Evaluating Parameterizations for Turbulent Fluxes over the Landfast Sea-Ice Surface in Prydz Bay, Antarctica

It is crucial to appropriately determine turbulent fluxes in numerical models. Using data collected in East Antarctica from 8 April to 26 November 2016, this study evaluates parameterization schemes for turbulent fluxes over the landfast sea-ice surface in five numerical models. The Community Noah Land Surface Model with Multi-Parameterizations Options (Noah_mp) best replicates the turbulent momentum flux, while the Beijing Climate System Model (BCC_CSM) produces the optimum sensible and latent heat fluxes. In particular, two critical issues of parameterization schemes, stability functions and roughness lengths, are investigated. Sensitivity tests indicate that roughness lengths play a decisive role in model performance. Based on the observed turbulent fluxes, roughness lengths over the landfast sea-ice surface are calculated. The results, which can provide a basis for setting up model parameters, reveal that the dynamic roughness length (z_0m) increases with the increase of frictional velocity (u_*) when u_* ≤ 0.4 m s⁻¹ and fluctuates around 10⁻³ m when u_* > 0.4 m s⁻¹; thermal roughness length (z_0t) is linearly related to the temperature gradient between air and sea-ice surface (ΔT) with a relation of lg(z_0t) = −0.29ΔT−3.86; and the mean water vapor roughness length (z_0q) in the specific humidity gradient (Δq) range of Δq ≤ −0.6 g kg⁻¹ is 10⁻⁶ m, 3.5 times smaller than that in the range of Δq ˃ −0.6 g kg⁻¹.