Systematic Ocean Ventilation Biases in Climate Models Indicated by Chlorofluorocarbon‐12

Abstract

An accurate representation of ocean ventilation in models is crucial for simulating ocean uptake of heat and carbon, as well as the ocean's supply of oxygen. Using the Inverse Gaussian Transit Time Distribution technique, we assessed ocean ventilation in current climate models via comparing water ages —defined as the time since last contact with the sea surface— derived from observed and simulated distributions of chlorofluorocarbon‐12. Our analysis covers 22 climate models, including 10 coarse‐resolution atmosphere‐ocean coupled models ( COUPLED , with horizontal resolution of ∼1), 8 coarse‐resolution forced ocean‐only models (Ocean Model Intercomparison Project), and 2 high‐resolution eddy‐rich models (with horizontal resolution of 0.1). We identify systematic ventilation biases consistent across most climate models: (a) the deep layer of Subantarctic Mode Water (SAMW, ∼800 m in the South Atlantic) is too old, indicating excessive isolation from the atmosphere; (b) the upper North Atlantic Deep Water (NADW, ∼2,500 m) is too young, whereas the lower NADW (>4,000 m) is too old; and (c) the Antarctic Bottom Water (AABW, >4000 m south of 50S) is too old. OMIP models using reanalyzed atmospheric forcings show similar biases to COUPLED models, except for a slightly better ventilated AABW. Increasing horizontal resolution reduces global ocean ventilation biases — particularly in the AABW — thereby enhancing deep‐ocean oxygen supply and increasing anthropogenic carbon storage, which in turn promotes longer‐term carbon sequestration.