[A-CG34] Projection and detection of global environmental change

[ACG34-04] Comparison of ocean deoxygenation in the last half century between CMIP models and an observational dataset in the North Pacific

*Yumi Abe¹, Shoshiro Minobe¹ (1.Hokkaido University)

Keywords:subarctic North Pacific, inter-ensemble, internal variability, the Sea of Okhotsk, the Bering Sea

Ocean deoxygenation is a phenomenon in which dissolved oxygen (O₂) decreases in the global oceans due to global warming. In the past half-century, according to global observational data analysis, the global O₂ inventory decreased by approximately 2% of the global O₂ (Schmidtko et al., 2017; Ito et al., 2017). No consensus is present on the accuracy of numerical experiments with respect to the observed O₂ decrease. Because discrepancies exist between observations and climate models in the Coupled Model Intercomparison Project Phase 5 and 6 (CMIP5/6), CMIP5 models indicate only 0.6% global O₂ decrease compared with the observations (Oschlies et al., 2018; Grégoire et al., 2021). By contrast, Kwiatkowski et al. (2020) suggested that the global O₂ concentration trend of multi-model ensemble mean of CMIP5/6 models is within the range of different observational estimates. The lack of consensus implies that further studies are needed to understand the relationships between the observed and simulated O₂ changes. It should be useful to compare observations and models in an ocean basin where data availability is relatively high and strong deoxygenation has been observed. Thus, this study investigated the relationship between the observed and simulated dissolved oxygen (O₂) inventory changes in the North Pacific by analyzing an observational dataset and CMIP5/6 model ensembles between 1958 and 2005. A total of 20 models were analyzed for CMIP5 and CMIP6, with one to three ensemble members for each model. While the multi-model ensemble mean is lower than the observed O₂ inventory decreasing trend, several model ensembles showed a decreasing trend higher than the observed value. This result suggests that instead of the forced response that is common for model ensembles, internal variability and model dependency are more important for reproducing the observed trend. An inter-ensemble empirical orthogonal function (EOF) analysis revealed that the different simulated magnitudes of the decreasing O₂ trend is closely associated with the first EOF mode, and ensembles with strong decreasing trends are characterized by large oxygen reduction in the subarctic North Pacific, especially around the boundaries between the North Pacific Ocean and the Okhotsk as well as the Bering Seas. The high O₂ decrease in the subarctic North Pacific is consistent with the spatial pattern of the observed O₂ trend. Therefore, the internal climate variability plays a vital role in the observed strong ocean deoxygenation in the North Pacific. Further analysis of climate models indicated that the O₂ decrease in the subarctic region was primarily caused by physical factors because a significantly high correlation is present between the potential temperature and O₂ inventory trend in the subarctic region, followed by sea ice volume, whereas a minimal correlation coefficient is present between dissolved organic carbon and the O₂ inventory trend. However, the observations have a larger ratio of O₂ inventory trend to temperature trend than any of the ensembles, and thus the relationship between O₂ and temperature change in the subarctic North Pacific seen in the CMIP5/6 simulations is inaccurate.