A Balanced Atmospheric Ensemble Forcing for Sea Ice Modeling in Southern Ocean

Abstract

Abstract To deal with the great challenges in simulating the highly non‐linear physics of Antarctic sea ice, a multivariate balanced atmospheric ensemble forcing is developed based on the high‐resolution component of ERA5, which considers the relationship between different variables and adjacent times. To validate the performance of this new forcing, experiments were conducted from 1 January 2016 to 28 February 2017. Compared to simulations forced with the ensemble component of ERA5, a more reasonable ensemble of simulations is produced by the atmospheric ensemble forcing developed in this study, which suppresses the sea‐ice concentration (SIC) simulation errors and produces a better estimation of SIC simulation uncertainties. Further sea‐ice thickness budget analysis reveals that this impact of atmospheric ensemble forcing on sea ice simulation is due to a modulation of atmosphere‐ocean and sea ice‐ocean thermodynamic processes. These results lay the foundation for further improvements in Antarctic sea ice data assimilation and probabilistic prediction.