Submarine volcanic microbiota record three volcano-induced tsunamis

Submarine volcanic activity can generate destructive tsunamis that affect distant coasts, as highlighted by the 15 January 2022 Hunga Tonga–Hunga Ha'apai eruption and subsequent warnings across the Pacific. While most tsunamis are triggered by large earthquakes, eruptions and submarine landslides can also generate tsunamis, often without precursory ground shaking, complicating prediction. Extensive research after the 2004 Sumatra and 2011 Tohoku-Oki events advanced recognition of tsunami deposits and introduced multiple proxies, yet interpretation remains challenging due to similarities with other coastal processes, emphasizing the need to constrain sediment provenance. Tsunamis triggered by submarine eruptions may carry distinctive biological signatures; for example, detection of specialized bacteria typically restricted to hydrothermal vent systems in coastal settings could provide compelling evidence of tsunami transport from deep-sea volcanic sources. In the semi-enclosed East Sea (Japan Sea), historical tsunami records over the last 2000 years are relatively sparse and predominantly earthquake-sourced, despite the presence of active volcanoes and volcanic islands such as Ulleung Island. The basin’s semi-closed nature reduces external perturbations and favors preservation of geological archives. The aim of this study is to identify geological and biological evidence for tsunamis triggered by volcanic eruptions in the East Sea and to propose a novel criterion—microbial and microfossil tracers—for tracing sediments of volcanic-tsunami origin.

Research on tsunami deposits following the 2004 Indian Ocean and 2011 Tohoku-Oki tsunamis established stratigraphic, sedimentologic, microfossil, and geochemical indicators for recognizing tsunami layers in coastal archives and highlighted challenges distinguishing them from storm deposits. Multiple proxies have been proposed, including diatom assemblages, non-destructive core analyses, and geochemical signatures. However, interpretive ambiguity persists where deposits resemble those of extreme storms or infragravity surges, prompting calls for additional discriminants such as paleogenetic approaches, microtextural and heavy mineral analyses. In the East Sea region, historical tsunamis are less documented and mainly attributed to earthquakes; records of volcanic sources are limited despite known submarine volcanoes and volcanic islands. On the western coast of Japan facing the East Sea, event-stratigraphic studies (e.g., Lake Kamo and Iwafune Lagoon) document multiple Holocene tsunami layers, providing a basis for regional correlation. This study builds on that literature by introducing deep-sea hydrothermal microbiota (e.g., Sulfurimonas) and pelagic silicoflagellates (Dictyocha byronalis) as provenance-specific indicators linking coastal deposits to submarine volcanic sources.

Field and coring: A 14 m-long sediment core (20HH01) was recovered on 08 December 2020 from Lagoon Hyangho (128°48′38.37″E, 37°54′40.09″N; water depth ~1.7 m) on the eastern coast of the Korean Peninsula. Coring was conducted from a barge; sediments were retrieved in a PVC liner, split lengthwise, and subsampled for AMS 14C dating, bacterial diversity, microfossil analyses, and XRF scanning. Computed tomography images were used to visualize internal structures (referenced in figures). Age dating: AMS radiocarbon dating (plants, shells, humic materials) was performed at KIGAM. Calibrations were done using OxCal (http://c14.arch.ox.ac.uk). Reported ages are calibrated (cal. kyr BP) with stated uncertainties (±20 where noted). Microbial analyses: DNA was extracted from 0.25 g sediment (DNeasy PowerSoil Pro Kit, Qiagen). The V3–V4 region of the 16S rRNA gene was amplified using primers 16S-F (CCTACGGGNGGCWGCAG) and 16S-R (GACTACHBGGGTATCTAATCC). PCR used KAPA HiFi HotStart ReadyMix with an initial denaturation at 95°C for 1 min; 34 cycles of 95°C 30 s, 55°C 30 s, 72°C 30 s; final extension at 72°C for 5 min. Products were purified with Agencourt AMPure XP. A second PCR added Nextera XT indices. Amplicons were quality-checked (gel electrophoresis), quantified (Qubit dsDNA HS on Qubit 3.0), pooled, checked (Agilent 2100 Bioanalyzer), and quantified by qPCR (Bio-Rad CFX96). Sequencing was on Illumina MiSeq (Bionics, Korea). Taxonomic profiling used EZBioCloud MTP pipeline with in-house QC; OTUs clustered by UCLUST at ≥97% identity; taxonomic assignment used the PKSSU4.0 database. Alpha diversity indices (Chao1, ACE, Shannon, Simpson) were calculated. To minimize sequencing bias, bacterial diversity comparisons were normalized to 10,000 reads. Representative 16S sequences from Sulfurimonas-rich horizons were BLASTed and assessed phylogenetically against reference sequences (including hydrothermal vent clones). Raw sequence data are in NCBI SRA SRR27371705–SRR27371753 (BioProject PRJNA1058190). Microfossil analyses: Forty-two samples (~1 g each) were dried at 60°C for 24 h, oxidized with 30% H2O2 (20 mL) to remove organics, rinsed, mounted in Pleurax, and examined via light microscopy (Nikon Eclipse Ni; images by DS-Ri2). Peroxide-cleaned residues were also filtered onto 2.0 μm polycarbonate membranes, coated with Au-Pd, and observed using FE-SEM (TESCAN MIRA 3). Identification employed standard taxonomic literature; morphometrics were measured using ImageJ/NIS-Elements. Geochemistry: Split-core surfaces were covered with SPEXCerti Ultralene film and scanned with an Avatech XRF core scanner (KIGAM): 10 kV/0.25 mA (no filter) for Al–Fe; 30 kV/1.0 mA (thick Pb filter) for Cu–Zr. Elemental ratios (e.g., Sr/Ti, Si/Al, Sr/Ca, Fe/S) were used. Fe/S values were interpreted after accounting for pyrite influence from sulfate-reducing conditions. Stratigraphy and sedimentology: CT imagery and visual core logging documented lithologies (silty clay with interbedded sands), laminations, and deformation features. Depth-age models integrated AMS dates with observed features to define three event intervals (Events I–III).

• Three discrete event intervals were identified in core 20HH01: Event I (11.77–11.10 m; ~8.8–8.3 ka), Event II (9.23–8.10 m; ~7.8–6.5 ka), and Event III (2.30–1.20 m; ~2.5–0.3 ka). Each exhibits transient physical and biological signatures consistent with tsunami deposition rather than gradual environmental change. - Hydrothermal-vent–associated bacteria (Sulfurimonas_f; Epsilonproteobacteria) peak in all three events: averages of 6.21% (Event I), 20.43% (Event II), and 8.69% (Event III) of bacterial communities, while background intervals are dominated by Thiomicrohabdus/Thiomicrospira (avg. 38.31%). Phylogenetic analysis of representative 16S sequences from Sulfurimonas-rich horizons showed closest matches to clones from marine hydrothermal vent fluids, supporting a deep-sea hydrothermal origin. - A pronounced occurrence of Alicyclobacillus ferrooxydans (Firmicutes) was detected only at 11.23 m (~8.3 ka) with 6.8% (normalized) and 9.8% (original) relative abundance, indicative of acidic geothermal environments. - The silicoflagellate Dictyocha byronalis, typical of warm, high-salinity pelagic waters, occurs at 11.23 m, anomalous for a coastal lagoon and consistent with offshore/pelagic input. - Physical sedimentary evidence includes a thin sand layer at 11.20 m (~8.3 ka; Event I), suggestive of tsunami sand incursion, and highly deformed laminations between ~8.8–8.1 m (Event II), potentially from earthquake/tsunami-induced gravitational instability. - Radiocarbon anomalies indicate reworked older materials during Events II and III: ages within these intervals (~10.0–9.6 ka) are ~2000 years older than adjacent sediments (Event II) and ~8000 years older (Event III), consistent with entrainment from pelagic areas with low sedimentation. - XRF-derived Fe/S excursions covary with Sulfurimonas_f peaks and with anomalous AMS 14C ages, implying co-transport of Fe-rich volcanic ash/tephric material and hydrothermal bacteria during events. - Event I (~8.3 ka) aligns temporally with Ulleung Island tephra U-3 (N-3), implicating contemporaneous volcanic activity on Ulleung Island and nearby seabed as the tsunami source. - Correlation with tsunami-event records from Japan’s western coast (Lake Kamo, Iwafune Lagoon) shows temporal matches: Event I (Ev24–Ev22; Ev9), Event II (Ev20–Ev16; Ev7–Ev5), Event III (Ev6–Ev2; Ev1). - Typhoons are unlikely to account for these signatures due to their frequency, limited capacity to affect deep offshore environments, and absence of comparable microbiological markers in typical storm deposits.

The study set out to identify and trace tsunami deposits generated by submarine volcanism using provenance-specific biological and geochemical indicators. The co-occurrence of hydrothermal vent–associated bacteria (notably Sulfurimonas), a geothermal acidophile (Alicyclobacillus ferrooxydans), and a pelagic silicoflagellate (Dictyocha byronalis) within discrete horizons of a coastal lagoon core is inconsistent with normal nearshore sedimentation and indicates episodic delivery from offshore deep-sea volcanic environments. Physical features (tsunami sand, deformation structures) and anomalously old radiocarbon ages corroborate event deposition and reworking/entrainment from pelagic reservoirs. The Fe/S excursions synchronous with Sulfurimonas peaks point to enhanced volcanic ash/iron delivery, reinforcing a volcanogenic tsunami mechanism. Source constraints favor the East Sea and Ulleung Basin; Event I’s synchrony with Ulleung Island tephra (U-3/N-3) ties the tsunami to a known eruption phase, while Events II and III likely reflect other submarine volcanic episodes in the basin. Alternative storm (typhoon) origins are disfavored due to depth limitations of storm influence and the specificity of the biological indicators. Regionally, correlations with event layers on Japan’s western coast suggest that volcanogenic tsunamis in the East Sea impacted both Korean and Japanese coasts during the Holocene, highlighting the broader hazard implications.

This work documents three previously unrecognized Holocene tsunamis recorded in a Korean coastal lagoon and introduces deep-sea hydrothermal microbiota and pelagic silicoflagellates as provenance-specific indicators for volcanic-tsunami deposits. Event I (~8.3 ka) coincides with Ulleung Island’s U-3/N-3 tephra, implicating that eruption and associated seabed activity as the source; Events II and III reflect additional submarine volcanic processes in the East Sea. The combined microbial (Sulfurimonas, Alicyclobacillus ferrooxydans), microfossil (Dictyocha byronalis), sedimentological (tsunami sand, deformation), geochemical (Fe/S), and chronological (anomalous 14C) evidence provides a robust multiproxy framework to identify volcanogenic tsunami deposits in semi-enclosed basins. Future work should (1) validate microbial-source linkages with culture/genome-based analyses, (2) expand spatial coverage to additional East Sea coastal and lacustrine sites for regional correlation, (3) integrate high-resolution tephrostratigraphy and geophysical mapping of submarine volcanic centers, and (4) refine chronological models to quantify recurrence and cascade hazards.

• Source attribution is constrained: multiple submarine volcanoes in the East Sea complicate pinpointing exact sources for Events II and III. - The inference that hydrothermal vent microbiota were tsunami-transported relies on amplicon-based taxonomic profiling; definitive proof requires culture-based and genome-resolved analyses to exclude alternative provenance or post-depositional contamination. - The study is based on a single core from one lagoon; broader spatial replication is needed to assess regional extent and to rule out site-specific processes. - Radiocarbon anomalies imply reworking; reservoir effects and incorporation of old carbon introduce uncertainties in exact event timing and sediment source. - Storm processes are argued against on mechanistic grounds, but without direct modern analog calibration at the site, some depositional overprinting cannot be entirely excluded.