The stable atmospheric boundary layer over snow-covered sea ice: Model evaluation with fine-scale ISOBAR18 observations

Abstract

A realistic representation of the stable atmospheric boundary layer in numerical weather prediction (NWP) and climate models is still a challenge. We study the evolution of a stable boundary layer over snow-covered sea ice in Bothnian Bay during wintertime in 2018. We perform high-resolution model experiments with the Weather Research and Forecasting model in its single-column model configuration and its default mesoscale configuration to assess which physical processes are essential to predict near-surface variables correctly. We evaluate our model runs against the unique observational dataset collected during ISOBAR18, which combines novel, upper-air measurements by an uncrewed aircraft system with wind lidar, sodar, and conventional meteorological mast data. By analysing surface fluxes in the single-column model, we demonstrate how the atmospheric cooling at the ground can be modelled more realistically than in the mesoscale set-up. We show that surface albedo and sea-ice thickness are essential for the surface energy balance in the model, and we demonstrate how the surface fluxes in the mesoscale downscaling with default settings are subject to strong biases. We also show that the ERA5 reanalysis is not capable of representing the observed surface meteorology in the stable atmospheric boundary layer. Our study illustrates the importance of surface albedo and sea-ice thickness for NWP models. Though a seasonal snow albedo is already in use in many NWP settings, the routine inclusion of sea-ice thickness, in particular, would be a great step forward for weather forecasts and regional climate simulations.