This study is an evaluation of the polar atmospheric boundary layer in an ensemble of Arctic regional weather and climate models using observations from the Multidisciplinary drifting Observatory for the Study of Arctic Climate (MOSAiC). This multi-model analysis takes advantage of the unique year-round observations available from MOSAiC and highlights a novel artificial learning-based approach for evaluating model simulations of boundary layer processes. The project is a part of ongoing model intercomparison efforts by the Arctic CORDEX (Coordinated Regional Downscaling Experiment) and MOSAiC research communities. An initial evaluation has been completed for the Coupled Arctic Forecast System (CAFS). Additional regional atmosphere-only and coupled model simulations, contributed by the international Arctic CORDEX community, are also evaluated. The collection of models includes multiple versions of the Weather Research and Forecasting (WRF) model and the Regional Arctic System Model (RASM). Regional output from global models, including the Department of Energy’s Energy Exascale Earth System Model (E3SM), will also be included in this multi-model evaluation.

Static stability within and just above the boundary layer was evaluated based on radiosonde profiles of virtual potential temperature. An artificial neural network data analysis approach known as self-organizing maps (SOMs) was used to objectively identify the range of virtual potential temperature profiles. The virtual potential temperature profiles, up to 1 km, are from ~1400 MOSAiC radiosonde observations and vertical profiles from the multi-model ensemble corresponding to the location of MOSAiC. Initial results indicate that CAFS reproduced the full range of observed stability profiles, but not necessarily with the correct frequency or at the correct time. Based on the SOM analysis, boundary layer stability regimes were defined by near surface stability and stability just above the boundary layer for both the radiosonde and model profiles. The results show that CAFS underrepresents the frequency of strong stability, particularly when the strong stability occurs just above the boundary layer. Radiation and wind observations were used to assess impacts of varying surface energy budget and wind shear on boundary layer stability and determine if these boundary layer forcings were accurately depicted in the models. Downwelling longwave radiation and 10 m wind speed corresponding to each stability regime were found to have too great a magnitude in CAFS compared to the observations, with the largest differences for the strongest stability regimes.