www.gsi.go.jp/uchusokuchi/20250910ioto-e.html (From the Geospatial Information Authority of Japan (GSI))