Surface-to-space atmospheric waves from Hunga Tonga-Hunga Ha'apai eruption

On 15 January 2022, the Hunga Tonga-Hunga Ha'apai submarine volcano erupted, producing a vertical plume more than 30 km tall with overshooting tops above 55 km, a record in the satellite era and probably longer. From surface pressure data, the single-event energy release from the initial explosion is estimated between 10 and 28 EJ, probably larger than the 1991 Mt Pinatubo eruption (around 10 EJ) and possibly comparable to Krakatoa in 1883 (around 30 EJ). Large explosions such as volcanoes and nuclear tests are theoretically understood to produce atmospheric waves across a range of scales, including external Lamb waves, acoustic waves and internal gravity waves. Volcanoes can also act as sustained wave sources after the initial eruption via plume convection and associated heating. However, direct observational evidence for long-distance propagation in the free electrically neutral atmosphere of Lamb or gravity waves triggered by volcanoes has been lacking, due to historical limitations in satellite resolution and coverage. The Hunga Tonga eruption presents a unique opportunity to quantify the global wave response to a point-source disruption using modern satellite and ground-based observations spanning from the surface to the ionosphere.

Past major eruptions such as Krakatoa (1883) and Pinatubo (1991) produced strong Lamb waves observable in surface pressure records. Internal waves in the boundary layer have been inferred from seismography, barometry and infrasound for eruptions including El Chichon (1982), Pinatubo, and Okmok (2008). In the free atmosphere, local gravity wave activity associated with plume convection has been observed in mesospheric nightglow (e.g., La Soufrière 2021; Calbuco 2015) and re-analyses of AVHRR cloud imagery over Pinatubo. Ionospheric electron-density disturbances are commonly observed after eruptions, and their magnitude has been proposed as a metric of explosive power. Nonetheless, before the Hunga Tonga event, there was no direct observational evidence of long-distance propagation of Lamb or gravity waves through the free neutral atmosphere from a volcanic source, largely because pre-2000s satellite systems lacked sufficient resolution/coverage and no comparable wave response had occurred since. This study builds on theoretical and observational work on Lamb waves, acoustic–gravity wave propagation, and climatologies of gravity waves from satellite limb and nadir sensors to provide the first comprehensive, global-scale characterization of surface-to-ionosphere wave responses from a single eruption.

The study integrates a global suite of satellite and ground-based measurements to track atmospheric wave propagation from the surface to the ionosphere following the 15 January 2022 eruption. Key datasets and processing steps include: - Geostationary infrared imagery: GOES (including 10.3 µm channel) and Meteosat SEVIRI were used to map brightness temperature (ΔBT) perturbations and track Lamb wavefronts and tropospheric/stratospheric gravity waves. Range rings at 500 km and 2,000 km intervals were overlaid. To reduce meteorological noise, global and antipodal panels applied a 200-km-radius Wiener filter; Andes-region panels used a 400-km boxcar plus a 72-km-radius Wiener filter. - Polar-orbiting thermal IR sounders: AIRS, CrIS, and IASI radiances were analyzed in channels sensitive to specific altitudes: 4.3 µm (~39 ± 5 km), and 15 µm channels sampling ~25 ± 5 km and ~42 ± 5 km separately. These provided height-resolved detection of Lamb and gravity waves and enabled phase-speed estimation at multiple atmospheric levels. - Mesospheric/upper-atmosphere measurements: CIPS Rayleigh albedo anomaly data (~55 ± 5 km) and ground-based hydroxyl airglow imaging over Hawai'i (~87 ± 4 km) identified wave phase fronts at high altitudes. - Surface pressure: A global network of barometric stations provided timing for Lamb wave passage; back-projection from pressure anomalies estimated the trigger time (04:28 ± 0:02 UTC) and near-surface phase speed (318.2 ± 6 m s−1). - Ionosphere: GNSS-derived total electron content (TEC) maps were analyzed to identify travelling ionospheric disturbances (TIDs), their phase speeds, horizontal wavelengths, amplitudes, and arrival times relative to lower-atmosphere signals. Analytical approach: The team identified Lamb waves by their high phase speed, large amplitude, and non-dispersive characteristics, and gravity waves by dispersive behavior with phase speeds below Lamb-wave speeds but near theoretical maxima before total internal reflection. Phase speeds were measured across instruments and locations; arrival times at antipodes and around-the-world transits were recorded. Vertical structure was inferred from the near-identical phase at 25 and 42 km (implying very long vertical wavelengths) and detections at 55 and 87 km. Spatial filtering and careful time–distance analyses distinguished primary from subsequent wavefronts. Observational limitations (e.g., intermittent low-Earth orbit sampling, ash obscuration) were addressed by combining multiple platforms and cross-validating signals across heights and distances.

• Energy and trigger: The eruption released an estimated 10–28 EJ of energy (from surface pressure), exceeding Pinatubo (~10 EJ) and possibly comparable to Krakatoa (~30 EJ). Back-projection indicates a primary trigger at 04:28 ± 0:02 UTC. - Lamb waves: Near-surface phase speed 318.2 ± 6 m s−1; in the stratosphere 308 ± 5 to 319 ± 4 m s−1. The wave circled the globe multiple times, reaching the antipode (Algeria) 18.1 h (±7.5 min) after the eruption and appearing as four distinct wavefronts due to deformation. Partial reflection from the Andes and slowing over South America were observed. Lamb wave phase fronts were coherent from the surface to at least the upper mesosphere/lower thermosphere (~87 km), with height-uniform phase speeds within instrument uncertainties. - Fast gravity waves: Following the Lamb wave, a dispersive packet of fast internal gravity waves with phase speeds 240–270 m s−1 (varying with horizontal wavelength) was observed. The leading phase front had ΔBT amplitude ~0.74 K and λ ~380 km, decreasing to ~0.15 K and ~100 km across the packet. The packet extended ~2,000 km with ~8 phase cycles visible ~7 h after generation. Vertical wavelength was very large (λz >> 110 km), with no phase difference between ~25 and ~42 km, and detections up to ~55 and ~87 km. The packet transited the antipode around 00:30–02:30 UTC on 16 January (20–22 h after eruption). - Propagation gap: A growing gap between the Lamb wave and subsequent gravity waves is consistent with a forbidden phase speed range imposed by total internal reflection. - Additional explosions: Two smaller-amplitude Lamb-like wavefronts within the gap traced back to later origin times, consistent with subsequent smaller explosions also seen in surface pressure. - Ionospheric response: Over New Zealand, three large TIDs were detected: (1) 667 m s−1, λh ~1,000 km, amplitude ~0.1 TECu; (2) 414 m s−1, λh ~700 km, amplitude ~0.4 TECu; (3) 343 m s−1, λh ~400 km, amplitude >0.3 TECu. These did not match the Lamb wave’s phase speed/arrival but a sharp TEC modulation >0.6 TECu at 06:15 aligned with the expected Lamb wave arrival. Over North America, TID 2 and an additional TID (4) with phase speed ~311 m s−1 were observed; strong TEC modulations again matched Lamb wave arrival times. - Sustained post-eruption waves: For nearly 15 h after the initial eruption, concentric gravity waves with ΔBT ~0.5–8 K and λ ~3–65 km were seen over the plume top (GOES) and radiating across the Pacific basin (AIRS/CrIS/IASI), with phase speeds often >100 m s−1 and reaching ~200 m s−1 to Japan before 16:00 UTC. These waves dominated the stratospheric gravity wave spectrum over a radius >9,000 km for >12 h—an exceptional, unique dominance by a single source. - Generation mechanisms: The shallow submarine vent likely enabled flash-boiling of seawater and rapid injection/condensation of water into the stratosphere, producing strong, vertically deep forcing across a broad spectrum. Observed enhancements in stratospheric H2O and H2SO4 (relative to expectations) support this. Subsequent waves were likely driven by convective/mechanical oscillator effects within the plume; tsunami-driven generation was deemed less likely given the regular concentric atmospheric wave morphology, though atmospheric waves did generate meteotsunamis. - Modeling relevance: The event provides a natural experiment for testing and improving gravity wave parameterizations and assessing model representations of winds, temperatures, and density via comparisons of observed vs simulated wave propagation.

The study demonstrates, for the first time, globally propagating Lamb and fast internal gravity waves from a single volcanic source observed continuously from the surface to the ionosphere. The findings address the key question of how the atmosphere responds to a sudden, point-source perturbation: the eruption generated a broad wave spectrum with Lamb waves propagating essentially nondispersively at near-sound speeds and gravity waves propagating at unusually high speeds close to theoretical limits for vertical propagation. The coherence across atmospheric layers, multiple global circumnavigations, and the dominance of a single source over >9,000 km for >12 h are unprecedented. Mechanistically, the shallow submarine context and explosive power likely enabled rapid stratospheric moistening and latent heat release across tens of kilometers in depth, producing vertically deep waves, while continued plume convection drove sustained concentric wave generation. Ionospheric responses, while not simple Lamb-wave counterparts, reveal complex coupling, including high-speed TIDs and sharp TEC modulations at Lamb-wave arrival times, consistent with waves coupling upward as acoustic/gravity waves and then propagating horizontally within the ionosphere. The results have high relevance for atmospheric dynamics and for evaluating and improving weather–climate models, particularly gravity wave source parameterizations and the bulk-atmosphere representation assessed via wave travel-time comparisons.

This work provides the first comprehensive, surface-to-ionosphere characterization of atmospheric waves from the Hunga Tonga-Hunga Ha'apai eruption, revealing globally propagating Lamb waves and unusually fast, vertically deep gravity waves that circled the Earth and dominated the stratospheric gravity wave spectrum over extraordinary spatial and temporal extents. The event serves as a natural experiment for testing models across the atmospheric column and for refining gravity wave parameterizations, with immediate applications to weather and climate simulations. Future research should include high-resolution simulations (e.g., large eddy or specialist models) of the eruption’s wave generation and propagation, systematic comparisons of observed vs simulated travel times to diagnose model winds/temperatures/densities, and studies of ocean–atmosphere coupling including meteotsunami generation and feedbacks. Improved multi-instrument coordination and higher-temporal-resolution observations around such rare events would further clarify generation mechanisms and vertical coupling pathways.

• The wave generation process could not be directly observed due to insufficient temporal resolution during the initial explosion and ash plume obscuration. - Tracking returning gravity waves beyond the antipode was limited by intermittent low-Earth-orbit satellite sampling and confusion with later waves from the eruption and other sources. - Some directional differences in measured Lamb wave speeds are within instrument uncertainty and background wind effects, limiting precise attribution. - Current global weather models lack the spatial/temporal resolution to directly resolve the observed waves (Courant–Friedrichs–Lewy constraints), restricting immediate model reproduction to regional high-resolution simulations. - Attribution of certain ionospheric TIDs to specific lower-atmosphere wave modes/pathways remains uncertain due to complex coupling and indirect propagation routes.