Search

The impacts of large terrestrial volcanic eruptions are apparent from satellite monitoring and direct observations1,2. However, more than three quarters of all volcanoes worldwide lie submerged beneath the ocean and the risks they pose to people and infrastructure remain poorly understood due to inaccessibility and a lack of detailed observations before and after eruptions3. Here, comparing data acquired in 2015-20174,5 and three months after the January 2022 eruption of Hunga Tonga Hunga Ha’apai, we document the wide-reaching and diverse impacts of the largest seaborne volcanic eruption in nearly 150 years6. Almost 10 km3 of seafloor material was removed during the eruption, most of which was redeposited within 20 km of the caldera by long run-out pyroclastic currents. These powerful currents damaged seafloor cables over a length of >100 km, reshaped the seafloor, and caused mass-mortality of seafloor life. Biological seafloor communities that survived the eruption were only observed where local topography provided a physical barrier to pyroclastic currents (e.g., on nearby seamounts). While the longer-term consequences of such a large eruption for human, ecological and climatic systems are emerging, we expect that these previously-undocumented refugia will play a key role in longer-term ecosystem recovery7–9.

One way viruses can affect coral health is by infecting their symbiotic dinoflagellate partners (Symbiodiniaceae). Yet, viral dynamics in coral colonies exposed to environmental stress have not been studied at the reef scale, particularly within individual viral lineages. We sequenced the viral major capsid protein (mcp) gene of positive-sense single-stranded RNA viruses known to infect symbiotic dinoflagellates (‘dinoRNAVs’) to analyze their dynamics in the reef-building coral, Porites lobata. We repeatedly sampled 54 colonies harboring Cladocopium C15 dinoflagellates, across three environmentally distinct reef zones (fringing reef, back reef, and forereef) around the island of Moorea, French Polynesia over a three-year period, and spanning a reef-wide thermal stress event. By the end of the sampling period, 28% (5/18) of corals in the fringing reef experienced partial mortality versus 78% (14/18) in the forereef. Over 90% (50/54) of colonies had detectable dinoRNAV infections. Reef zone influenced the composition and richness of viral mcp amino acid types (‘aminotypes’), with the fringing reef containing the highest aminotype richness. The reef-wide thermal stress event significantly increased aminotype dispersion, and this pattern was strongest in the colonies that experienced partial mortality. These findings demonstrate that dinoRNAV infections respond to environmental fluctuations experienced in situ on reefs. Further, viral productivity will likely increase as ocean temperatures continue to rise, potentially impacting the foundational symbiosis underpinning coral reef ecosystems.