Venting of a separate CO₂‐rich gas phase from submarine arc volcanoes: Examples from the Mariana and Tonga‐Kermadec arcs

<jats:p>Submersible dives on 22 active submarine volcanoes on the Mariana and Tonga‐Kermadec arcs have discovered systems on six of these volcanoes that, in addition to discharging hot vent fluid, are also venting a separate CO<jats:sub>2</jats:sub>‐rich phase either in the form of gas bubbles or liquid CO<jats:sub>2</jats:sub> droplets. One of the most impressive is the Champagne vent site on NW Eifuku in the northern Mariana Arc, which is discharging cold droplets of liquid CO<jats:sub>2</jats:sub> at an estimated rate of 23 mol CO<jats:sub>2</jats:sub>/s, about 0.1% of the global mid‐ocean ridge (MOR) carbon flux. Three other Mariana Arc submarine volcanoes (NW Rota‐1, Nikko, and Daikoku), and two volcanoes on the Tonga‐Kermadec Arc (Giggenbach and Volcano‐1) also have vent fields discharging CO<jats:sub>2</jats:sub>‐rich gas bubbles. The vent fluids at these volcanoes have very high CO<jats:sub>2</jats:sub> concentrations and elevated C/<jats:sup>3</jats:sup>He and <jats:italic>δ</jats:italic><jats:sup>13</jats:sup>C (CO<jats:sub>2</jats:sub>) ratios compared to MOR systems, indicating a contribution to the carbon flux from subducted marine carbonates and organic material. Analysis of the CO<jats:sub>2</jats:sub> concentrations shows that most of the fluids are undersaturated with CO<jats:sub>2</jats:sub>. This deviation from equilibrium would not be expected for pressure release degassing of an ascending fluid saturated with CO<jats:sub>2</jats:sub>. Mechanisms to produce a separate CO<jats:sub>2</jats:sub>‐rich gas phase at the seafloor require direct injection of magmatic CO<jats:sub>2</jats:sub>‐rich gas. The ascending CO<jats:sub>2</jats:sub>‐rich gas could then partially dissolve into seawater circulating within the volcano edifice without reaching equilibrium. Alternatively, an ascending high‐temperature, CO<jats:sub>2</jats:sub>‐rich aqueous fluid could boil to produce a CO<jats:sub>2</jats:sub>‐rich gas phase and a CO<jats:sub>2</jats:sub>‐depleted liquid. These findings indicate that carbon fluxes from submarine arcs may be higher than previously estimated, and that experiments to estimate carbon fluxes at submarine arc volcanoes are merited. Hydrothermal sites such as these with a separate gas phase are valuable natural laboratories for studying the effects of high CO<jats:sub>2</jats:sub> concentrations on marine ecosystems.</jats:p>