A Model of Baroclinic Instability and Waves between the Ventilated Gyre and the Shadow Zone of the North Atlantic Ocean

Abstract

Distributions of eddy kinetic energy in the North Atlantic Ocean show that enhanced mesoscale activity exists in the Cape Verde Frontal Zone (the eastern part of the Central Water Boundary). This variability and its associated length and time scales are investigated with a three-dimensional numerical model that uses primitive equations in a hybrid (quasi-isopycnic) coordinate system in which the Coriolis parameter varies linearly. The model has a horizontal resolution of 15.625 km in 11 isopycnic layers. The domain comprises a 1000-km square centered at 20°N having idealized bottom topography and zonally periodic boundary conditions. A zonal, geostrophically balanced jet representing an undisturbed Canary Current is used for the initial conditions. Potential vorticity is used to distinguish between water masses that are reconstructed from a hydrographic section. The growth of meanders is stimulated by an ageostrophic perturbation field consisting of white noise. Integration is carried out for 200 model days. Baroclinic instability leads to a growth of meanders and the subsequent shedding of eddies. The most unstable wave has a length of 200 km, attains an amplitude of about 100 km after 80 model days, and sheds its first eddy around model day 140. Eddies have typical horizontal length scales of 100 km. Instabilities associated with the westward jet are confined to a zonal strip about 300 km wide and occur over time scales of 100?300 days in the uppermost layers, 100?125 days in the intermediate layers, and between 50 days and infinity in the layers close to the bottom. Baroclinically unstable waves, Rossby waves, and topographic Rossby waves contribute to the variability, but the latter two can be observed only in the layers near the bottom owing to their small amplitude relative to the growing unstable waves elsewhere. There is good agreement between model predictions and observations with respect to the spatial scales of variability. In comparing the temporal scales predicted by the model with spectra from mooring velocity records, we find consistency within the limits imposed by statistical constraints. Phase spectra from these records indicate that the observed eddy activity in the CVFZ is due to baroclinic instability.