[A-OS15] Marine ecosystems and biogeochemical cycles: theory, observation and modeling

[AOS15-10] Interannual variability of sea-air carbon fluxes in the Tropical Pacific Ocean simulated by CMIP5 and CMIP6 models

*Yangchun LI¹, Yongfu XU¹ (1.Institute of Atmospheric Physicas, Chinese Academy of Sciences)

Keywords:Interannual variation, air-sea carbon flux, CMIP, the tropical Pacific

The tropical Pacific plays an important role in the global ocean carbon cycle, firstly because it is one of the strongest carbon sources, and secondly because the strongest interannual fluctuation in sea-air carbon fluxes occur here. Air-sea carbon fluxes (FCO₂) in 18 models of the Coupled Model Intercomparison Project Phase 5 (CMIP5) and 20 models of CMIP6 are examined. Compared with CMIP5, CMIP6 has improved the simulation performance of interannual variations (IAV) of global FCO₂. 13 of 20 models in CMIP6 can simulate the dominant role of the Tropical Pacific (TP) in the IAV of global FCO₂, which is related to El Niño–Southern Oscillation events (ENSO), whereas 7 of 18 models in CMIP5 can represent this phenomenon. There are also many similarities between CMIP5 and CMIP6 models. Those models which failed to simulated the dominate role of the TP in the IAV of global FCO₂ in both CMIP5 and CMIP6 are mainly due to the stronger interannual variability in the Southern Ocean or underestimated interannual fluctuation of dissolved inorganic carbon (DIC) in the surface ocean of the TP. Even for those models which can reproduce the responses of EP FCO2 to ENSO, most CMIP6 models inherit some biases of their CMIP5 versions. For example, the simulated interannual fluctuation of precipitation leads to excessive interannual variations of the DIC concentration and then the partial pressure of CO₂ at the sea surface (pCO_2sea) in the western tropical Pacific, especially in the warm pool. Another inherited disadvantage is the underestimated interannual fluctuation of FCO₂ in the central tropical Pacific from 120ºW to 180ºW. There are two main reasons for the weak interannual fluctuation strength of FCO₂ in the models. One is that most models cannot represent the strong interannual variation of 10m wind speed (vm10) in the central tropical Pacific, and the other is that the interannual fluctuation of pCO_2sea in the central tropical Pacific in most models is too weak, which is related to the overestimated interannual variation of sea surface temperature here. Overall, the shortcomings of CMIP5, regarding the simulation performance of IAV of FCO₂ in the TP, have not been well resolved in CMIP6.