Giant offshore pumice deposit records a shallow submarine explosive eruption of ancestral Santorini

Submarine explosive eruptions are less well understood than their terrestrial counterparts, yet shallow submarine calderas are common along island arcs and can produce highly destructive events. In the South Aegean Volcanic Arc, prior studies documented onland eruptive histories and tephra layers but lacked robust records of submarine volcanism due to limited access to marine deposits. IODP Expedition 398 drilled deep into marine rift basins around the Christiana-Santorini-Kolumbo Volcanic Field (CSKVF) to ground-truth seismic stratigraphy and recover a continuous Neogene–Quaternary record of volcanism. The purpose of this study is to identify, date, characterize, and quantify a newly discovered giant submarine pumice deposit, assess its emplacement mechanisms and water depths, correlate submarine and onland facies, and evaluate implications for volcanic hazards and the eruptive history of the arc.

Previous work has highlighted that shallow submarine eruptions can be extremely violent (e.g., 2022 Hunga Tonga-Hunga Ha'apai) and that pyroclastic currents entering the ocean can transform into water-supported gravity flows that deposit pumice- and ash-rich units on the seafloor. In the South Aegean Volcanic Arc, onland mapping, marine tephrochronology, and seismic imaging provided broad constraints on volcanic and tectonic evolution, including recognition of multiple basins (Anhydros, Amorgos, Anafi, Christiana) and the activity of Santorini, Kolumbo, and Christiana. However, the detailed offshore record of submarine explosive activity remained poorly constrained. Prior Santorini studies distinguished older (>650–550 ka; Early Centres of Akrotiri) and younger (<530 ka to present) geochemical groups and documented numerous Plinian eruptions since 360 ka, including the Minoan. Seismic studies suggested thick basin-filling units, variably interpreted as pyroclastic or mass-transport deposits, but lacked definitive core validation. Experimental, field, and theoretical studies of subaerial flows entering water and purely submarine explosive eruptions provided models for transformation to turbidity currents, pumice rafting and saturation, and dispersal mechanisms, framing the hypotheses this study tests.

• Deep-sea drilling: IODP Expedition 398 (Dec 2022–Feb 2023) drilled 12 sites in and around the CSKVF. Seven sites recovered the newly identified Archaeos Tuff. Multiple holes per site were cored using Advanced Piston Coring or Half-Length APC (core diameter 6.2 cm). Core descriptions followed standard pyroclastic terminology, accounting for coring disturbances. Samples of pumice and ash were collected for petrological and geochemical analyses. Bulk sediments from core catchers were used for micropaleontology and biostratigraphy.
• Bathymetry compilation: A high-resolution DEM merged ASTER, EMODnet, R/V Aegaeo GEOWARN, and R/V Marcus G. Langseth PROTEUS data to map basin morphology and site locations.
• Seismic data and core–seismic integration: Single- and multi-channel seismic profiles from three cruises (2006, 2019) were processed (filtering, deconvolution, SRME, migration). The Archaeos Tuff forms an acoustically chaotic to transparent layer traceable across basins. Shipboard P-wave velocity measurements on wet core samples (mean 1864.8 m/s, SD 168.0 m/s) enabled conversion of two-way travel time to thickness and calculation of deposit volume.
• Volume estimation and DRE conversion: Integration of mapped thickness within the seismic network yielded a bulk volume. Moisture and density measurements (n=74) provided bulk, grain, and solids densities. Accounting for vesicles and intergranular porosity, a DRE conversion factor of 0.341 +/- 0.009 was determined, used to convert bulk to DRE volume.
• Geochemistry and mineralogy: Glass major elements by electron microprobe (15 kV, 6 nA, 10 µm beam; standards Lipari obsidian, VGA99), and trace elements by LA-ICP-MS (193 nm excimer, 24–30 µm spots; internal standard 43Ca, external NIST 612). Phenocrysts (e.g., amphibole) analyzed by EMP. Comparative analyses were performed for onland and offshore samples to establish correlation and petrogenesis.
• Textures and granulometry: Due to core disturbance and fines segregation, minimally affected intervals were targeted for sieving at 1-phi intervals. Pumice vesicularities (connected and isolated) were measured by envelope volume (Geopyc) and helium pycnometry (AccuPyc) to quantify total, connected, and isolated porosity. Hydraulic equivalence was assessed by measuring maximum clast sizes of pumice and lithics on high-resolution core images and estimating settling equivalence in water for waterlogged pumice.
• Biostratigraphy and paleobathymetry: Calcareous nannofossils and planktonic foraminifera were prepared and analyzed using established taxonomies and datums (Geologic Time Scale 2020 and regional schemes). Age-depth models constrained the age of the tuff’s upper contact. Benthic foraminiferal assemblages in adjacent sediments constrained emplacement water depths at each site.

• Discovery and characterization: A laterally extensive rhyolitic pumice deposit, the Archaeos Tuff, was recovered at seven drill sites around Santorini, forming a massive to diffusely bedded pumice- and ash-rich unit with minor lithics. Proximal sites are pumice lapilli–rich; distal sites are ash-dominated.
• Age: Biostratigraphic datums immediately above the unit constrain the eruption age to 520 +/- 10 ka. A 510 ka datum occurs within meters of the upper contact at multiple sites; extrapolated sedimentation rates give ~520 ka for the top of the unit.
• Water depth of emplacement: Paleobathymetry from benthic foraminifera indicates emplacement at water depths of approximately 200–700 m (Christiana and Anhydros basins and Anafi margin) and 500–1000 m (Anafi basin axis).
• Thickness distribution: Core and seismic data show thicknesses up to 150 m in the basin between Christiana and Santorini and in the Anafi Basin. Site thicknesses include: U1593 75 m; U1591 65 m; U1592 50 m; U1599 32 m; U1589 8 m; U1600 6 m; U1598 >46 m.
• Volume: Integration of thickness across the seismic network yields an observed bulk volume of 89 +/- 8 km^3. Shipboard measurements determine a DRE conversion factor of 0.341 +/- 0.009, giving an observed DRE volume of 30 +/- 3 km^3. Extrapolation of log(thickness) vs area to 1 m suggests a distal bulk volume of ~105 km^3, implying the total erupted volume likely exceeds the observed minimum.
• Physical properties: P-wave velocity averages 1864.8 m/s (SD 168.0 m/s). Mean deposit porosity is ~65.9%, consistent with the DRE factor. Pumice lapilli have high vesicularity (total ~75.9%, with ~63.5% connected and ~12.4% isolated). Grain-size metrics show median diameters between -1.9 and 3.3 phi and Inman sorting coefficients of 1.4–2.9 phi in minimally disturbed samples, better sorted and depleted in fines relative to typical subaerial ignimbrites.
• Composition and minerals: Glass compositions are high-silica rhyolite (~78 wt% SiO2) and are essentially identical among 38 offshore and 7 onshore samples. They are distinct from Christiana, Kolumbo, and younger (<530 ka) Santorini products and most similar to Early Akrotiri (>650–550 ka) compositions. Phenocrysts include plagioclase, quartz, cummingtonite, augite, hypersthene, magnetite, ilmenite, and zircon.
• Onland correlatives: Thin, poorly sorted ignimbrite veneers (few meters or less) on Christiani, Santorini, and Anafi islands match the submarine unit geochemically and mineralogically (including cummingtonite, tubular pumice, and lithic assemblages), indicating eruption column breach of the sea surface with deposition on nearby islands.
• Emplacement processes: The unit is a volcaniclastic megaturbidite emplaced by water-supported gravity flows derived from collapse of a powerful shallow submarine eruption column. Erosional basal contacts and basinward thickening support gravity-flow emplacement; sorting and fines depletion imply water-supported flow transformation. Hydraulic equivalence between waterlogged pumice and lithics is consistent with transport in turbidity currents and slurries.
• Source and magnitude: The uniform composition and stratigraphy favor a single large event from shallow submarine vents associated with the ancestral Akrotiri complex. The eruption footprint (>3000 km^2 submarine deposit plus island veneers) and volumes indicate one of the largest events in the South Aegean Volcanic Arc, comparable in DRE magnitude to the Kos Plateau Tuff when using a consistent DRE factor, and six times larger (bulk) than the Minoan pyroclastic-current deposit offshore near Santorini.

The study provides the first core-validated documentation of a giant, Middle Pleistocene submarine pumice deposit in the CSKVF and resolves longstanding ambiguities in seismic interpretations by demonstrating a pyroclastic origin for a basin-wide unit. The integrated biostratigraphy, geophysics, petrology, and textural data show that a high-intensity, shallow submarine eruption around 520 ka produced a dominantly submarine deposit by rapid water entrainment and collapse of the eruption jet into turbidity currents and slurries. The observed sorting, fines depletion, basin-thickening patterns, and erosive bases align with emplacement by water-supported gravity flows rather than primary fallout or prolonged gas-supported currents. The presence of thin, poorly sorted ignimbrites on nearby islands demonstrates episodic breaching of the sea surface, likely during peak discharge or vent shallowing, enabling pyroclastic currents to overrun water and deposit subaerially. Geochemical affinity to Early Akrotiri implies the eruption culminated development of a submarine Akrotiri system and marks a transition to younger Santorini magmatism. The volumes and spatial extent highlight that shallow submarine caldera-forming eruptions can confine most products to the seafloor, leading to substantial under-representation in subaerial records. This recognition recalibrates hazard assessments for the South Aegean arc by revealing capacity for very large submarine explosive events with potential for widespread seafloor impacts and tsunamigenic hazards, analogous to but much larger than recent events such as Hunga Tonga-Hunga Ha'apai.

This work identifies, dates, maps, and characterizes the Archaeos Tuff, a giant submarine volcaniclastic megaturbidite emplaced at 520 +/- 10 ka by a shallow submarine explosive eruption of the ancestral Santorini (Akrotiri) system. Core–seismic integration constrains an observed bulk volume of 89 +/- 8 km^3 (30 +/- 3 km^3 DRE), with likely larger total volume when accounting for distal deposits, ash, pumice rafts, and intracaldera fill. Geochemistry and mineralogy tie submarine deposits to thin onland ignimbrites, demonstrating eruption-column breach of the sea surface. The findings extend the explosive history of the CSKVF, reveal a major buried submarine caldera system, and underscore substantial submarine eruption hazards in a densely populated region. Future work should focus on precisely locating and imaging the source caldera, quantifying distal and co-ignimbrite ash components and pumice raft contributions, refining temporal relationships with Akrotiri and younger Santorini magmatic systems, and incorporating these findings into arc-wide hazard models.

• Total eruptive volume remains a minimum estimate confined to the seismic survey area; contributions from distal flow deposits beyond the study area, co-ignimbrite ash, pumice rafts, and intracaldera tuff are not fully quantified.
• Core recovery and disturbance (e.g., ash liquefaction, fines segregation, fall-in) complicate detailed sedimentological logging and granulometric analyses, necessitating selective sampling of minimally disturbed intervals.
• Biostratigraphic age constraints have depth imprecision due to one-per-core sampling and possible reworking at some sites; age estimates rely on extrapolated sedimentation rates near contacts.
• Paleobathymetry reconstructions may be affected by sediment remobilization and downslope transport of shallow-water species in complex volcanotectonic settings.
• The precise location and geometry of the source caldera remain uncertain; further targeted seismic imaging is needed.