Assessment of the Atlantic water layer in the Arctic Ocean in CMIP5 climate models

Qiang.Wang [ at ]


Atlantic water (AW) plays an important role in the thermal balance of the Arctic Ocean, but thus far there has been no comprehensive assessment of the AW layer in the Arctic Ocean simulated by coupled climate models in the framework of Coupled Model Intercomparison Project (CMIP). In this study we assessed the climatology and the trend of the Arctic AW layer in the historical simulations of 41 CMIP5 climate models. The results show that the CMIP5 intermodel spread is large in terms of simulated hydrography, AW core temperature (AWCT) and AW core depth (AWCD) in the Arctic Ocean. The CMIP5 multimodel means are found to be able to reproduce the main climatological spatial patterns of both the AWCT, which is warm near the Fram Strait and decreases along the AW pathways, and the AWCD, which deepens along the AW pathways. However, similar to standalone ocean-ice models, the CMIP5 climate models also face the common problems of too deep and too thick AW layer. AWCT bias in the Arctic Ocean is related to simulated water properties near the Fram Strait and in the Kara and Barents seas. Models with large AWCT biases are those with large biases in AW temperature in the Fram Strait. The biases of AWCT are also significantly correlated with the ocean temperature in the Kara Sea, which is modulated by winter cooling, hence the mixed layer depth and sea ice cover in the Barents Sea. The CMIP5 models largely underestimate the interannual variability of the AWCT, and the CMIP5-simulated increasing trend of the AWCT in the Arctic Ocean is considerably lower than the observed one since the late 1970s.

Item Type
Primary Division
Primary Topic
Research Networks
Peer revision
ISI/Scopus peer-reviewed
Publication Status
Eprint ID
DOI 10.1007/s00382-019-04870-6

Cite as
Shu, Q. , Wang, Q. , Su, J. , Li, X. and Qiao, F. (2019): Assessment of the Atlantic water layer in the Arctic Ocean in CMIP5 climate models , Climate Dynamics, 53 (9-10), pp. 5279-5291 . doi: 10.1007/s00382-019-04870-6


Download (2MB) | Preview



Research Platforms


Edit Item Edit Item