Sensitivity of iceberg drift and deterioration simulations to input data from different ocean, sea ice and atmosphere models in the Barents Sea

Abstract. Iceberg observations in the Barents Sea are scarce. Numerical simulations of iceberg drift and deterioration help fill this gap. The quality of these simulation results depends, among other factors, on the accuracy of the environmental data (e.g., wind, waves, currents, salinity, temperature), often derived from ocean, sea ice and atmosphere models. In this study, we conduct a numerical experiment simulating the drift and deterioration of a large number of synthetic icebergs. We force the iceberg model with two atmospheric reanalyses (ERA5, CARRA) and two ocean and sea ice models (Topaz, Barents-2.5) in the Barents Sea for the years 2010–2014 and 2020–2021. The differences in iceberg model output are statistically quantified, illustrated using an exemplary trajectory, and explained based on variations in environmental input. We conclude that simulation results of iceberg drift and deterioration are highly sensitive to the choice of the environmental input, depending on the simulation goal, time frame, area of interest and input characteristics. Iceberg simulations using input from Barents-2.5 yielded a distinct regional distribution of iceberg density, 8 d longer drift duration, and −6.2 × 104 kg lower deterioration trend. These differences are primarily attributed to lower sea surface temperature (−0.41 °C), higher ice concentration (4 %), larger exposure to sea ice (23 %), larger water speeds (0.05 m s−1) and the representation of tides and topographically-steered currents in Barents-2.5, compared to Topaz. Atmospheric input has little impact on most iceberg characteristics. However, the iceberg pathways and their southern extent remain largely insensitive to variations in environmental inputs.