Submarine cores record magma evolution toward a catastrophic eruption at Kikai Caldera

Caldera-forming eruptions, characterized by massive pyroclastic flows and widespread ash fallout, result from the evacuation of voluminous magma from crustal reservoirs. These events often recur over tens to hundreds of thousands of years. Understanding the magma recharge and differentiation processes leading up to these catastrophic events is crucial. A generalized model proposes a 'caldera cycle' with incubation, maturation, and eruption phases. However, the specifics of magma evolution vary between calderas. Studying past volcanic activity through analysis of volcanic materials provides valuable insights into eruption styles and magma evolution. While proximal volcanic materials offer detailed information, obtaining a chronological sequence can be difficult due to subsequent eruptions. Distal pyroclastic materials, sampled through coring techniques, offer an alternative approach, especially in submarine environments where preservation is better than in subaerial settings. This study focuses on Kikai Caldera, Japan, known for its catastrophic eruptions at 140, 95, and 7.3 ka. The 7.3 ka Kikai-Akahoya eruption (K-Ah), a VEI 7 event, produced extensive pyroclastic flows and ash fallout. The 95 ka Kikai-Tozurahara eruption (K-Tz) was similarly large. Post-K-Ah activity includes subaerial and submarine eruptions. However, the preparatory stages between the K-Tz and K-Ah eruptions are poorly understood due to limited inter-caldera material occurrences, though the Komoriko Tephra provides some evidence of pre-K-Ah volcanism.

Previous research on caldera-forming eruptions has highlighted the challenges in reconstructing magma evolution due to the overprinting of subsequent events. Studies on Yellowstone, Toba, and other calderas have attempted to model the precursory processes. These studies have suggested that magma recharge and differentiation processes play key roles in building up to the next catastrophic eruption. While several models for caldera cycles have been proposed, the specific processes vary significantly from one caldera to another. This underscores the need for detailed studies of specific cases, leveraging different geological records such as submarine cores to complement the often fragmented subaerial records. The Kikai Caldera, with its history of large-scale eruptions, provides an excellent opportunity to investigate these processes.

This study used a submarine core (Site C9036, Hole-B) sampled near Kikai Caldera using the hydraulic piston coring system (HPCS) of the D/V *Chikyu*. The core, 95 m below sea floor (mbsf), was analyzed to reconstruct volcanic activity between the K-Tz and K-Ah eruptions. The core was divided into three units: the AK unit (K-Ah deposits, 19.1–22.4 mbsf), the BW unit (inter-caldera sediments, 22.4–42.5 mbsf), and the TZ unit (K-Tz deposits, 48.5–78.0 mbsf). Radiocarbon dating of foraminifera and seashells in the BW unit provided chronological constraints. Twelve samples were selected for geochemical analysis using laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). This technique allowed for the analysis of glass fragments and minerals to determine their major and trace element compositions. Careful consideration was given to potential alteration effects on the samples. Samples exhibiting anomalously low K₂O or high P, Pb, and U relative to Nb were excluded from the analysis. The study also analyzed the size fractions of samples (>500, 250-500, 120-250 µm) to check for any systematic differences in chemical composition.

Analysis of the submarine core revealed key insights into magma evolution at Kikai Caldera. The K-Tz eruption deposits (TZ unit) contained both felsic and mafic glass fragments, indicating the involvement of mafic magma in this predominantly felsic eruption. This was further supported by the presence of high-An# plagioclase, unlike the low-An# plagioclase found in the subaerial K-Tz tephra. The inter-caldera volcanic activity (BW unit) spanned at least 43,000 years, exhibiting a shift from mafic to felsic magma dominance. However, in the final stage (~9000 years) before the K-Ah eruption, the proportion of highly felsic magma, geochemically similar to K-Ah rhyolite, decreased. This suggests that a significant reservoir of felsic magma accumulated during this period. The chemical compositions of felsic glass fragments in the BW unit mirror those of the K-Ah eruption, indicating the presence of similar magma long before the final eruption. The BW unit also contained a larger proportion of mafic glass fragments compared to the TZ and AK units. The most mafic compositions in the BW unit have an estimated SiO2 content of approximately 62 wt%, indicating a significant range of magma compositions in the inter-caldera period. In the final phase (BW-6 to BW-8), the proportion of highly felsic magma (SiO2 between 73 and 76 wt%) decreased, supporting the hypothesis of felsic magma accumulation.

The findings demonstrate that the K-Tz eruption involved interaction between felsic and mafic magmas, challenging the previous understanding based on subaerial samples alone. The extended inter-caldera volcanic activity reveals a complex history of magma supply and differentiation, with a clear shift towards felsic dominance but a final decrease in the most felsic component before the K-Ah eruption. The observed geochemical similarity between the BW unit (inter-caldera) and the AK unit (K-Ah) suggests that the magma responsible for the K-Ah eruption was being supplied and stored for a considerable period, supporting the concept of magma reservoir growth over time. This study emphasizes the importance of submarine records in revealing the complete picture of magma evolution prior to large-scale caldera-forming eruptions. The similar isotopic features between mafic and felsic rocks suggest that they originated from the same source material but underwent differentiation at shallow crustal levels, possibly with occasional mixing.

This study provides compelling evidence for the role of mafic magma interaction in the K-Tz eruption and reveals a complex, long-term evolution of magma composition leading up to the K-Ah eruption. The decrease in highly felsic magma in the final stages before the K-Ah eruption suggests a significant build-up of felsic magma in a reservoir. Future research could focus on detailed modeling of magma reservoir dynamics and better constraining the sources of mafic and felsic magmas to provide a more comprehensive understanding of caldera cycle processes at Kikai Caldera and other similar systems.

The study acknowledges potential limitations related to the core sampling location and the potential for alteration in some samples. The absence of complete core recovery may have resulted in gaps in the volcanic stratigraphy. While efforts were made to mitigate the effects of alteration, some degree of alteration may still affect the interpretation of geochemical data. Further investigation to better constrain the source locations of mafic and felsic magmas would enhance our understanding of the system.