Seasonal variability of Red Sea mixed layer depth: the influence of atmospheric buoyancy and momentum forcing

Abstract

The seasonal and spatial evolution of the mixed layer (ML) in the Red Sea (RS) and the influence of atmospheric buoyancy and momentum forcing are analyzed for the 2001–2015 period using a high-resolution (1/100°, 50 vertical layers) ocean circulation model. The simulation reveals a strong spatiotemporal variability reflecting the complex patterns associated with the air–sea buoyancy flux and wind forcing, as well as the significant impact of the basin’s general and mesoscale circulation. During the spring and summer months, buoyancy forcing intensifies stratification, resulting in a generally shallow ML throughout the basin. Nevertheless, the results reveal local maxima associated with the influence of mesoscale circulation and regular wind induced mixing. Under the influence of surface buoyancy loss, the process of deepening of the ML commences in early September, reaching its maximum depth in January and February. The northern Gulf of Aqaba and the western parts of the northern RS, exhibit the deepest ML, with a gradual shoaling toward the south, primarily due to the surface advection of relatively fresh water that enters the basin from the Gulf of Aden. The mixed layer depth (MLD) variability is primarily driven by atmospheric buoyancy forcing, especially its heat flux component. Although evaporative fluxes dominate the annually averaged surface buoyancy forcing, they exhibit weak seasonal and spatial variability. Wind induced mixing exerts a significant impact on the MLD only locally, especially during summer. Of particular importance are strong winds channeled by topography, such as those in the vicinity of the Strait of Bab-Al-Mandeb and the straits connecting the two gulfs in the north, as well as lateral jets venting through mountain gaps, such as the Tokar Jet in the central RS. The analysis highlights the complex patterns of air-sea interactions, thermohaline circulation, and mesoscale activity, all of them strongly imprinted on the MLD distribution.