Stochastic dynamics of Arctic sea ice Part I: Additive noise

Abstract: We analyze the numerical solutions of a stochastic Arctic sea ice model with constant additive noise over a wide range of external heat-fluxes, $\Delta F_0$, which correspond to greenhouse gas forcing. The variability that the stochasticity provides to the deterministic steady state solutions (perennial and seasonal ice states) is illustrated by examining both the stochastic paths and probability density functions (PDFs). The principal stochastic moments (standard deviation, mean and skewness) are calculated and compared with those determined from a stochastic perturbation theory described previously by Moon and Wettlaufer (2013). We examine in detail the competing roles of the destabilizing sea ice-albedo-feedback and the stabilizing long-wave radiative loss contributions to the variability of the ice cover under increased greenhouse-gas forcing. In particular, the variability of the stochastic paths at the end of summer shows a clear maximum, which is due to the combination of the increasing importance of the albedo-feedback and its associated "memory effect" that accumulates from early spring to late summer. This is counterbalanced by the stabilization of the ice cover due to the longwave loss of energy from the ice surface, which is enhanced during winter, thereby focusing the stochastic paths and decreasing the variability in winter. The organizing principle for the interpretation of these effects is a generalized Ornstein-Uhlenbeck process with a time dependent potential.