A Model for Waves-in-Ice and Sea Ice Dynamics in the Marginal Ice Zone

Abstract

Abstract Waves can drastically transform the properties and behavior of the ice cover in a very short time. Consequently, they potentially represent a severe threat for personnel and make offshore operations challenging (for example during platform evacuation) if not anticipated. Waves-in-ice and floe sizes also have implications for the design of marine structures and coastal management. Forecasting waves and their effects on sea ice thus becomes very important for risk assessment, safety and preparedness in the MIZ. In this paper, we present a strategy for including waves-in-ice processes in a sea ice model for a coherent description of the marginal ice zone dynamics. Open ocean wave forecasts are used to provide boundary conditions. Waves are then propagated and attenuated as they travel in sea ice with an attenuation coefficient that is computed by a wave scattering model. A floe breaking parameterization and a renormalization group method are used to provide information about the floe size distribution. Finally, the link between waves-in-ice and dynamics is done through the implementation of a floe size-dependent rheology. Preliminary results are obtained from a high-resolution regional model of the Fram Strait, showing that floe size can be used as to discriminate the dynamical regimes. This model is designed to give access tothe waves-in-ice energy spectrum and information about the floe size distribution asprognostic outputs. Introduction Ocean waves can propagate remarkable distances into ice-coveredseas and contribute to ice breakup a few hundreds kilometers from the ice edge (Squire 2007). By altering the size andshape of individual floes, waves can significantly change the dynamical response of sea ice, especially in the marginal ice zone (MIZ). A MIZ is found where the ice edge is exposed to open ocean as opposed to land. In the Southern Ocean, a MIZ is maintained all around Antarctica by long swell coming from surrounding oceans. The relatively thin first-year ice allows the MIZ to extend few hundreds of kilometers from the ice edge (Lange et al. 1989). In the Arctic, the MIZ is mostly encountered in the Greenland and the Barents Seas where the ice cover is exposed to the open ocean all year round. During summer, ice retreats and waves can affect a larger portion of the central pack. During the recent record minima of ice extent, the Beaufort Sea and even the North Pole areas were transformed into MIZs, a situation that is likely to repeat over the next century. Current sea ice models and forecasting systems are not designed to adequately simulate MIZ processes. Sea ice dynamics are traditionally modeled using a viscous-plastic rheology (Hibler 1979) characterized by a discontinuous flow regime, whereas the ice flow in the MIZ is continuous and is better represented by a granular or collisional than a plastic material (Shen et al.1987, Feltham 2005). This can be partly explained by the coarse resolution of the models used to test sea ice dynamics, where the MIZ was represented by only a few cells, and partly because of a general lack of observational knowledge of ice-wave interactions in natural waters. With a higher resolution, today's coupled ice-ocean models manage to resolve the oceanic mesoscale, which is very prominent near the polar front, and the MIZ, which now cover up to a few tens of grid cells. Despite this gain in resolution, models still miss an adequate representation of MIZ processes.