Global Coupled Climate Response to Polar Sea Ice Loss: Evaluating the Effectiveness of Different Ice‐Constraining Approaches

Coupled ocean‐atmosphere models have been utilized to investigate the global climate response to polar sea ice loss using different approaches to constrain ice concentration and thickness. The goal of this study is to compare two commonly used methods within a single model framework: ice albedo reduction, which is energy conserving, and ice‐flux nudging, which is not energy conserving. The two approaches generate virtually identical equilibrium global climate responses to the same seasonal cycle of sea ice loss. However, while ice‐flux nudging is able to control the sea ice state year‐round, albedo reduction is most effective during summer and lessens the effects of climate change in winter due to the underestimation of sea ice loss. These evaluations have implications for the Polar Amplification Model Intercomparison Project (PAMIP), which proposes a set of coordinated coupled model experiments but without a defined protocol on how to constrain sea ice. Plain Language Summary What does polar amplification mean for global climate? This question has been addressed by using different arbitrary methods to isolate melting of Arctic and/or Antarctic sea ice in coupled ocean‐atmosphere modeling experiments. However, these experiments sometimes obtained different answers. It is still unclear whether the divergent results arise from models' structural differences or from the different methods used to melt sea ice. In this study, two commonly used approaches (albedo reduction and nudging) are applied to a single climate model to generate the same seasonal cycle of ice loss and polar amplification. Nearly identical responses are found, including slowing down the thermohaline circulation in the Atlantic Ocean, weakened westerly winds over high latitudes, increased precipitation over the west coast of North America associated with a deepened Aleutian Low, and tropical upper tropospheric warming. However, ice albedo reduction is found to be ineffective in winter, unlike nudging, which can control sea ice year‐round. We therefore recommend nudging to be adopted for the Polar Amplification Model Intercomparison Project (PAMIP) coupled experiments. Key Points Two ice‐constraining approaches in coupled models (albedo and nudging) to study the climate responses of future ice loss are evaluated Given the same seasonal cycle of ice loss, ice albedo reduction and ice‐flux nudging generate very similar global climate responses Ice albedo reduction underestimates projected winte