The propagation of large, storm-generated waves through sea ice has so far not been measured, limiting our understanding of how ocean waves break sea ice. Without improved knowledge of ice breakup, we are unable to understand recent changes, or predict future changes, in Arctic and Antarctic sea ice. Here we show that storm-generated ocean waves propagating through Antarctic sea ice are able to transport enough energy to break sea ice hundreds of kilometres from the ice edge. Our results, which are based on concurrent observations at multiple locations, establish that large waves break sea ice much farther from the ice edge than would be predicted by the commonly assumed exponential decay. We observed the wave height decay to be almost linear for large waves--those with a significant wave height greater than three metres--and to be exponential only for small waves. This implies a more prominent role for large ocean waves in sea-ice breakup and retreat than previously thought. We examine the wider relevance of this by comparing observed Antarctic sea-ice edge positions with changes in modelled significant wave heights for the Southern Ocean between 1997 and 2009, and find that the retreat and expansion of the sea-ice edge correlate with mean significant wave height increases and decreases, respectively. This includes capturing the spatial variability in sea-ice trends found in the Ross and Amundsen-Bellingshausen seas. Climate models fail to capture recent changes in sea ice in both polar regions. Our results suggest that the incorporation of explicit or parameterized interactions between ocean waves and sea ice may resolve this problem.

[1] We present a model for wave attenuation in the marginal ice zone (MIZ) based on a two-dimensional (one horizontal and one vertical dimension) multiple floating elastic plate solution in the frequency domain, which is solved exactly using a matched eigenfunction expansion. The only physical parameters that enter the model are length, mass, and elastic stiffness (of which, the latter two depend primarily on thickness) of the ice floes. The model neglects all nonlinear effects as well as floe collisions or ice creep and is therefore most applicable to floes which are large compared to the thickness and to wave conditions which are not extreme. The solution for a given arrangement of floes is fully coherent, and the results are therefore dependent on the exact geometry. We firstly show that this dependence can be removed by averaging over a distribution of floe lengths (we choose the Rayleigh distribution). We then show that after this averaging, the attenuation coefficient is a function of floe number and independent of floe length, provided the floe lengths are sufficiently large. The model predicts an exponential decay of energy, just as is shown experimentally. This enables us to provide explicit values for the attenuation coefficient, as a function of the average floe thickness and wave period. We compare our theoretical predictions of the wave attenuation with measured data and other scattering models. The limited data allows us to conclude that our model is applicable to large floes for short to medium wave periods (6 to 15 seconds). We also derive a floe breaking model, based on our wave attenuation model, which indicates that we are under-predicting the attenuation coefficients at long periods.Citation: Kohout, A. L., and M. H. Meylan (2008), An elastic plate model for wave attenuation and ice floe breaking in the marginal ice zone,

Key Points:• Long waves attenuate with distance traveled in proportion to the inverse of wave period squared • Significant wave height attenuates with distance traveled into the MIZ • Peak periods increase with distance traveled into the MIZ Abstract In situ measurements of ocean surface wave spectra evolution in the Antarctic marginal ice zone are described. Analysis of the measurements shows significant wave heights and peak periods do not vary appreciably in approximately the first 80 km of the ice-covered ocean. Beyond this region, significant wave heights attenuate and peak periods increase. It is shown that attenuation rates are insensitive to amplitudes for long-period waves but increase with increasing amplitude above some critical amplitude for short-period waves. Attenuation rates of the spectral components of the wavefield are calculated. It is shown that attenuation rates decrease with increasing wave period. Further, for long-period waves the decrease is shown to be proportional to the inverse of the period squared. This relationship can be used to efficiently implement wave attenuation through the marginal ice zone in ocean-scale wave models.

A large collaborative program has studied the coupled air‐ice‐ocean‐wave processes occurring in the Arctic during the autumn ice advance. The program included a field campaign in the western Arctic during the autumn of 2015, with in situ data collection and both aerial and satellite remote sensing. Many of the analyses have focused on using and improving forecast models. Summarizing and synthesizing the results from a series of separate papers, the overall view is of an Arctic shifting to a more seasonal system. The dramatic increase in open water extent and duration in the autumn means that large surface waves and significant surface heat fluxes are now common. When refreezing finally does occur, it is a highly variable process in space and time. Wind and wave events drive episodic advances and retreats of the ice edge, with associated variations in sea ice formation types (e.g., pancakes, nilas). This variability becomes imprinted on the winter ice cover, which in turn affects the melt season the following year.

A precise linear mathematical theory is reported to model the response of a solitary ice floe in ocean waves, allowing the floe to bend with the passing wave. Both infinite and finite water depths are considered, and the model is also extended to include a pair of separated floes of different length, the case of n‐floes then being a natural and straightforward development. For a single ice floe, perfect transmission is achieved whenever the wavelength beneath the ice couples perfectly to the length of the ice floe. Then the strains induced in the bending ice floe reach a maximum, and because multiple‐cycle tuning can occur in floes which are long compared to the wavelength, strain response amplitude operators (RAOs) are complicated. By considering the case of infinite stiffness, heave and roll RAOs are also found for typical ocean wave periods, and these agree well with two‐dimensional rigid‐body models. Finally, ice floes of different diameters but constant 1‐m thickness are subjected to spectral forcing, and the strain spectral density for each is found. Spectral density envelopes increase gradually with floe diameter, achieving a maximum value for a floe of about 102m. Thereafter strains never decrease below those for the 80‐m curve owing to multiple‐cycle tuning. This may explain the presence of zones within an ice field where floe size never exceeds some prescribed value. Results obtained from the complete theory for two adjacent floes do not differ significantly from those found by applying the single‐floe model serially except when the separation is very small.

Wave attenuation coefficients (α, m−1) were calculated from in situ data transmitted by custom wave buoys deployed into the advancing pancake ice region of the Weddell Sea. Data cover a 12 day period as the buoy array was first compressed and then dilated under the influence of a passing low‐pressure system. Attenuation was found to vary over more than 2 orders of magnitude and to be far higher than that observed in broken‐floe marginal ice zones. A clear linear relation between α and ice thickness was demonstrated, using ice thickness from a novel dynamic/thermodynamic model. A simple expression for α in terms of wave period and ice thickness was derived, for application in research and operational models. The variation of α was further investigated with a two‐layer viscous model, and a linear relation was found between eddy viscosity in the sub‐ice boundary layer and ice thickness.

Analysis of field measurements of ocean surface wave activity in the marginal ice zone, from campaigns in the Arctic and Antarctic and over a range of different ice conditions, shows the wave attenuation rate with respect to distance has a power law dependence on the frequency with order between two and four. With this backdrop, the attenuation‐frequency power law dependencies given by three dispersion relation models are obtained under the assumptions of weak attenuation, negligible deviation of the wave number from the open water wave number, and thin ice. It is found that two of the models (both implemented in WAVEWATCH III®), predict attenuation rates that are far more sensitive to frequency than indicated by the measurements. An alternative method is proposed to derive dispersion relation models, based on energy loss mechanisms. The method is used to generate example models that predict power law dependencies that are comparable with the field measurements.

No abstract