Ocean Waves Quietly Push Sea Ice Around, and New Buoy Data Finally Show How Much

The edge of the polar sea is one of the most violent battlegrounds in the climate system, and for decades scientists have argued about how hard the ocean waves actually hit the ice. A new observational study published in Ocean Dynamics now delivers some of the most direct field-based estimates to date of both how quickly waves decay as they travel into sea ice and how much momentum those waves hand over to the ice cover. Drawing on two independent networks of drifting buoys, one deployed in the Barents Sea in winter 2021 with six instruments and another comprising 65 buoys around Svalbard between 2022 and 2023, the study combines the buoy records with ERA5 reanalysis forcing and daily sea ice concentrations from the AMSR2 satellite sensor to build a quantitative picture of the marginal ice zone as a coupled wave-ice system.

The marginal ice zone, the sprawling and fractured ribbon of floes that separates the open ocean from the consolidated pack, is where wave-ice interactions matter most. Waves arriving from distant storms can flex and fracture floes, change their size distribution, drive currents along the ice edge, and alter the drag between the atmosphere, the ice, and the ocean. Yet despite decades of theoretical modeling, direct observational estimates of the magnitude of wave-induced momentum transfer to sea ice have remained scarce. The new work addresses this gap by treating the buoy network not merely as a wave detector but as a distributed measurement of the mechanical forcing that the ice actually experiences during storms.

One of the study’s central technical contributions is a new estimator of the attenuation power law, the relationship that describes how wave energy or amplitude decreases with distance traveled through ice. Attenuation in the marginal ice zone is commonly approximated by a power law in frequency, in which the attenuation coefficient grows with wave frequency raised to some exponent. Traditional methods for fitting this law require analysts to decide in advance whether wave energy is increasing or decreasing along the propagation path, an assumption that can bias results when buoys drift, storms evolve, or the ice itself changes. The new estimator removes the need for any sign selection and was validated against synthetic spectra before being applied to the field data, a precaution designed to guard against the spurious rollover artifacts that have plagued earlier field campaigns.

The resulting exponents are striking in their regime dependence. In winter pack ice, the study finds a frequency exponent of 2.55, with a 95 percent confidence interval spanning 2.28 to 2.82. In spring pack ice the estimate is 2.37, with a wider interval of 1.58 to 3.46 reflecting the smaller or noisier sample. But in a single autumn campaign over new, thin ice, the exponent jumps to 5.29, with a confidence interval of 4.84 to 5.60. That dramatic difference means there is no single universal attenuation law for the marginal ice zone. Thin new ice, which offers little mechanical resistance, attenuates high-frequency wave energy far more steeply with distance than the thicker, more coherent pack of winter and spring. For wave forecasters and coupled climate modelers, the message is that attenuation parameterizations must be conditioned on ice regime, not treated as constants.

The second major thread of the study concerns momentum, and it is here that the findings are perhaps most consequential. In essence, the author asked how much of the mechanical push that sets the ice in motion comes not from the wind directly but from the waves. Within 25 kilometers of the ice edge during storm conditions, the ratio of wave-induced momentum input to wind momentum input reaches as high as 0.32 at the 90th percentile, with a range of 0.02 to 0.46. In other words, at the outer margin of the ice, waves can contribute on the order of a third of the momentum that the wind alone would deliver, a share large enough to matter for ice drift, floe fracture, and the exchange of heat and salt across the ice-ocean interface.

Crucially, this wave contribution does not simply persist deep into the ice cover. The ratio decays inward from the ice edge with an exponential scale of 12 kilometers in the Barents Sea network, with a confidence interval of 9 to 22 kilometers, and 10 kilometers around Svalbard, with an interval of 8 to 12 kilometers. The author tested the robustness of these decay scales by allowing the drag coefficient to depend on both ice concentration and wind speed, and the results held. What makes the decay rate surprising is that it is faster than what the measured spectral attenuation of wave energy would predict. The waves lose energy at one rate, but the momentum they deliver to the ice falls off more quickly, a discrepancy that hints at additional physics, including directional spreading, scattering by individual floes, and the possibility that energy is dissipated locally within the ice rather than transmitted as organized motion.

That discrepancy becomes even more revealing when the direction of the forcing is considered. Where the wave stress and the wind stress are misaligned by more than 45 degrees, the residual of a wind-only model of ice motion grows systematically with the magnitude of the wave stress, with a Spearman rank correlation of 0.50 and a confidence interval of 0.14 to 0.70. This is a statistical fingerprint: when waves and winds point in different directions, the ice goes where a wind-only model says it should not, and the size of the error tracks the wave forcing. The implication is sobering for operational forecasting. Sea ice drift models that neglect wave momentum transfer will be most wrong precisely in the mixed, confused seas near the ice edge, where the misalignment between swell and local wind is largest and where shipping, fishing, and offshore operations are most active.

The study also tackles the dramatic process thatCaptures headlines: ice breakup. By scanning the buoy records for abrupt jumps in spectral variance, the author built a catalogue of 85 candidate breakup events. Of these, 43 occurred in ice with concentrations of at least 50 percent, and these in-ice events showed wave strains in the upper tail of the background strain distribution, indicating that the ice was being flexed harder than usual when the events occurred. Within three days of these events, the daily AMSR2 satellite concentration at the buoy locations dropped by a median of 9 percentage points. Nine of the ten strongest in-ice events show the same pattern in case-by-case validation: an energy jump while the ice concentration is still high, followed by a measurable concentration decline. While a concentration drop of a few points over three days cannot be attributed to waves with absolute certainty given the confounding influences of melt, divergence, and advection, the statistical consistency of the pattern across dozens of events strengthens the case that wave-induced flexure is a genuine driver of ice edge fragmentation.

The methodological transparency of the work deserves attention in its own right. Every sample count can be audited from candidate matches to final estimates through a published attrition budget, the analysis was conducted in MATLAB with a purpose-written ice fluid-structure interaction package, random seeds were fixed in the scripts, and bootstrap resampling with sizes between 500 and 2000 underpins all the confidence intervals quoted. The observational buoy datasets, ERA5 products, and AMSR2 concentrations are openly available from public repositories, and the derived data and processing scripts are documented and can be made available on reasonable request. In an era when wave-ice models are being coupled into operational sea ice forecasting systems and Earth system models, studies of this kind provide the empirical anchor that the models need. As the Arctic warms and the marginal ice zone widens in summer and shifts poleward in winter, the ocean’s waves are likely to reach further into the ice cover than at any time in the satellite record. Quantifying exactly how much push they deliver, and how quickly that push fades with distance into the pack, is a step toward predicting the future of the ice edge itself.

Subject of Research: Wave attenuation and wave-induced momentum transfer to sea ice in the Arctic marginal ice zone

Article Title: An observational study of wave attenuation and wave-induced momentum transfer to sea ice in the marginal ice zone

Article References: Yoladi, M. (2026). An observational study of wave attenuation and wave-induced momentum transfer to sea ice in the marginal ice zone. Ocean Dynamics, 76(10), Article 103. https://doi.org/10.1007/s10236-026-01855-y

Image Credits: AI Generated

DOI: 10.1007/s10236-026-01855-y

Keywords: marginal ice zone, wave attenuation, momentum transfer, sea ice, ice breakup, drifting buoys, Barents Sea, Svalbard, AMSR2, ERA5, Arctic, ocean waves