Summary: | It was found (Garzoli et al., 1992) that the general circulation model of Semtner and Chervin (1992) provides accurate descriptions of the Brazil-Malvinas and the Kuroshio/Oyashio confluence systems, except for the fact that the model prediction shows less variability than that present in observations. This paper investigates the problem of model variability by analyzing the mean and the eddy kinetic energy from the model and comparing the values with the Geosat altimeter observations for the South Atlantic Ocean and for the Kuroshio system. It is found that, while the model shows transient eddy activity in the areas that overlap the Geosat observations, the energy level of the model transient motions is considerably smaller following an arch along the bottom topography. The same was found from the comparisons made with values obtained from FGGE and surface drifters. It is suggested that the model is poorly resolving instabilities in the confluence front, and is not resolving other transients appearing in regions of marked topography.
|