The gas transfer velocity in marginal sea ice zones exerts a strong control on the input of anthropogenic gases into the ocean interior. In this study, a sea state‐dependent gas exchange parametric model is developed based on the turbulent kinetic energy dissipation rate. The model is tuned to match the conventional gas exchange parametrization in fetch‐unlimited, fully developed seas. Next, fetch limitation is introduced in the model and results are compared to fetch limited experiments in lakes, showing that the model captures the effects of finite fetch on gas exchange with good fidelity. Having validated the results in fetch limited waters such as lakes, the model is next applied in sea ice zones using an empirical relation between the sea ice cover and the effective fetch, while accounting for the sea ice motion effect that is unique to sea ice zones. The model results compare favorably with the available field measurements. Applying this parametric model to a regional Arctic numerical model, it is shown that, under the present conditions, gas flux into the Arctic Ocean may be overestimated by 10% if a conventional parameterization is used.

We present 34 profiles of radon‐deficit from the ice‐ocean boundary layer of the Beaufort Sea. Including these 34, there are presently 58 published radon‐deficit estimates of air‐sea gas transfer velocity (k) in the Arctic Ocean; 52 of these estimates were derived from water covered by 10% sea ice or more. The average value of k collected since 2011 is 4.0 ± 1.2 m d−1. This exceeds the quadratic wind speed prediction of weighted kws = 2.85 m d−1 with mean‐weighted wind speed of 6.4 m s−1. We show how ice cover changes the mixed‐layer radon budget, and yields an “effective gas transfer velocity.” We use these 58 estimates to statistically evaluate the suitability of a wind speed parameterization for k, when the ocean surface is ice covered. Whereas the six profiles taken from the open ocean indicate a statistically good fit to wind speed parameterizations, the same parameterizations could not reproduce k from the sea ice zone. We conclude that techniques for estimating k in the open ocean cannot be similarly applied to determine k in the presence of sea ice. The magnitude of k through gaps in the ice may reach high values as ice cover increases, possibly as a result of focused turbulence dissipation at openings in the free surface. These 58 profiles are presently the most complete set of estimates of k across seasons and variable ice cover; as dissolved tracer budgets they reflect air‐sea gas exchange with no impact from air‐ice gas exchange.

is constrained to 10 9 satellite and in situ observations. 12• Strict adherence to conservation laws ensures all sources/sinks can be accounted 13 for, enabling application for meaningful budget analyses. 14• ASTE captures the large-scale dynamics of the Arctic ocean-sea ice system includ-15 ing variability and trends in heat and freshwater storage.

In ice-covered regions it is challenging to determine constituent budgets -for heat and momentum, but also for biologically and climatically active gases like carbon dioxide and methane. The harsh environment and relative data scarcity make it difficult to characterize even the physical properties of the ocean surface. Here, we sought to evaluate if numerical model output helps us to better estimate the physical forcing that drives the air-sea gas exchange rate (k) in sea ice zones. We used the budget of radioactive 222 Rn in the mixed layer to illustrate the effect that sea ice forcing has on gas budgets and air-sea gas exchange. Appropriate constraint of the 222 Rn budget requires estimates of sea ice velocity, concentration, mixed-layer depth, and water velocities, as well as their evolution in time and space along the Lagrangian drift track of a mixed-layer water parcel. We used 36, 9 and 2 km horizontal resolution of regional Massachusetts Institute of Technology general circulation model (MITgcm) configuration with fine vertical spacing to evaluate the capability of the model to reproduce these parameters. We then compared the model results to existing field data including satellite, moorings and ice-tethered profilers. We found that mode sea ice coverage agrees with satellite-derived observation 88 to 98 % of the time when averaged over the Beaufort Gyre, and model sea ice speeds have 82 % correlation with observations. The model demonstrated the capacity to capture the broad trends in the mixed layer, although with a significant bias. Model water velocities showed only 29 % correlation with point-wise in situ data. This correlation remained low in all three model resolution simulations and we argued that is largely due to the quality of the input atmospheric forcing. Overall, we found that even the coarse-resolution model can make a modest contribution to gas exchange parameterization, by resolving the time variation of parameters that drive the 222 Rn budget, including rate of mixed-layer change and sea ice forcings.

A number of feedbacks regulate the response of Arctic sea ice to local atmospheric warming. Using a realistic coupled ocean-sea ice model and its adjoint, we isolate a mechanism by which significant ice growth at the end of the melt season may occur as a lagged response to Arctic atmospheric warming. A series of perturbation simulations informed by adjoint model-derived sensitivity patterns reveal the enhanced ice growth to be accompanied by a reduction of snow thickness on the ice pack. Detailed analysis of ocean-ice-snow heat budgets confirms the essential role of the reduced snow thickness for persistence and delayed overshoot of ice growth. The underlying mechanism is a snow-melt-conductivity feedback, wherein atmosphere-driven snow melt leads to a larger conductive ocean heat loss through the overlying ice layer. Our results highlight the need for accurate observations of snow thickness to constrain climate models and to initialize sea ice forecasts. Plain Language Summary In this study we explore the relationship between Arctic sea ice growth and near-surface air temperature in a modeling framework. By mapping the time-and space-evolving sensitivity of the Arctic ice volume to changes in air temperature, we show that warming at the end of the melting season can lead to thicker ice at the end of the following winter. Warmer air temperatures can melt snow and remove the insulation it provides, exposing the ice surface to subfreezing air temperatures. We show that removal of this insulating snow layer is essential for enhancing sea ice growth later on, by allowing more heat to be conducted up and out of the underlying ocean, supporting seawater freezing. Our results highlight the importance of measuring snow thickness for accurately forecasting Arctic sea ice.