Are 100 Ensemble Members Enough to Capture the Remote Atmospheric Response to +2°C Arctic Sea Ice Loss?

• DOI: 10.1175/jcli-d-20-0613.1
• Publication Date: 2021-02-01T20:45:37Z
• Authors: Yannick Peings; Zachary M. Labe; Gudrun Magnusdottir
• Keywords: Middle latitudes; Ensemble average

Abstract This study presents results from the Polar Amplification Multimodel Intercomparison Project (PAMIP) single-year time-slice experiments that aim to isolate atmospheric response Arctic sea ice loss at global warming levels of +2°C. Using two general circulation models (GCMs), ensemble size is increased up 300 members, beyond recommended 100 members. After partitioning in groups reproducibility evaluated, with a focus on midlatitude jet streams North Atlantic and Pacific. Both atmosphere-only coupled ocean–atmosphere PAMIP are analyzed. Substantial differences found among different experiment subsets, suggesting 100-member ensembles still significantly influenced by internal variability, which can mislead conclusions. Despite an overall stronger response, runs exhibit greater spread due additional ENSO-related variability when ocean interactive. The lack consistency true for anomalies statistically significant according Student’s t false discovery rate tests. problematic multimodel assessment as some may be attributed model sensitivities whereas it variability. We propose method overcome this issue allows more robust conclusions only members used.