Reassessing sea ice drift and its relationship to long-term Arctic sea ice loss in coupled climate models

• Published in: Journal of Geophysical Research: Oceans
• Main Authors: Tandon, N.F., Kushner, P.J., Docquier, D., Wettstein, J.J., Li, C.
• Format: Article in Journal/Newspaper
• Language: English
• Published: 2018
• Subjects: Arctic; Fram Strait; Sea ice
• Online Access: https://www.vliz.be/imisdocs/publications/329632.pdf

Results of an earlier study suggest that sea ice drift in climate models is unrealistic, and this has undermined confidence in model projections of long-term (i.e., secular) Arctic sea ice loss. We revisit this by analyzing 22 models participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5). It is shown that, when consistent temporal sampling is applied, sea ice drift speed in models and observations come into closer agreement than previously suggested. There is still considerable intermodel scatter in climatological drift speed, and we show that much of this likely relates to prescribed parameters in the sea ice models. Since 1979, observations show a long-term positive trend of annual mean Arctic average sea ice drift speed resulting primarily from sea ice thinning, and most of the CMIP5 models qualitatively reproduce this. The simulated annual mean drift speed trends reflect strong cancellation between winter trends (which are positive in most models and in good agreement with observations) and summer trends (which are negative in most models and in poor agreement with observations). Positive Arctic average drift speed trends do not consistently coincide with positive trends of Fram Strait outflow. The simulated regional relationship between sea ice strength and drift speed changes dramatically as the Arctic transitions from full to partial ice cover, and this sea ice extent effect likely influences simulated summer drift speed trends. Altogether, these results highlight aspects in which models show encouraging agreement with observations, while pinpointing aspects in which models require improvement.