Emergent Constraint on CMIP6 Model Projections of Winter Ice‐Free Periods in the Barents–Kara Sea

Abstract

Severe winter ice loss in the Barents–Kara Sea (BKS) affects the local environment and climate in other regions. To reduce sea ice projection uncertainties and reliably identify future ice‐free periods, we apply a time‐varying emergent constraint method to model simulations from the Coupled Model Intercomparison Project Phase 6 (CMIP6). For SSP5‐8.5, the constrained ice‐free date is the multi‐model ensemble mean (MMEM) estimate of 2071–2075 with a 90% confidence range (equivalent to IPCC's “very likely range”) of [2047, 2100+], while it shifts to 2086–2090 with the very likely range of [2054, 2100+] for SSP3‐7.0. For SSP2‐4.5 and SSP1‐2.6, the very likely range for the ice‐free dates are [2080, 2100+] and [2095, 2100+], while ice remains in the optimal constrained MMEM sea ice concentration. Relative to the unconstrained results, the optimal constrained maximum probability of an ice‐free BKS increases notably under SSP5‐8.5 and SSP3‐7.0, but decreases under SSP1‐2.6 and SSP2‐4.5.