Abstract

Traditionally, the mechanism driving the seasonal restratification of the Southern Ocean mixed layer (ML) is thought to be the onset of springtime warming. Recent developments in numerical modelling and North Atlantic observations have shown that submesoscale ML eddies (MLE) can drive a restratifying flux to shoal the deep winter ML prior to solar heating at high latitudes. The impact of submesoscale processes on the intra-seasonal variability of the Subantarctic ML is still relatively unknown. We compare five months of glider data in the Subantarctic Zone to simulations of a 1-D mixing model to show that the magnitude of restratification of the ML cannot be explained by heat, freshwater and momentum fluxes alone. During early spring, we estimate that periodic increases in the vertical buoyancy flux by MLEs caused small increases in stratification, despite predominantly down-front winds that promote the destruction of stratification. The timing of seasonal restratification was consistent between 1-D model estimates and the observations. However, during up-front winds, the strength of springtime stratification increased over two-fold compared to the 1-D model, with a rapid shoaling of the MLD from >200 m to <100 m within a few days. The ML stratification is further modified under a negative Ekman buoyancy flux during down-front winds, resulting in the destruction of ML stratification and deepening the MLD. These results propose the importance of submesoscale buoyancy fluxes enhancing seasonal restratification and mixing of the Subantarctic ML.

Link to Full Article
Marcel_F1

a) NCEP wind stress (N m-2) co-located to the time and position of glider profiles and used for the PWP model run. b) MERRA 6-hourly net heat flux (QNET ) at the ocean surface (red (blue) indicates ocean heating (cooling)) indicates a strong diurnal structure with a net positive heat flux into the ML. c) Time series of stratification (N2) averaged for the upper 100 m for SG573 (green line) and the PWP model (magenta line). Shaded areas are as in Figure 3.