Wave Attenuation by Sea Ice in the Marginal Ice Zone: Estimation and Rollover Behavior

Abstract

Ocean surface waves propagating into the marginal ice zone influence the extent and properties of sea ice through wave–ice interactions. However, the attenuation of wave energy by sea ice remains poorly represented in wave models used for climate projections and operational forecasting, and no universally applicable parametric formulation has yet been established. Here, we analyze four in situ datasets collected by wave buoys deployed around Svalbard between 2020 and 2023 to estimate spectral wave attenuation coefficients using buoy pairs. All datasets exhibit a power-law frequency dependence in the median attenuation coefficient profiles with exponents n = 1.32 ∼ 2.57. Deviations from this trend occur primarily at high frequencies, where the attenuation flattens or decreases—a phenomenon known as the “rollover”. Instrument noise can contribute to such spurious rollover and bias attenuation estimates; after applying noise control, our results suggest that neglecting local wind energy input may also contribute to it. The rollover tends to occur within the frequency range corresponding to spectral wave ages c / U 10 ≲ 1.2 (where c is the wave phase speed and U 10 is the wind speed at 10 m above the sea surface), and is more pronounced under relatively strong winds. Removing wind energy input prior to estimating attenuation is found to substantially reduce the rollover signal. These results suggest that, to mitigate spurious rollover in attenuation coefficient estimates, it is advisable to filter instrument noise and account for the wind input source term prior to attenuation estimation.