[A-CG43] Science in the Arctic Region

[ACG43-06] Relationship between Barents-Kara sea-ice decrease and surface warming in the Gulf Stream in CMIP6 models

*Yoko Yamagami¹, Masahiro Watanabe², Masato Mori³, Jun Ono¹ (1.Japan Agency for Marine-Earth Science and Technology, 2.Atmosphere and Ocean Research Institute, University of Tokyo, 3.Research Institute for Applied Mechanics, Kyushu University)

Keywords:Barents-Kara Sea, Gulf Stream, CMIP6, climate model

Winter sea-ice in the Barents-Kara Sea has been rapidly decreasing and affects the Northern Hemisphere climate variability (e.g., Mori et al. 2014). In CMIP5 models, the multi-model mean of the negative sea-ice trend in the Barens-Kara Sea is smaller than the observation, and there remains a large inter-model variance (Li et al. 2017). Recently, "Atlantification" of seawater in the Barents Sea has been observed (Polyakov et al. 2017). Also, atmospheric anomalies related to the Gulf Stream variability affect sea-ice anomalies in the Barents Sea (Sato et al. 2014). We reported that underestimated warming of the SST in the Gulf Stream region leads to the less negative trend of winter sea-ice in the Barents-Kara Sea in a climate model (MIROC6). Therefore, in this study, we further examine the relationship between linear trends of the Barents-Kara sea-ice and Gulf Stream SST using the CMIP6 multi-model ensemble.
First, we analyze linear trends of the sea-ice concentration (SIC) in the Barents-Kara Sea and the Gulf Stream SST (hereafter, we refer to these trends as the SIC and SST trends) in 39 CMIP6 models for 1970-2014. It is found that the multi-model mean of the SIC (SST) trend is less negative (positive) than in observations. We calculate the correlation coefficient for the inter-model variations of the SIC and SST trends. A significant negative correlation (r=-0.59) is found, suggesting that the Gulf Stream warming leads to the sea-ice decrease in the Barents-Kara Sea in CMIP6 models.
To investigate whether the relationship between the SIC and SST trends is due to the external forcing or due to internal variability, we examine 14 models that have more than 10 ensemble members. Forced linear trends in 14 models (as defined by the respective ensemble mean) indicate a significant negative correlation (r = -0.75). In contrast, the SST-SIC correlation is insignificantly low in nine-fourteenths of the other models. This result indicates that the Gulf Stream warming driven by external forcing could explain 56% of the Barents-Kara SIC decrease from 1970–2014.