A 1-D model study of Arctic sea-ice salinity

Abstract. We use a 1-D model to study how salinity evolves in Arctic sea ice. To do so, we first explore how sea-ice surface melt and flooding can be incorporated into the 1-D thermodynamic SAMSIM sea-ice model presented by Griewank and Notz (2013). We introduce flooding and a flushing parametrization which treats sea ice as a hydraulic network of horizontal and vertical fluxes. Forcing SAMSIM with 36 years of ERA-interim atmospheric reanalysis data, we obtain a modeled Arctic sea-ice salinity that agrees well with ice-core measurements. The simulations hence allow us to identify the main drivers of the observed mean salinity profile in Arctic sea ice. Our results show a 1.5–4 g kg−1 decrease of bulk salinity via gravity drainage after ice growth has ceased and before flushing sets in, which hinders approximating bulk salinity from ice thickness beyond the first growth season. In our simulations, salinity variability of first-year ice is mostly restricted to the top 20 cm. We find that ice thickness, thermal resistivity, freshwater column, and stored energy change by less than 5% on average when the full salinity parametrization is replaced with a prescribed salinity profile. We conclude that for earth system models the impact of fully parametrizing the Arctic temporal salinity evolution is too small to justify the increase in computational cost and model complexity.