As I write this article, Ioto/Iwo-Jima erupts anew, ever since October 27, and I turn my attention back to this historically renowned island and dangerously overlooked volcanic system south of Japan, and I turn back my attention to quests I left unfinished and dive yet again into ancient records frozen in ice. This volcano should need no introduction here; it has been featured before in a number of articles. The first to recognize Ioto’s dangerous potential was Henrik article, which put Ioto at the very top of Volcanocafe’s New Decade Volcano program:

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

Albert also wrote about this volcano, adding interesting historical aspects related to its activity:

Iwo Jima in 45 eruptions

About a year ago, on account of another eruption at the volcano, Tallis and I made extensive analyses of the volcanic system:

Crisis on Ioto

Nonetheless, let’s sum up some important aspects about Ioto. The volcano is located near the southern end of the Bonin Arc (sometimes also called the Ogasawara Arc), a chain of mostly submarine volcanoes formed where the oceanic lithosphere of the Pacific dives under the oceanic lithosphere of the Philippine Sea Plate. To its south lies Fukutoku-Oka-no-Ba volcano, and to the south of the latter volcano the Mariana Through starts, an area of ridge spreading in the back-arc of the Mariana Arc, where new crust is being created. Ioto features a very rare chemistry for an arc volcano in that it lies well into the alkaline field of volcanic rocks, erupting lavas like trachyte and trachyandesite, making it part of an exclusive club of main-arc alkaline volcanoes that includes some of the Italian volcanoes, Tolbachik, Tambora, and few others.

The unimposing edifice emerges above the water as one of the small Bonin islands, and like another noteworthy volcanic island of the arc, Nishinoshima, it has been growing in size during the past few decades. But while Nishinoshima grows through eruptive activity, Ioto grows through uplift driven by magma accumulation below the surface. The island is floored with pyroclastic deposits from its caldera-forming eruption around 2,700 years ago, yet its top forms the summit of a dome that stands more than 100 meters above the highest point of the caldera rim, and some 200-300 m above the deepest parts of the caldera floor near the edges of the caldera (which themselves are probably not as deep as the center of the caldera was left after the major eruption), showing how over the past 2,700 years the central part of the caldera has risen a few hundred meters, probably well over 300, as magma accumulated beneath. A magma volume of definitely more than 10 cubic kilometers, and probably more than 20 km3, was supplied over this period. Over the past 10 years, inflation rates have been of about a meter per year in parts of the island, a rate that puts other caldera systems of the planet to shame. On account of its prodigious uplift, its former history, and recent change in activity from magma accumulation-phreatic eruptions to magmatic eruptions, we consider Ioto to be a prime suspect for the next major caldera-forming eruption on the planet.

Ioto has been erupting again since October 27 from the same location off the SE coast of the island that the 2023 eruption came from. The eruption is also similar. Thus far, weak phreatomagmatic eruptions have built a tuff cone and are producing incandescent ejecta. The main danger is that the eruption could intensify and initiate a runaway towards a caldera collapse. I will discuss more about the future at the end of the article. But now I want to look at the past, which is the key to the future. One of the unresolved questions I’ve been dragging around for a year is where exactly Ioto’s caldera-forming eruption lies in the ice cores. And I say ice cores, because snowfall in the polar regions traps past volcanic chemistry signatures within the ice caps of Greenland and Antarctica; these environmental records have been recovered through drilling and possess invaluable information. Knowing exactly what Ioto’s last eruption was like might give an idea of the scale of the event that is upon us, and if this fateful piece of land could change history once again.

The radiocarbon date of Ioto’s CFE

To search for Ioto’s signal, figuring out the eruption age is fundamental. There’s radiocarbon data available, so that should be the first step. Two radiocarbon samples are available (Nagai & Kobayashi, 2015), which come from good material: the outermost ring of carbonized wood and a carbonized twig found within the explosive eruption deposits themselves. The standard deviation is low, which helps constrain the age better. I calibrated the radiocarbon age, a necessary step due to the changing carbon-14 concentration in the atmosphere through time.

The image above shows the calibration with the OxCal software of the average of the two samples. There is a 95.4% probability that the eruption happened in one of the following periods: 795-752 BC, 683-668 BC, 632-623 BC, 612-592 BC. The years with the highest probability are those around 780-770 BC. Despite the data being so accurate, there is scatter because the date grazes the Hallstatt plateau, or as archeologists call it, the first millennium BC radiocarbon disaster. This “plateau” refers to an 800–400 BC flat segment in the calibration curve where, no matter how accurate a radiocarbon date, the possible age span covers this whole period. This was due to a period of lowering C14 concentrations in the atmosphere; 800-BC-old plant remnants would have decayed away carbon-14 at about the same rate as it was being depleted from the atmosphere, so for these four centuries any plant remains would now have the same C14. Luckily, it’s now exactly in the plateau, so the number of possible years is relatively restricted, and there’s a way around this too as we will see now.

The 2.7 ka mystery (or not-so-mystery-anymore) Asian tephra

The other key evidence as to the age of the Ioto eruption came while I was looking into other volcanic topics. I stumbled upon an article about cryptotephra layers in Jeju Island (Sun, C., Plunkett, G., Park, J., Guo, Z., & Liu, J.; 2024) that mentioned a mystery East Asian tephra marker. A cryptotephra is a tephra so thin that it’s not visible but can be detected through laboratory techniques. The mystery cryptotephra in question is a layer of trachyte ash that has been correlated across three different locations, using its similar age and chemistry; those locations being Lake Suigetsu in Japan, Jeju Island in South Korea, and Huguangyan Lake in China.

When I saw trachyte and 2.7 ka (2,700 years ago) mentioned alongside each other, I immediately thought of Ioto, which I was able to rapidly confirm with chemistry. Below, I compare the average compositions of some Ioto eruptive units (Nagai & Kobayashi, 2015) that are part of the CFE (caldera-forming eruption) or closely related, and Lake Suigetsu tephras (D. McLean et al. 2018), including the 2.7 ka layer. Given the good chemical match and the similarity in their ages, both events, the pyroclastic deposits of Ioto and the 2.7 ka East Asian ash layer, can be linked together as the same event: the caldera-forming eruption of Ioto.  The fact that none of the three articles that detected this cryptotephra at their respective sites thought to link it to Ioto’s CFE even though the chemical and chronological data for Ioto’s event were already available, shows how dangerously overlooked a volcano Ioto is.

	SiO2 wt%	TiO2 wt%	Al2O3 wt%	FeOt wt%	MnO wt%	MgO wt%	CaO wt%	Na2O wt%	K2O wt%	P2O5 wt%
Ioto chemistry av.	62.11	0.79	16.20	5.67	0.21	1.17	2.49	6.15	4.88	0.34
Lake Suigetsu 2.7 ka ash	61.76	0.79	16.66	5.65	0.23	1.09	2.57	6.06	4.60	0.36

Being recorded in Lake Suigetsu provides an advantage, because this lacustrine record is extensively radiocarbon-dated, and when these ages are combined with the stratigraphy, they build their own age model. This way, the tephra has been dated to 787- 670 BC in Lake Suigetsu, and this interval should be reliable. When combined with the radiocarbon ages, we can discard the two later, smaller potential time windows, which means that the eruption must almost certainly be limited to 787- 752 BC or 683-670 BC, a very narrow interval to search in the ice cores.

Searching the ice cores

The first thing I did after constraining the potential time windows was to look at the HolVol database for potential signals. I was already aware that there weren’t any particularly robust signals around the time of Ioto’s eruptions because that’s the reason that the search stagnated the first time I tried; however, the size of the potential signals has turned out to be very underwhelming. There’s definitely no Tambora-level signal anywhere around the age range. In fact, between a 1174 BC event and the 426 BC signal that has been suggested to come from Ambrym, none of the eruptions reach sulfate outputs comparable to the strongest events of the last 1,000 years, namely the 1230, 1257, 1458, and 1815 AD eruptions. Within the high-confidence windows of 787- 752 BC or 683-670 BC, there are no powerful volcanic signals in the ice cores AT ALL, they are all below that of a Pinatubo-level eruption. The most powerful signal that could have a remote shot at being Ioto’s is the 572 BC event, which has a sulfate load roughly similar to Huaynaputina 1600 AD eruption, but the radiocarbon of the two charred samples casts a near-zero possibility of this event belonging Ioto’s, and the dating in Lake Suigetsu’s sediments also probably precludes this possibility. Ioto was a “phantom colossus”.

How is it possible that an eruption that created a 10-km-wide caldera, that made an 80-meter-thick pyroclastic density current deposit, and that dusted China, Japan, and South Korea in ash did not leave some massive, very obvious sulfate spike? Well, it’s not the first time I’ve encountered this, and while at first it was shocking, now I see it as more of a common thing. The Millennium Eruption of Changbaishan volcano is one well-dated event that nearly reached VEI-7 size, yet it’s practically invisible in the ice core records of both Greenland and Antarctica, the eruption having been seemingly depleted of sulfur dioxide gas, and therefore could not produce much sulfate aerosol either. The Taupo Ignimbrite and Ilopango Tierra Blanca Joven eruption were two borderline VEI 6-7 events that are chemically linked via ash shards to their respective ice core signals, yet have estimated sulfate loads that are smaller than those of a usual Pinatubo-level eruption.

Why does this paradox occur? It appears to reflect complex dynamics of gas sulfur dioxide content in eruptions. The more evolved a magma is, the lower its SO2 solubility is, to the point that trachytes and rhyolites have very little SO2 solubility compared to more mafic magmas; this likely explains why trachyte-rhyolite eruptions like Ioto, Changbaishan, Taupo, or Ilopango can sometimes produce underwhelming SO2 signals relative to their sizes. If the SO2 becomes exsolved and stays a long time stored close to the surface in a shallow magma chamber, it makes sense that it could escape and become depleted very rapidly. Other volcanoes like Pinatubo-style volcanoes probably have deeper chambers and high volatile fluxes from depth that allow a lot of exsolved gas to still be contained within the rhyolite melt at the time of the eruption. Other volcanoes, like Okmok in its 43 BC eruption, can have a lot of dissolved SO2 gas due to their mafic-intermediate melts that have high SO2 solubility and produce sulfate outputs several times larger than volumetrically bigger silicic eruptions. Added to this, or most of the SO2 might react with water to immediately form sulfuric acid when an eruption takes place through the sea or a lake, as is the case with Ioto’s 2.7 ka CFE and the Taupo Ignimbrite.

Another contributing factor may have been that the caldera collapse was relatively shallow. Not all calderas are equally deep, and some large calderas can collapse only to a few hundred meters deep or less; in fact, I imagine some explosive caldera collapses can be only a few meters deep, though in such situations the explosive eruption would not enter an ignimbrite phase. While I’m confident Ioto’s 2.7 ka event was a VEI 6, it was probably not as voluminous as the 10-km-wide caldera might suggest. On top of that, the sea would have absorbed part of the sulfur output (though the eruption did start subaerially and pumice does float in water, so it seems unlikely this alone was the main factor), and perhaps most importantly, the system may have been SO2-degassed due to storage at shallow depth of highly evolved magma. As such, it was likely a combination of factors, while I lean towards SO2 degassing before the eruption being the main one.

But do I have an actual candidate for the Ioto eruption? I do have a candidate that I cannot fully confirm, but I think it has some good chances. Looking at South Pole ice core data in Antarctica (Dominic A. Winski; 2021), which shows signals very neatly, and some characteristics that are only local, like the nitrate effects, I found a 769 BC event to be the most probable candidate. This is a modest sulfate signal with an amplitude of only about 40 ppm, very little compared to the largest sulfate spikes, which reach amplitudes of hundreds of ppm and have longer durations. However, it does have an unusually strong nitrate drop compared to the modest sulfate spike. Many of the largest sulfate signals in the south pole data show in association to major drops in nitrate concentration. I argued in previous articles that a large nitrate drop was the one distinctive characteristic of Krakatau’s submarine eruption in 1883, and that such potent drops were also characteristic of some of the 12th-13th century events, some of which must have been submarine eruptions, including the likely source of the Loisels Pumice in New Zealand, and that the mechanism through which submarine eruptions cause a disproportionate reduction in nitrate would be ozone depletion due to HCl gas sourced from magma-seawater interaction (similar to the haze in lava ocean entries), ozone depletion would let more UV radiation through which destroys nitrate in the stratosphere:

The largest volcanic explosions from 1500 to 1000 AD

The not very deep, but sharp and long, nitrate drop of the 769 BC event, combined with its position in the highest probability years of the radiocarbon date, makes me think this could be Ioto’s signal, though it is difficult to know for sure.

 

Future

This is actually good news since the same factors that caused Ioto’s 2.7 ka eruption to lack a major sulfate signal might also come into play if Ioto goes for a repeat. Which doesn’t seem at all unlikely right now. There are some sources of uncertainty, though, one being that the exact spike still remains uncertain; another is that the magma characteristics of a repeat could not be identical. For example, if the magma influx has been higher post-2.7 ka than before, then the magma may have had less time to degas this time (though I still think degassing is probably quite intense), or less likely that the current chamber, instead of being homogeneous trachyte like last time, could have a lower pool of more primitive trachyandesite magmas, with potentially more dissolved SO2. Even so, and even if the next one is more SO2-rich and does produce some noticeable solar dimming, or temporarily halts air traffic over a good portion of the globe, I highly doubt the eruption would be catastrophic on a global level. The main threat would be the island itself being obliterated, along with anyone on it, and tsunamis of unclear and variable magnitude. The tsunamis would definitely be destructive and life-threatening in the inhabited Ogasawara Islands, though it’s unclear how destructive they would be further away.

At present, the eruption continues at a low intensity, and as long as it keeps this way, the volcano should stay stable, but this is a situation that can change quickly. There are multiple potential scenarios that I see:

Inflation resumes: This is the scenario where the current active vent stops erupting, and inflation of Motoyama resumes. This inflation would keep the magma supply underground and thus postpone any potential eruptive escalations. But how long the inflation can continue without significant eruptive activity is unclear. In recent years, we’ve seen the first magmatic eruptions, and they’ve come from both sides of the island; it appears the volcano is struggling to keep the magma underground, and that anywhere along the edges of the resurgent dome has become a potential pathway. So it’s unlikely this scenario can hold for long before switching to one of the following two,

Sustained low-intensity activity: I cannot fully rule out that eruptive activity manages to stay at a low level where it can’t initiate the final runaway, for example, if it starts effusing lava flows and builds a shield or lava platform. I struggle to see a trachyte magma with so much water around, and that has already formed a polygenetic conduit, really becoming a stable, sustained eruption, but again wouldn’t fully rule it out.

Explosive runaway: This is the most likely to happen eventually; maybe in weeks or maybe in decades; it’s hard to tell. All it takes is for the magma to build some inertia up the conduit and start a paroxysm to send things flying out of control. Multiple feedbacks would jump into place, first mechanical erosion of the conduit walls, widening the conduit and increasing ascent speed, leading to even more violent explosive activity. If deflation of the volcano is strong enough, it would eventually initiate ring dike formation, which would immediately start an ignimbrite eruption phase with a line of explosive vents cutting open the ring fault of the caldera, and make part of the caldera faults frictionless; the erupting lava would collapse the caldera further, which would open the ring dike further, which in turn would elevate eruption rates and also accelerate mechanical erosion of the walls. Once the ignimbrite phase is entered, all hell breaks loose, and the island ceases to exist in the blink of an eye.

Are there ways in which even substantial explosive activity wouldn’t be able to evolve into a CFE? Maybe, there are, after all many things we ignore about caldera eruptions and how they start, but I do see a very unstable volcano reaching a tipping point right before our eyes.

 

References

Sigl, Michael, et al. “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, vol. 14, no. 7, 12 July 2022, pp. 3167–3196, https://doi.org/10.5194/essd-14-3167-2022

Winski, D. A. (2021) “South Pole Ice Core Sea Salt and Major Ions” U.S. Antarctic Program (USAP) Data Center. doi: https://doi.org/10.15784/601851

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

Sun, C., Plunkett, G., Park, J., Guo, Z., & Liu, J. (2024). Major Holocene cryptotephras layers identified from Jeju Island, Republic of Korea: Implications for regional volcanic eruptions and environmental changes. Palaeogeography Palaeoclimatology Palaeoecology, 655, 112530. https://doi.org/10.1016/j.palaeo.2024.112530

Sun, C., Wang, L., Plunkett, G., Zhang, E., & Liu, J. (2021). An Integrated Late Pleistocene to Holocene Tephrostratigraphic Framework for South‐East and East Asia. Geophysical Research Letters, 48(5). https://doi.org/10.1029/2020gl090582

McLean, D., Albert, P. G., Nakagawa, T., Suzuki, T., Staff, R. A., Yamada, K., Kitaba, I., Haraguchi, T., Kitagawa, J., & Smith, V. C. (2018). Integrating the Holocene tephrostratigraphy for East Asia using a high-resolution cryptotephra study from Lake Suigetsu (SG14 core), central Japan. Quaternary Science Reviews, 183, 36–58. https://doi.org/10.1016/j.quascirev.2017.12.013