Modeling the Winter Heat Conduction Through the Sea Ice System During MOSAiC

Models struggle to accurately simulate observed sea ice thickness changes, which could be partially due to inadequate representation of thermodynamic processes. We analyzed co‐located winter observations of the Arctic sea ice from the Multidisciplinary Drifting Observatory for the Study of the Arctic Climate for evaluating and improving thermodynamic processes in sea ice models, aiming to enable more accurate predictions of the warming climate system. We model the sea ice and snow heat conduction for observed transects forced by realistic boundary conditions to understand the impact of the non‐resolved meter‐scale snow and sea ice thickness heterogeneity on horizontal heat conduction. Neglecting horizontal processes causes underestimating the conductive heat flux of 10% or more. Furthermore, comparing model results to independent temperature observations reveals a ∼5 K surface temperature overestimation over ice thinner than 1 m, attributed to shortcomings in parameterizing surface turbulent and radiative fluxes rather than the conduction. Assessing the model deficiencies and parameterizing these unresolved processes is required for improved sea ice representation. Plain Language Summary Numerical sea ice models rely on conceptual simplifications of the sea ice and snow conditions observed in the Arctic and Southern Ocean. In particular, we cannot account for the variations in sea ice thickness and snow depth at the meter scale because we have limited computing capabilities and lack a detailed understanding of the processes defining how the sea ice system evolves. Furthermore, when designing sea ice models, we assume that the heat can flow only vertically, while in reality, this exchange is also horizontal and depends on the local topography. Thanks to observations collected during the Multidisciplinary Drifting Observatory for the Study of Arctic Climate expedition, we can better quantify model errors when using thermodynamic approximations. Our findings suggest that slightly more heat flows through the sea ice system than we can simulate with a simplified sea ice model and that the surface temperature of the model is too warm for thin ice and snow conditions. Learning the nature of these errors is useful because we could formulate corrections for our models, investigate the occurrence of climate feedback mechanisms, and possibly provide more reliable predictions about the current state of the sea ice and its future evolution, which is currently heavily i