Over the last few decades, the polar regions have become an increasing focus for the weather and climate science community, and there is clear evidence that the interactions between waves and sea ice are important across various time scales.
In the present and past ECMWF operational forecasting systems, wave-sea ice interaction has not been represented. In the ECMWF Ocean Wave Model (ecWAM), waves are simulated only up to 30% sea ice cover (SIC). Beyond this limit, they are instantly attenuated (i.e. set to 0), and most model output variables are masked, thereby providing no wave information in areas with higher levels of ice cover.
In Cycle 50r1, scheduled for operational implementation in early 2026, waves have been introduced into the sea ice. First, the wave model is allowed to operate under all sea ice conditions. To describe the attenuation rate, a state-of-the-art model is used that accounts for ice thickness and wave frequency. With this implementation, consistent wave forecasts across the polar regions can be provided.
With recent advances in hardware and software, wave buoys have reduced in cost. This has enabled the Norwegian Meteorological Institute (Meteorologisk Institutt), which has largely been spearheading this work, to deploy many more buoys in their campaigns. Because each buoy provides only a single observation point, and sea ice is inhomogeneous, many buoys are needed to build a representative picture. Thanks to collaborations with colleagues in Norway, the Cycle 50r1 model can now be tested and verified against such campaign data.
Taking the Svalbard Marginal Ice Zone (MIZ) 2024 (SvalMIZ-24) campaign as an example, to verify the Cycle 50r1 model, a wave model hindcast was run over the campaign period, using atmospheric information from ERA5, and ocean and sea ice information from ORAS6. The model predictions of wave height (HS) were then compared with those measured by the buoys.
The top two panels of the figure show model HS for Cycle 49r1 and Cycle 50r1 globally during the SvalMIZ-24 campaign. As waves enter the polar regions in the Cycle 50r1 model, there is a sharp but smooth gradient as they are attenuated by the sea ice. For Cycle 49r1, no wave information is provided for SIC>30%. Focusing on the SvalMIZ-24 campaign, the bottom left panel shows model and observed HS for the campaign's most intense storm. Within the MIZ (between the 15% and 85% SIC contours, marked by the solid and dashed contour lines, respectively), the buoys measure waves of up to HS =5.37 m. The Cycle 50r1 model also shows large waves in this region but is low biased (HS =3.84 m at this location). In the pack ice north of Svalbard however, (>85% SIC), the Cycle 50r1 model and the buoys agree: the waves have been fully attenuated (HS ≈0 m).
The bottom right panel shows a statistical verification of the model in the MIZ zone for the campaign: for the bulk of the measurements (HS <2 m), the model is unbiased with respect to the observations, albeit random errors remain significant. For larger waves (HS >2 m), the model is low biased compared to the observations. Although there is room for improvement, the Cycle 50r1 model can capture the basic evolution of waves in sea ice. This represents a significant qualitative advancement compared to Cycle 49r1, which is unable to provide wave information in these regions.
The Cycle 50r1 implementation represents a one-way coupling between waves and sea ice (attenuation of waves by sea ice). This implementation, along with the development of the relevant verification infrastructure, lays the foundations for a larger body of work that is currently underway: the representation of two-way coupled interactions between the waves and sea ice.
Additionally, in Cycle 50r1, refraction by ocean surface currents will be activated, using ocean information from the coupled ocean model NEMO4. Refraction can alter the direction of wave propagation, resulting in wave focusing, meaning waves are larger in some areas and smaller in others, as seen most clearly in the tropics through the increased small scale variability in the contour lines, shown in the top two panels of the figure, due to the strong currents here.
Help shape our wave products
Are you interested in helping shape wave products at ECMWF? If so, get in touch through this QR code or at https://tinyurl.com/ecmwf-wave-survey
