Hunga’s Seafloor Collapsed at Staggering Speed During 2022 Eruption

When Hunga volcano in Tonga erupted on 15 January 2022, it produced the most explosive volcanic event of the modern instrumental era, a blast heard thousands of kilometres away and a pressure wave that circled the globe several times. New research published in Nature Geoscience now documents a decisive and previously under-resolved element of that eruption: the rapid collapse of the volcano’s submarine caldera. By combining bathymetric surveys acquired before and after the climactic phase with seismological and pressure-sensor records, the study reconstructs, minute by minute, how a vast volume of seafloor beneath the ocean surface gave way in the opening hours of the eruption. The work transforms the Hunga event from a spectacular but poorly constrained catastrophe into one of the best-characterized submarine caldera collapses ever recorded, and it carries broad implications for how scientists assess the hazards posed by shallow-marine volcanoes worldwide.

Hunga volcano, which rises from the Tonga-Kermadec arc about 65 kilometres north of Tongatapu, had already signalled its unrest in December 2021, with a sequence of eruptions that built a small temporary island and peppered the surrounding sea with floating pumice. Then, in the early afternoon of 15 January local time, the system transitioned almost instantaneously into a climactic phase. Explosions tore through the shallow edifice, whose vent lay only a few tens of metres below sea level, and the resulting plume punched through the tropopause into the mesosphere, the highest volcanic plume ever reliably documented. What the new analysis makes clear is that the extraordinary violence of the eruption was tightly coupled to the structural failure of the volcano itself: as the eruption evacuated magma from a shallow reservoir, the roof of the system foundered, and the caldera floor dropped by hundreds of metres in a remarkably short interval.

The quantitative core of the study lies in the comparison of multibeam echo-sounder surveys. A regional bathymetric mapping campaign in 2016 had captured the pre-eruption shape of Hunga’s submarine flanks and summit in detail, and repeated surveys carried out in the aftermath of the eruption, together with additional high-resolution mapping, allowed the researchers to compute a precise difference map of the seafloor. The results show that the collapse removed on the order of several cubic kilometres of rock from the edifice, deepening the caldera and reshaping the inner walls. Vertical changes of as much as several hundred metres were measured across broad areas of the caldera floor, confirming that this was a coherent piston-like subsidence of the reservoir roof rather than scattered slumping of loose material on the flanks.

To place that structural change on a timeline, the team turned to the geophysical signals recorded as the collapse unfolded. Seismic stations around the Pacific recorded an extraordinary sequence of very-long-period signals, a slow, rhythmic pulsing superimposed on the eruption’s broadband shaking that previous work had linked to repeated pressure oscillations in the coupled magma-seawater system. The new study shows that these pulses align closely in time with the inferred stages of caldera subsidence, indicating that the collapse did not occur as a single instantaneous drop but proceeded in discrete steps as the caldera block settled incrementally onto the diminishing magma body. Hydroacoustic and barometric records, including the Lamb wave that swept around the atmosphere, provide independent timing anchors, allowing the collapse history to be synchronized with the eruption’s global signals.

The mechanics inferred from these data are consistent with classical models of caldera formation but with an emphatically submarine twist. As magma withdrew rapidly from a shallow reservoir, the overlying roof lost support and failed along ring-shaped fractures. Seawater flooded into the vent region, and the intimate mixing of magma and water drove the eruption’s exceptional explosivity, a process long recognized in theory but seldom observed at this scale. Crucially, the collapse itself may have sustained the eruption: each increment of subsidence decompressed the remaining magma, accelerating volatile exsolution and feeding the next explosive pulse. The result was a positive feedback between structural failure and explosive discharge that explains how a volcano whose vent lay underwater could outshine, in atmospheric effect, any eruption of the satellite era.

The volume budget that emerges from the study also resolves a puzzle about the eruption’s products. Estimates of the ash and pumice deposited on land, on the seafloor, and dispersed through the atmosphere fall short of the volumes implied by the erupted magma in some reconstructions, and the new accounting of collapse volume helps close that balance. The subsided caldera accounts for a large fraction of the material removed from the reservoir, while a substantial volume of fragmented juvenile rock remained trapped as dense deposits within the deepened caldera basin, never reaching the surface as visible ash. This partitioning matters for interpretation: it shows that the caldera acts as a sediment trap that can hide a large share of an eruption’s output from conventional tephra inventories, a bias that likely affects volume estimates at many submarine volcanoes.

Beyond Hunga itself, the findings sharpen hazard assessment for the dozens of shallow submarine volcanoes in the Tonga-Kermadec arc and in comparable settings worldwide, from the Aeolian arc to the Kermadec subduction system’s northern neighbours. Shallow-water vents are notoriously difficult to monitor, and Hunga demonstrated that such systems can escalate from low-level activity to a globally significant climax with little warning. The recognition that caldera collapse can proceed in rapid, stepped fashion within hours, generating devastating tsunami through both displacement of seawater by the collapse and by the pyroclastic flows that raced across the seafloor, gives tsunami modellers a concrete, physically grounded scenario to test. The Hunga event generated tsunami waves that claimed lives as far away as Peru, underscoring that submarine collapse hazards are not confined to the volcano’s immediate shoreline.

Methodologically, the study exemplifies the power of merging datasets that individually capture only fragments of a fast-moving natural experiment. Pre- and post-eruption bathymetry fixes the geometry of collapse; very-long-period seismology constrains its tempo; atmospheric pressure waves and satellite imagery bound the eruption’s explosive energy; and petrological work on the erupted juvenile clasts ties the geophysical record to the magma’s storage conditions and volatile content. None of these threads alone could distinguish between competing interpretations, such as phreatomagmatic excavation versus gravitational sliding, but together they converge on a coherent narrative of reservoir-roof failure synchronously coupled to explosive discharge. The approach offers a template that can be applied retrospectively to other recent submarine eruptions and prospectively in monitoring network design.

The 2022 Hunga eruption severed Tonga’s international submarine cable, blanketed agricultural land in corrosive ash, and displaced communities, even as the exact interplay of processes driving its ferocity remained uncertain for months afterwards. By demonstrating that the seafloor of the volcano collapsed with extraordinary speed during the eruption’s climactic hours, the new research closes a central gap in that understanding and turns a catastrophic event into a benchmark for volcano science. As mapping technology and ocean-bottom instrumentation spread across the world’s volcanic arcs, events like Hunga will increasingly be captured in real time, and the lessons recorded here, about stepped collapse, magma-water feedback and hidden deposit volumes, will shape both the science and the preparedness of coastal nations living above active submarine volcanoes.

The Hunga collapse also provides a rare modern analogue for historical events that shaped early volcanological thinking. The 1883 eruption of Krakatau in Indonesia, which likewise involved a shallow marine vent and generated deadly seaborne tsunami, has long been interpreted as involving caldera-scale destruction of the edifice, but it was documented only through aftermath surveys and eyewitness accounts. The minute-by-minute geophysical record assembled for Hunga offers what those earlier events never could: a direct, instrumentally timed view of how a volcanic edifice fails during a climactic submarine eruption, allowing reinterpretation of historical catastrophes through a modern observational lens.

The stepped character of the subsidence carries particular significance for seismologists. Very-long-period signals of the kind recorded during Hunga are produced by slow, volumetric pressure changes in shallow magmatic and hydrothermal systems, and their rhythmic repetition during the collapse suggests that each increment of roof failure perturbed the underlying magma-seawater system in a repeatable way. Such signals, once calibrated against the bathymetric difference maps, could in principle serve as real-time indicators that a submarine edifice is actively foundering, giving warning of tsunami potential while an eruption is still under way.

The study also illustrates how much of an eruption’s story remains hidden beneath the sea surface. Because Hunga’s vent lay underwater, much of the fragmentation, transport and deposition of juvenile material occurred out of sight, and only the combination of seafloor mapping and geophysical inversion could recover the full mass balance. This hidden component helps explain why eruption magnitude scales derived from atmospheric and tephra observations alone can underestimate submarine events, and why repeated bathymetric surveys should be treated as an essential component of post-eruption response in volcanic arcs.

For the Tonga-Kermadec arc specifically, the results underscore the value of baseline mapping. The 2016 pre-eruption survey proved decisive, and the study implicitly argues for systematic, repeated multibeam coverage of shallow submarine volcanoes before crises occur, since without such baselines the geometry of any future collapse would be irrecoverable.

Subject of Research: Rapid submarine caldera collapse during the 2022 climactic eruption of Hunga volcano, Tonga

Article Title: Rapid submarine caldera collapse during the 2022 climactic eruption of Hunga volcano (Tonga)

Article References: Ribó, M., Cronin, S. J., Park, S.-H., Garvin, J., Yeo, I. A., Clare, M. A., Watson, S. J., Kang, S.-G., Sielfeld, G., Slayback, D. A., Choi, Y., Jung, J., Yoo, J., Hutton, B., Stern, S., White, J. D. L., Brenna, M., Hunt, J., Mackay, K., … Kula, T. (2026). Rapid submarine caldera collapse during the 2022 climactic eruption of Hunga volcano (Tonga). Nature Geoscience. https://doi.org/10.1038/s41561-026-02099-7

Image Credits: AI Generated

DOI: 10.1038/s41561-026-02099-7

Keywords: Hunga volcano, Tonga, submarine caldera collapse, 2022 eruption, caldera formation, bathymetry, very-long-period seismicity, phreatomagmatic eruption, tsunami, magma-water interaction, volcano monitoring, Tonga-Kermadec arc