Modeled Stokes drift in the marginal ice zones of the Arctic Ocean

Abstract

This study modeled Stokes drift in the marginal ice zone (MIZ) of the Arctic Ocean using WAVEWATCH III. Applying two viscoelastic and one empirical frequency-dependent wave–ice models, modeled wave parameters and spectra were compared with field observations from the Beaufort–Chukchi Sea. The three wave–ice parameterizations showed similar ability in reproducing the surface Stokes drift estimated using buoy measurements. Comparison with buoy drift showed the estimation of total surface drift was improved by considering Stokes drift during significant wave events. Based on 5-year (2015–2019) hindcasted directional spectra of the summer-autumn Arctic, we investigated monthly mean surface Stokes drift (1–10 cm/s), e-folding depth (1–14 m), and vertically integrated transport (< 0.23 m²/s) in the MIZ. The results showed that Stokes drift in the MIZ increased in strength from July to October. When bulk wave parameters were adopted to estimate Stokes drift fields, the surface Stokes drift was underestimated by about 34–50%, while the Stokes e-folding depth was overestimated by approximately 1.4–5.0 times, increasing from the interior to the edge of the ice cover. We compared the modeled surface Stokes drift with the Eulerian mean flow from reanalyzed data. It showed the mean surface Stokes drift is typically approximately 30% of the Eulerian current over large parts of the Arctic Ocean MIZ with mean ice concentration < 60%, and is the same order of magnitude as (or even larger than) some areas of the Chukchi, East Siberian, and Laptev Seas. It implies Stokes drift should be considered to improve modeling of the dynamic processes of sea ice, especially ice floe drift.