Introduction

Phreatomagmatic eruptions, caused by the interaction of magma and external water, are one of the most common, and hazardous, types of volcanic eruptions on Earth1,2. Such eruptions can severely impact areas around volcanoes by generating high-energy pyroclastic density currents (PDCs), strong pressure waves, and tephra-laden jets3,4,5,6. The explosivity of phreatomagmatic eruptions increases when a certain amount of external water is incorporated, and higher energy-exchange efficiency from magmatic heat to mechanical energy is triggered7,8,9,10,11,12. Therefore, the mixing conditions and the ratio of external water interacting with magma are essential to understanding the explosivity and eruption style of subaqueous eruptions. Explosivity and eruption style are also related to the water depth of the eruption and the magma discharge rate13,14,15,16,17. In general, shallow-water environments, in which external water is efficiently involved, yield more-explosive eruptions than do deep-water or dry subaerial ones. Even eruptions with low magma discharge rates can be explosive, as often observed in Surtseyan-type eruptions3,4,5,6. For high-discharge eruptions, the explosivity and eruption style may change dramatically where eruptions occur in a shallow-water environment in which external water is efficiently involved17,18,19,20,21,22,23; conversely, if the eruptions occur in a deep-water environment, the vesiculation of magma is inhibited and the explosivity is suppressed by high pressure14,15,16. Direct observational data from substantial phreatomagmatic eruptions are, unfortunately, limited, and the geological record has provided the most information about the possible surface phenomena of these eruptions17,18,19,20,21,22,23. Recent remote-sensing tools, including satellites and the global infrasound monitoring network have, however, captured some eruptions24,25,26,27,28 with volumes of ~0.01 km3. The explosive submarine eruption that occurred at Hunga Tonga-Hunga Ha’apai volcano (HT-HH) in Tonga on 15 January 2022 may be an example of a large-scale (on the order of 1 km3)29 phreatomagmatic eruption in the shallow sea, but at present, the physical and chemical processes, and related parameters, of that eruption remain unclear. Accordingly, how magma interacts with seawater in the shallow water environment, the relationship between mass discharge rate and plume height in such conditions, and their transition in real space and time are most poorly constrained. The 2021 Fukutoku-Oka-no-Ba (FOB) eruption may be a rare case to reveal the process of explosive interaction between the magma with a high discharge rate and ambient seawater involved with an eruption of this type. Although the 2021 FOB eruption has been studied from geochemistry30 and satellite observations31, our high-resolution time series records and quantitative analyses of eruption parameters allow us to provide a more comprehensive understanding of this explosive shallow-sea phreatomagmatic eruption.

The FOB volcano is one of the active volcanoes in the Izu-Ogasawara arc (Fig. 1a). On 13–15 August 2021, a large explosive eruption occurred at this volcano (Fig. 1b). Prior to the eruption, the depth of the summit of the volcano was 40–50 m b.s.l. (below sea level) with a slightly deeper fissure (~70 m b.s.l.), where the 2021 eruption began, on its northern side (Fig. 1c). We analyse the surface phenomena such as plume shape and their transitions during this eruption using satellite imagery, aerial photos, infrasound data, plume modelling, and geochemistry, and discuss the interaction between magma and ambient seawater, which produced a water-rich eruption column. The eruption began with a sustained plume, which breached the overlying seawater and reached a tropopause height of 16 km a.s.l. (above sea level) for at least nine hours32. As a result, the eruption reclaimed the shallow sea and produced a subaerial tuff cone (Fig. 1c). The eruption also produced a 300 km2 pumice raft seen immediately after the major phase of the eruption (Fig. 1d), which was dispersed more than 1000 km west of the volcano by ocean currents30,31. The pumice raft reached the Pacific Ocean coasts of the Japanese archipelago and caused damage to coastal infrastructure such as ports and harbours in Okinawa. Chemical analyses of the pumice clasts indicate that the eruption consisted of trachytic magma with glass compositions reaching 68 wt.% SiO2. Although some phreatomagmatic explosions with intermediate to silicic compositions have been documented or recorded in the last few decades33,34,35, they have been smaller-volume eruptions with lower magma discharge rates. The FOB eruption was a substantial explosive phreatomagmatic eruption, and was recorded in rare detail.

Fig. 1: Outline of the 2021 Fukutoku-Oka-no-Ba (FOB) eruption.
figure 1

a Location map for FOB volcano. Bathymetry data from NOAA’s National Centres for Environmental Information. b Photo of the eruption at ~15:00 JST on 13 August 2021, taken from 6000 m above sea level and ca. 90 km north of FOB by the Japan Coast Guard. c Bathymetry around FOB prior to the 2021 eruption55. The new islets and the vent area are indicated by red areas and the dashed line, respectively. Photo by GeoEye-1 (Maxar Technologies) on 22 August 2021. d Distribution of the pumice raft (grey zone) extending from the source; image from photo by Himawari-8 (JMA, processed by the Digital Typhoon Project of NICT) at 09:57:30 JST.

Full size image

Results

Timeseries data

Himawari-8 satellite and infrasound remote observations suggest that the FOB eruption began at 05:55 JST on 13 August 2021 (Fig. 2a, b). The eruption occurred in four phases. Phase 1 consisted of a sustained plume phase that began at 05:55 JST on 13 August and lasted for ~14 h with some fluctuations of plume diameter and infrasound signal until ~20:00 JST on 13 August. The period 12:00–19:00 JST was more intense with steady plumes which reached the tropopause. In early Phase 1 (Phase 1a), strong infrasound signals were repeatedly observed, although the plume diameter was still small. After 12:00 JST (Phase 1b), the plume became larger; contrarily, infrasound became relatively weak, except at 14:30–15:00, but continuous. Phase 2 was a 14 h pulsating unsteady phase, characterised by frequent strong infrasound signals and smaller eruption plumes ((iv) with red arrows in Fig. 2b). The period 05:30–06:30 JST on 14 August was more intense with a sustained plume. Phase 3 was unsteady and consisted of 24 h of intermittent weak explosions and sparse strong explosions until 09:00 on 15 August. In Phase 4, the activity decayed with the decrease of plume diameter to less than a few km and mostly ceased in 15 hours since the onset of this phase.

Fig. 2: Timeseries data from remote observations of the 2021 FOB eruption.
figure 2

a Developing eruption plume captured by Himawari-8 (JMA, delivered by NICT) at five times. b Change in the diameter (north-south direction) of the eruption plume captured by Himawari-8 (upper) and infrasound data at 5–15 Hz recorded at Chichijima (bottom). Time is JST. Inverted triangles indicate infrasound signals confirmed to be from the FOB direction. Blue bars indicate times of development of the sustained eruption columns. Red arrows indicate signals of a typical Surtseyan eruption occurred at 8:30 on 14 August 2021 [(iv) in figure (a)]. The eruption is divided into four major phases: Phase 1, the initial intense phase with sustained plumes; Phase 2, the pulsating phase; Phase 3, intermittent and sparse strong explosions; and Phase 4, decay, which is characterised by weak infrasound and little or no plume development from the source, with only ambient clouds being captured. c Magnification of the source area at the beginning of Phase 1. The Himawari-8 images obtained by JMA and processed by the Digital Typhoon Project of NICT.

Full size image

Phase 1 was the most powerful and produced the major eruptive products. It generated a continuous, white-coloured vigorous plume directed to the west. At approximately 08:00 JST on 13 August, the pumice raft was confirmed to be spreading circularly from the source (Fig. 2c). Development of the pumice raft was observed for the first 4 h until the eruption plume covered the source area in the satellite view. After the cessation of Phases 1–3, the pumice raft drifted westward, carried by ocean currents.

Plume characteristics of Phase 1

Plume diameter was at first consistently small, and weak fluctuations correlated with infrasound signals (Phase 1a). From 12:00–20:00 JST on 13 August and 05:30–06:30 JST on 14 August, eruption columns developed (blue bars in Fig. 2b). The eruption columns repeatedly reached ~16 km a.s.l., which corresponds to the measured, local tropopause height (Supplementary Note 1). The most vigorous eruption columns formed in the 1 h following 14:00 JST and in the 20 min from 12:45 JST on 13 August during Phase 1b. These eruption columns were steady and formed thin laterally spreading clouds with a 15–20 km radius at ~16 km a.s.l. and a slightly higher overshooting portion (Figs. 1b and 2a). Infrasound became stronger, corresponding to the development of the large eruption plume at ~14:30; however, it became weak and continuous when the eruption plume reached a steady condition at 15:00–20:00 JST on 13 August. The eruption column was entirely white-coloured, indicating a water-rich eruption. The thin laterally spreading clouds at the tropopause level (Fig. 1b) were similar to an ‘anvil cloud’ or ‘incus’, which is often observed when the upper portion of a strong cumulonimbus spreads out in the shape of an anvil along the tropopause36. These clouds differ from those observed in (Sub-)Plinian eruptions in which more vigorous, thicker pyroclast-laden ‘umbrella clouds’ with a grey/brown colour develop37. The FOB eruption column may have contained fine-grained ash; however, its shape, colour, and spreading behaviour along the tropopause did not provide evidence for a large amount of pyroclasts being suspended in the column. The area covered by the entire eruption cloud reached ~105 km2 at approximately 15:00 JST on 13 August. There is, however, no report or evidence of ashfall on any ships or boats in the downwind area or on the neighbouring Minami Io To island, 6 km south–southwest of the source.

Thermal anomaly

No thermal anomaly was detected in the near-source region, even when Himawari-8 clearly observed the base of the eruption column. An aerial observation by the Japan Coast Guard (JCG) with an infrared camera at 15:00–15:30 JST on 13 August captured hot pyroclastic material being ballistically ejected (Supplementary Fig. 1), but this thermal anomaly was probably too small to be detected by satellites. This situation is similar to the 2018 Anak Krakatau eruption38, in which a thermal anomaly was undetected during the most explosive phase of the shallow submarine eruption that took place after the collapse event. A reasonable interpretation for these cases is that most of the heat issued from the rising magma was consumed in the rapid vapourisation of seawater before it could be detected by thermal monitoring.

Tuff cone and pumice raft formation

The eruption formed a new tuff cone having a crater ~1 km diameter around the vent (Fig. 1 and Supplementary Fig. 2); this cone was rapidly eroded by waves and separated into western and eastern islands. The cone can be seen in a satellite image taken on 14 August. The height of the cone was ~15 m at maximum, and the cone components were massive, poorly sorted loose pyroclastic units, suggesting multiple depositional processes (Supplementary Fig. 2). These islands had disappeared by early 2022. From 15:00–15:30 JST, 13 August, during the intense, sustained plume phase, people on the JCG airplane repeatedly observed laterally spreading PDCs at the source (Supplementary Fig. 1). A major component of the new islands was likely formed via such as partial collapses of the eruption column, perhaps aided by other near-vent depositional processes mainly during Phase 1b. The volume of the tuff cone at the end of eruption is estimated to have been 0.04–0.07 km3.

A brown-coloured pumice raft began to form at 08:00 JST on 13 August and spread against the direction of the wind and ocean currents. The edge of the pumice raft reached ~4 km southeast of the source by 12:00 JST on 13 August (Fig. 2c), suggesting a spreading speed of ~1 km/h upstream when observed before the eruption plume obscured the source area. The raft probably continued to spread during Phases 1 and 2 and then was more passively drifted by ocean currents. The area of the pumice raft reached ~300 km2 at 10:00 JST on 15 August. Most of the pumice raft appeared to originate from the vent location. The growth of the pumice raft cannot be explained by direct fallout deposition from the eruption plume because the plume never covered the upwind area due to the strong easterly wind during the observation period. In the downwind direction, not pumice raft was seen to form, indicating that the eruption plume did not carry a large amount of pumice clasts to be deposited into the pumice raft in the downwind direction. Therefore, most of the pumice raft is thought to have been directly generated from the vent, which was subaqueous during Phase 1. The volume of the pumice raft is estimated to have been 0.1–0.3 km3, based on thickness estimates of 0.3–1.0 m, and the sum of the tuff cone and pumice raft volumes was 0.1–0.4 km3 (0.03–0.1 km3 dense rock equivalent, DRE).

Chemical compositions of magma and SO2 emissions

The whole-rock chemical compositions of the FOB products are 61.7–64.0 wt.% SiO2 and 9.6–11.1 wt.% Na2O + K2O and they are classified as trachyte (Fig. 3). The range of chemical compositions for the 2021 products from our analyses and others30 is the same as those of past eruptions. The chemical compositions of groundmass (GM) glass vary linearly from 56 to 68 wt.% SiO2, reflecting mixing-in of mafic magma, though silicic GM is the main component. The silicic melt inclusions in plagioclase (Pl-MIs) mainly have 65–67 wt.% SiO2. We also measured SO3 for the silicic GM and silicic MIs in plagioclase phenocrysts. The SO3 concentrations of the silicic GM and silicic MIs were 334 ppm (1σ = 336 ppm) and 425 ppm (1σ = 178 ppm), respectively (Fig. 3). The SO3 concentrations were recalculated as SO2 concentrations, and the degassed SO2 was estimated to be 73.3 ppm from the difference between the SO2 concentrations of the silicic GM and silicic MIs.

Fig. 3: Chemical compositions of the 2021 FOB products.
figure 3

a Whole-rock and glass (groundmass and melt inclusion, MI) major element compositions of the 2021 FOB products. Dots connected with a line indicate different portions of the same pumice clast. Black symbols indicate products of previous FOB eruptions. b Groundmass and MI glass SO3 compositions. Inset figure indicates the frequency of the composition for groundmass (red) and MI glasses (blue).

Full size image

Atmospheric SO2 emissions were observed by the TROPOMI instrument on the Sentinel-5 Precursor satellite. The mass of SO2 emitted during the 15 h of activity was 2.1 × 107 kg. Using the SO2 concentration estimated by geochemical analysis and the amount of SO2 measured by satellite, the mass of the erupted magma is estimated to be 2.9 × 1011 kg, corresponding to ~0.11 km3 DRE. This estimate of total erupted volume calculated from the SO2 balance is consistent with the range inferred from estimated deposit and pumice-raft volumes.

Modelling a water-rich eruption plume

The magma discharge rate required to form the 16-km-high eruption column in Phase 1 was estimated using the one-dimensional eruption plume model “Plumeria”39,40, which includes the effect of phase changes on the external water involved in the eruption, as well as for magmatic water. The effect of an amount of pumice that should have provided thermal energy to the plume but did not rise in the plume was also considered in this study (Supplementary Note 1). The relations between the total magma discharge rate Me and the plume height Hp and the given exit velocity are shown in Fig. 4a, b. The corresponding relation between added water fraction \(({m}_{w}^{{{{{{\rm{ext}}}}}}})\) and magma fraction in the plume is shown in Fig. 4c, where the water mass fraction at the vent (mw) varies from 0.5 to 0.95. The results indicate that a magma discharge rate Me of 3–6 × 105 kg/s is sufficient to explain the observed plume height of 16 km if a 5–50% fraction of erupted magma (0.3–3 × 105 kg/s) goes into the plume and contributes as the thermal component released via pumice entrainment. Assuming a nine-hour sustained plume (blue bars in Fig. 2b), the erupted mass of magma is estimated to be 1–2 × 1010 kg, corresponding to 0.004–0.008 km3 in DRE. Because the tuff cone formed by deposition in the proximal area, the volume of tephra deposited as distal fallout tephra into the ocean is considerably reduced compared with these values.

Fig. 4: The relationships between plume heights, eruption rates, and source conditions for the 2021 FOB eruption.
figure 4

a The relation between eruption rate (total magma discharge rate) Me and plume height Hp. b The relation between eruption rates and given exit velocities. Exit diameter was fixed at 100 m for all cases. The water mass fraction (mw) varies from 0.5 to 0.95 as the colours and numbers in c indicate (black: 0.5, light blue: 0.6, orange: 0.7, green: 0.8, blue: 0.9, and magenta: 0.95). The marker types are common among all of a–c, indicating the values of added water fraction \(({m}_{w}^{{{{{{\rm{ext}}}}}}})\) as shown on the horizontal axis of c. The relatively light-coloured cases did not yield 16 km high plumes. The plume height of 16 km is achieved with Me of 3–6 × 105 kg/s (the yellow arrow in a).

Full size image

Discussion

The eruptive volume of the FOB eruption has large uncertainties. Still, the volumes estimated from observations of tephra deposits and pumice raft and the SO2 emissions indicate that it was ~0.1 km3 DRE. This erupted mass, the high eruption plume, and the voluminous pumice raft caused by sustained magma discharge differ from the features of well-observed phreatomagmatic explosions, such as typical Surtseyan eruptions, which are characterised by a series of discrete events with a relatively low magma discharge rate6,9. Phase 1a is represented by frequent infrasound signals, similar to Phases 2 and 3 explosions, and a small-scale eruption plume suggesting that this phase was dominated by the eruptions with a relatively low magma discharge rates. The later, vigorous Phase 1b generated a sustained large eruption column with umbrella and overshooting portions, suggesting that this main part of Phase 1 involved different eruptive processes with a higher magma discharge rate. Plumes from Surtseyan eruptions rarely rise to high altitudes (generally less than tropopause) because their thermal flux is generally low2,9. The magma discharge rate and plume behaviour during Phase 1b are similar to those of subplinian eruptions, characterised by sustained explosive eruptions discharging hot pyroclast and gas mixtures with a tall eruption column and resulting in the widespread dispersion of pyroclasts41,42,43. However, the style of the FOB eruption also differs from these types because most ejecta accumulated near the vent where they formed the tuff cone and pumice raft.

The features of Phase 1 of the FOB eruption may be explained by effective decoupling of coarse pyroclasts from the eruption jet, which we infer to have resulted from interaction of the fragmented magma with seawater. In this process, the gas–pyroclast mixture above the submarine vent (50–100 m b.s.l.) may penetrate the seawater and reach the atmosphere when the magma discharge rate is relatively high ~105–106 kg/s because the effects of water pressure and density change of the jet centre due to mixing with ambient water are limited in such situations44. First, the lighter mixture ingests ambient seawater into the developing eruption jet. As the jet reaches the sea surface that acts as the boundary between two ambient fluids with different densities, it may split into two parts (Fig. 5): (1) a breaching part that enters the atmosphere driven by jet momentum, in which juvenile particles transfer heat rapidly to seawater that changes to steam and contributes to rapid subaerial acceleration of the jet, and (2) a part that remains under water, which has increased density because the jet margins have ingested and mixed with large amounts of seawater. In water, lighter parts of this mixture may form buoyantly spreading currents composed of a slurry of hot water and pyroclasts just beneath the sea surface, while denser parts may form eruption-fed density currents along the seafloor45,46,47, resulting in the formation of the tuff ring on the ocean floor. After the jet mixture enters the atmosphere, there may be a further decoupling of large, poorly fragmented clasts from the cooled gas–pyroclast mixture, resulting in partial column collapse and the generation of PDCs. The remaining central jet entrains additional air to form a buoyant vapour-rich convective plume that carries highly fragmented fine material. A large amount of floating pumice might result from sedimentation from subaerial PDCs and/or directly from the jet margins (Fig. 5). The circular spread of the pumice raft from the source indicates that the slurry discharge was sufficient to drive an upstream current. The volume of the pumice raft is estimated to have reached 50–90% of the total erupted volume. This is similar to a silicic deep submarine eruption in which ~70% of the eruptive products were dispersed as a pumice raft14.

Fig. 5: Idealised near-source processes of the phreatomagmatic explosions during the 2021 FOB eruption.
figure 5

Surtseyan eruptions in the later intermittent phase might have been controlled by the availability of external water vs. shallow eruptions with lower mass discharge rate, due to the build-up of the cone.

Full size image

Phase 1 was followed by the Phase 2 and 3 Surtseyan eruptions, reflecting both shallowing of the vent and a decrease in magma discharge rate (Fig. 5). The transition of eruption style is likely to have been strongly affected by the eruption depth and magma discharge rate. The build-up of the cone to an emergent edifice was probably a key in controlling eruptive transition behaviour from Phase 1 to Phases 2 and 3. Surtseyan eruptions in the later phase might be governed by the more-limited availability of external water vs. shallow eruptions in the open water fuelling subaerial plumes. A similar eruption transition was also proposed partly for the 2016–2017 Bogoslof28,48, the 2009 and 2014–2015 HT-HH24,49, and the 1650 Kolumbo eruptions22. Contrarily, the 2022 HT-HH event has a reverse phase behaviour to the 2021 FOB eruption, where eruptive activity started as emergent Surtseyan intermittent explosions, followed by the enormous explosion and plume production on 15 Jan 202229. In this case, due to more intense explosions than the FOB eruption, the deepening of the submarine vent in the early stage might make the rising magma more explosively interact with seawater in a high magma discharge rate and cause the powerful phreatomagmatic explosion in the later climactic stage. These comparisons indicate that both availability of external water, represented by eruption depth and magma discharge rate are critical in understanding similar or different eruptive behaviours at similar sea depths, particularly for the very shallow-water system.

Phase 1b of the FOB eruption exhibited an eruption style that appears when certain environmental and eruptive conditions (depth of <100 m, magma discharge rate of <106 kg/s, corresponding exit velocity of 100–120 m/s for the vent of 100 m diameter) are met. This generated a buoyant vapour-rich subaerial plume, but we infer that tephra dispersal processes were strongly affected by interaction of magma with seawater in the 100 m between vent and sea surface. It is difficult to categorise this eruption style into previously defined eruption styles such as ‘Surtseyan’ or ‘Plinian’41,42, though other studies have considered the 2021 FOB eruption as typical Plinian30 or Surtseyan31. ‘Phreatoplinian’ may potentially describe this type of eruption; however, ‘phreatoplinian’ was defined based on deposit characteristics, such as extensive dispersion of voluminous fine-grained ash, inferred to reflect intense fragmentation and perhaps higher magma discharge rate19,20,21. The FOB eruption was a phreatomagmatic eruption with a relatively high magma discharge rate compared with the typical ones satellite-observed in recent history24,48; however, we have no evidence that a large amount of fine-grained ash was generated in this eruption. It was probably not as energetic as to cause more intense fragmentation, and the inferred dominance of near-vent depositional processes, with low ash content in the high plume, strongly suggests that widespread dispersal of ejecta was minimal.

The FOB eruption has provided an important opportunity to explore the processes of substantial explosive submarine eruptions that form tropospheric plumes from shallow-water environments. We have comprehensively advanced our understanding of the 2021 FOB eruption using high-resolution time series records of infrasound and satellites and quantitative analyses of eruption parameters such as volumes and mass discharge rates. Our key finding is that even if the major portion of the pyroclastic materials ejected from the submarine vent is not released into the atmosphere, it can cause the eruption plume to grow very large when the magmatic heat is efficiently consumed to vapourise the seawater in the shallow sea environment. Therefore, the relationship between the erupted mass and the plume height in the explosive submarine eruptions is not as straightforward as proposed for on-land eruptions. Applications of Volcano Explosivity Index (VEI) scale50 relative to varying eruptive volume and plume height may not hold. How the mass and heat are partitioned across the boundary between the ambient external water and the atmosphere will be important in understanding the dynamics of submarine eruptions forming high subaerial plumes, such as the 2022 HT-HH event. The surface phenomena, eruptive products, and emplacement processes of such eruptions are likely to significantly differ from both deep submarine eruptions and from dry subaerial explosive eruptions. To examine these differences, it will be necessary to survey submarine deposits thoroughly. Tephra stratigraphy, thickness variation, and sedimentary structure of submarine deposits will be crucial in approaching these issues. Such information is essential to better constrain interpretations of the FOB eruption and enhance our knowledge of submarine and emergent volcanism in general.

Methods

Satellite and infrasound monitoring

A timeseries analysis of the eruption was performed using optical and infrared images captured by the Himawari-8 satellite operated by Japan Meteorological Agency (JMA). For the analysis of the entire shape of the plumes (Fig. 2a), the images obtained from the Himawari Real-time website (https://himawari8.nict.go.jp/en/himawari8-image.htm) operated by National Institute of Information and Communications Technology, Japan (NICT) were used. For the analysis of the proximal area (Figs. 1d and 2c), the data processed from the JMA’s Himawari-8 data distributed via the NICT Science Cloud by the Digital Typhoon Project of NICT were used.

An infrasound station at Chichijima, Ogasawara, 330 km north of the Fukutoku-Oka-no-Ba (FOB) volcano, detected signals of the explosions originating from the source (Figs. 1a and 2b). Infrasonic signals from the FOB eruption were identified by the simultaneous detection at the infrasound station and another seismic station, which is ~1 km away and sensitive to infrasound51, with a corresponding time lag.

Near real-time SO2 estimates (max. column and mass loading) by Support to Aviation Control Service (SACS, Royal Belgian Institute for Space Aeronomy; https://sacs.aeronomie.be)52 using the TROPOMI instrument installed on the Sentinel-5 Precursor satellite (S5P) were used to constrain the SO2 flux. SO2 data were retrieved from TROPOMI backscattered radiance measurements using the S5P operational processing algorithm53. The SO2 slant column densities (SCDs), which represent the effective optical path integral of the SO2 concentration, is first retrieved by applying differential optical absorption spectroscopy (DOAS) to ultraviolet spectra. To deal with possible biases in the spectral retrieval step, a post-processing correction is performed to generate background corrected SCDs. The corrected SCDs are then converted into vertical column densities by means of air mass factors obtained from suitable radiative transfer calculations, accounting for the presence of clouds and aerosols, surface properties and best-guess SO2 vertical profiles53. The SO2 notification was issued by SACS at 4:15 (UTC, 13:15 JST) and it indicated that 3.0 DU (Dobson Units, 1 DU = 2.69 × 1016 mol/cm2) for max. SO2 column, corresponding to 9.873 kt for SO2 mass loading, was emitted by that time, assuming a 15 km eruption plume54. Therefore, considering the onset time of the FOB eruption at ~6:00 JST, the SO2 flux can be estimated to be 1.4 × 106 kg/h for the first 7 h of the eruption. This flux was used to estimate the mass of the SO2 emissions during the 15 h intense phase of the eruption (14 h of the entire Phase 1 and 1 h of the intense period of Phase 2), although it should have an uncertainty caused by estimates of SO2 mass loading and emission duration.

Volume estimation

The volume of the tuff cone was estimated from the topographic change at the source before and after the eruption based on the bathymetry data55 and the location and dimension of the new islands according to the JCG airplanes and satellite observations (Fig. 1c). The dimensions and components of the tuff cone were inferred from photos taken on 12 October by JCG (Supplementary Fig. 2). The volume of the pumice raft was estimated based on its area and thickness immediately after the major eruptive phases. The thickness of the pumice raft was assumed to be 0.3–1 m based on recent submarine eruptions33,56,57. Note that there is a large uncertainty in this raft thickness because little or no validated groundtruthed measurements for a raft over time. In the DRE volume calculation, the bulk densities of the tuff cone (tephra deposit) and the pumice raft were assumed to be 1000–1500 kg/m3 and 500–800 kg/m3, respectively. The sum of the tuff cone and pumice raft volumes is estimated at 0.1–0.4 km3 (0.04–0.1 km3 DRE). On the other hand, the DRE magma volume calculated from the satellite SO2 observation and geochemical data is ~0.11 km3, although this estimation also includes uncertainties caused by observational and analytical processes as mentioned in other subsections. Caution should be paid when these volumes are compared. In many arc volcanoes, the remotely measured SO2 mass is larger than estimated based on geology and petrology that use the tephra volume and the difference between melt inclusion and groundmass glass S concentrations. This discrepancy is generally called ‘excess sulfur’ or ‘excess degassing’58. In the case of the FOB eruption, a portion of S exsolved from melt may have been dissolved into the water. Although the satellite-based estimation and the deposit-based estimation agree in our study, the dissolution of S might have affected this balancing. We discussed this issue in the subsection ‘Errors in SO2 estimation’.

Plume modelling

The plume rise height is primarily determined by the heat flux or mass discharge rate from the source because the development of a plume is driven by the conversion of thermal energy to potential energy. Under high humidity conditions in the atmosphere, latent heat is released during the phase change from steam to liquid water and then to ice as the eruption plume rises, causing an increase in plume height. Another essential feature of this eruption was the existence of a large amount of pumice that should have provided thermal energy to the plume but did not rise in the plume. In plume modelling, we applied the user-friendly plume model ‘Plumeria’39,40, which includes the effect of the phase change of water, to calculate the relation between the eruption parameters and plume height (Supplementary Note 1). To incorporate the thermal energy from the pumice that did not rise in the plume, we adjusted the input parameters of the software to represent the water-rich high-enthalpy mixture at the vent (Supplementary Figs. 3–5 and Supplementary Table 1). To achieve a plume height of 16 km, the necessary total magma discharge rate Me was estimated to be 3–6 × 105 kg/s. The dependence of external water temperature is small compared with that of the mass discharge rate.

Chemical analysis

Pumice samples from the 2021 FOB eruption were taken by the Japan Meteorological Agency at 25° 30.3 N, 138° 53.3 E on 22 August 2021 during a survey cruise and at Minami Daito on 4 October 2021; samples were also acquired by Minami Daito Village at Minami Daito on 8 October 2021 and by our research group on Okinawa Island on 20 November 2021 (Supplementary Fig. 6). We performed microscope observations and conducted whole-rock major element analyses using X-ray fluorescence spectrometry (ZSX Primus II, Rigaku Co., Ltd., Tokyo, Japan) and groundmass and mineralogical analyses using an electron probe microanalyser (EPMA, JXA-8800R, JEOL Ltd., Tokyo, Japan) with an acceleration voltage of 15 kV, a beam current of 12 nA, and a beam diameter of 10 μm at the Earthquake Research Institute, University of Tokyo. On the basis of the microscope observations, all the products from the 2021 FOB eruption include phenocrysts of plagioclase, clinopyroxene, Fe–Ti oxides, olivine, and apatite. Plagioclase is the most abundant phenocryst. Most of the pumice clasts (more than 90%; this value is the same as Yoshida et al.30) are white-grey, although black-coloured portions and individual black pumice are also present. The whole-rock chemical compositions for the 13 FOB pumice samples with textural and colour variations are 61.7–64.0 wt.% SiO2, 1.3–2.5 wt.% MgO, and 9.6–11.1 wt.% Na2O + K2O (Supplementary Data 1) classified as trachyte59. The ranges of the chemical compositions of the 2021 products are the same as those of past eruptions60,61,62,63,64,65, and the result of Yoshida et al.30. The chemical compositions of the groundmass (GM) glass vary over a linear trend ranging from 56 to 68 wt.% SiO2, reflecting a mixture of mafic magma, even though silicic GM is the main component (Supplementary Data 2). The silicic melt inclusions (MIs) in plagioclase have 65–67 wt.% SiO2. We also measured SO3 for the silicic GM (n = 95) and silicic MIs in a plagioclase phenocryst (n = 32) using EPMA. The SO3 concentrations of the silicic GM (Cs_gm) and silicic MIs (Cs_mi) were 334 ppm (1σ = 336 ppm) and 425 ppm (1σ = 178 ppm), respectively. The SO3 concentrations were recalculated as SO2 concentrations, and finally, the degassed SO2 was estimated to be 73.3 ppm (+409.1 ppm/−73.3 ppm if 1σ error is considered) from the difference between the SO2 concentrations of the silicic GM and silicic MIs (Cs_mi − Cs_gm).

Errors in SO2 estimation

The errors in SO2 estimation are caused by some magmatic and eruptive processes and are evaluated below. (1) Effects of uptake during crystallisation and analytical errors: For melt inclusions, we assumed it is unnecessary to consider the change of S concentration during crystallisation because there is no evidence of post-entrapment crystallisation of host minerals, such as systematic compositional change near the melt-crystal contact at least for the analysed crystals. On the other hand, the average value of the S concentration of melt inclusions (Cs_mi) may underestimate the initial S concentration in the magma reservoir because melt inclusions can experience gas leakage and may not reflect the original concentration. For groundmass, the average value of the S concentration of groundmass glass (Cs_gm) may overestimate the residual sulfur after the eruption because only glassy parts were analysed; crystallised parts where sulfur was thoroughly degassed cannot be analysed. Therefore, our petrological estimation of the degassed sulfur (ΔCs = Cs_mi − Cs_gm) indicates a minimal case. The relatively large errors (1σ errors in SO3 measurement by EPMA) may include all these effects during crystallisation and degassing processes. If ΔCs is larger, the volume of erupted magma will be smaller. In the case of the largest error with ΔCs = 409.1 ppm, the volume of erupted magma will be reduced to 1/5–1/6; however, this estimate would be inconsistent with that of eruptive products. (2) Effect of mafic minerals: Mafic minerals such as pyroxenes and olivine contain glassy inclusions (52–54 wt.% SiO2) with higher sulfur concentrations than those of plagioclase shown in Fig. 3. However, the mafic minerals are not major components of the erupted magma, and their contribution to the degassed sulfur is very limited. Even if we consider the effect of mafic minerals, the result is not changed.

Although there are some error sources in the S estimation, we assumed that the S concentration in the magma reservoir just before the eruption is represented by the S concentration of melt inclusions in plagioclase based on the above considerations. Another possible source of the estimation error in S is the process after degassing. Although a portion of S exsolved from groundmass glass during magma ascent might have been dissolved into the seawater, the eruption site is a shallow water environment, and the detailed mixing process between the magma with a high discharge rate and surrounding water cannot be evaluated. In the early stages, the eruption may have been so intense that the timescale for the S migration into seawater was shorter than the timescale for dissociation of SO2 into SO4. At this stage, it is difficult to quantitatively estimate the amount of S loss due to dissolution from erupted products into the seawater. This study assumes the S loss occurs only during magma ascent.