Evaluation of atmospheric boundary layer schemes in the Arctic during clear-sky conditions

A single-column model is used to evaluate the performance of five atmospheric boundary layer (ABL) schemes used in large-scale atmospheric models over the Arctic Ocean pack ice during clear-sky conditions. Four of the schemes use first-order closure and a combination of local and nonlocal parameterizations in their prediction of eddy diffusivity coefficients. A fifth scheme uses level 2.5 closure to parameterize eddy diffusivity coefficients as a function of turbulence kinetic energy. Two short duration (four to five-day) case studies taken from the Surface Heat Budget of the Arctic Ocean field experiment are used for their evaluation. Simulations of these periods reveal inconsistencies between the observed vertical transfer of heat during these periods and its parameterization within these schemes. In general, the schemes mixed too strongly in parts of the domain and too weakly in others, rarely prescribing mixing of intermediate strength, as was commonly observed. Additionally, the schemes showed little vertical or temporal variability in mixing intensity within either strongly or weakly-mixed regimes, even under significant changes in forcing. The schemes also had trouble correctly entraining the heat produced at the top of the ABL by the interaction of subsiding motions with the strong overlying temperature inversion. The nonlocal schemes generally performed better than the local schemes under stable conditions, prescribing mixing of greater vertical and temporal variability, and showing greater sensitivity to changes in the forcing. The absence of parameterization of the effects of breaking waves on mixing in the lower atmosphere is identified as a crucial omission in all of the schemes. Incorporation of this source of mixing would be hypothesized to increase the mean vertical transfer of heat within and above the ABL, improving the representation of entrainment, and leading to a more dynamic interaction between stabilizing and destabilizing influences throughout the lower atmosphere and ABL. The importance of vertical motion and adequate representation of the overlying temperature inversion to lower atmospheric dynamics and thermodynamics is also demonstrated.