Hazardous explosive eruptions of a recharging multi-cyclic island arc caldera

Large silicic volcanic systems can produce catastrophic caldera-forming eruptions with global impacts. Many such systems exhibit recurrent evolutionary paths involving rejuvenation and recharge, a build-up stage, and a climactic eruption with caldera collapse. Eruptions during the early stages, while smaller, still pose significant hazards. Understanding the reactivation timescale of magmatic systems after caldera formation is crucial but challenging due to the burial of early-stage eruption products. Santorini, a multi-cyclic caldera, provides a valuable case study. Its latest caldera-forming eruption occurred around 1600 BCE, followed by intra-caldera volcanism at Kameni Volcano. Historical records document mainly effusive eruptions since 197 BCE. This study uses high-resolution seismic reflection data and IODP Expedition 398 core samples to investigate the submarine architecture and volcanic history of Kameni Volcano, aiming to improve our understanding of the hazard potential of early-stage caldera eruptions.

Previous research on Santorini has focused on the Minoan eruption (~1600 BCE) and the subsequent effusive activity of Kameni Volcano. Studies have identified stratigraphic units within the caldera infill, but knowledge of early-stage post-collapse eruptions remained limited. Existing seismic reflection studies identified three stratigraphic units related to Minoan and post-Minoan deposits. While historical accounts describe Kameni eruptions, the scale and impact of these events, particularly submarine ones, remained poorly understood, leading to underestimation of the potential hazard.

IODP Expedition 398 drilled four sites within Santorini's caldera, recovering cores that revealed five lithologic units (L1-L5). High-resolution seismic reflection data, integrated with core lithologies and physical property measurements (P-wave velocities and densities), allowed for detailed mapping of the caldera infill. Three seismic units (S1-S3) were identified and correlated with the lithologic units. The age of these units was constrained by the integration of historical records and the stratigraphic relationships between lava flows and pyroclastic units. Volume estimations of the seismic units were conducted using P-wave velocity data and DRE conversion factors, which were calculated based on moisture and density analyses of the core samples. Geochemical analyses (major and trace element compositions of glass shards) were performed using electron microprobe and LA-ICP-MS to determine the origin and potential source of the erupted material. The data were processed using VISTA, KingdomSuite, and Petrel software.

The study identified a significant submarine explosive eruption in 726 CE, represented by seismic unit S2 (lithologic unit L2), with a volume of 2.0 ± 0.7 km³ (bulk) and 0.53 ± 0.18 km³ (DRE). Combining this with a previously identified distal ash layer (0.4 km³ DRE), the total volume of the 726 CE eruption reaches up to 3.1 km³ (bulk) and ~1 km³ (DRE), resulting in a VEI of 5. This eruption is significantly larger than previously assumed worst-case scenarios (VEI 3-4). The eruption deposits are largely submarine, explaining why they have gone largely unnoticed until now. The base of seismic unit S3 is significantly deeper in the northern basin than the southern basin, indicating a pre-existing topographic step which coincides with the Kameni line, a volcano-tectonic lineament. Seismic unit S3 shows clear onlap structures which suggests that its formation post-dates the Minoan collapse. Analysis of the glass shard compositions revealed consistency with the onshore Kameni compositions of the 726 CE eruption. This is also consistent with the age of seismic unit S2, indicating an age between 47 CE and 726 CE. Lithologic unit L3 contains ash and tuffaceous mud possibly from earlier Kameni eruptions. Units L4 and L5 likely represent material from older phases of Kameni Volcano, potentially predating 197 BCE. Seismic unit S1, with a bulk volume of 0.30 ± 0.04 km³ (0.09 ± 0.01 km³ DRE), correlates with post-726 CE eruptions of Nea Kameni. The unconsolidated pumice and ash deposits from the 726 CE eruption form significant parts of the submarine cone of Kameni volcano.

The discovery of the large 726 CE explosive eruption challenges the prevailing view of predominantly effusive activity at Kameni Volcano since the Minoan eruption. It highlights the potential for significant explosive eruptions even in the early stages of a caldera cycle. The unexpected magnitude of this event underscores the limitations of relying solely on onshore eruption records for assessing the hazard potential of submarine eruptions. The 726 CE eruption, if repeated today, would have severe consequences for Santorini and the wider eastern Mediterranean region, including tsunamis, extensive pumice rafts, and large ash plumes. The location of Kameni Volcano on a topographic step and the unconsolidated nature of the 726 CE deposits increase the risk of flank instability and sector collapse, as seen in other volcanoes. The study demonstrates that hazard assessments must consider the possibility of larger-than-expected explosive events, even in seemingly less hazardous phases of caldera evolution.

This study reveals a previously unrecognized, large explosive submarine eruption (VEI 5) in 726 CE at Kameni Volcano, Santorini. This finding significantly alters our understanding of the volcanic hazard at Santorini and other recharging silicic calderas. The magnitude of the 726 CE eruption, combined with the volcano's location and the unconsolidated nature of its deposits, highlights the need for improved monitoring and more sophisticated hazard assessments that incorporate the full range of potential eruptive behavior. Future research should focus on improving our understanding of the processes that trigger such unexpectedly large eruptions in recharging calderas, refine volume estimates through further research, and develop enhanced early warning systems.

The study's interpretations are based on available data, and some uncertainties remain. The acoustic basement was not penetrated during drilling, limiting the understanding of the pre-Minoan deposits. The low recovery rates in some parts of the core slightly reduces the precision of core-seismic integration. While the study provides a comprehensive analysis, unknown quantities of pumice transported in rafts and deposited across the Aegean seafloor could affect the total eruption volume.