Simulating and forecasting southern Africa's ocean

Why model the ocean?

Numerical ocean models allow us to 'sample' a model ocean at a much higher space and time frequency than is possible by either satellite or in situ monitoring programs in the real ocean. With the advent of satellite observations in the 1960s, and the continual advancement of satellite technology since then, great insight into large-scale features and dynamics of the global ocean has been achieved. Unfortunately, most satellite sensors do not 'see' deeper than the surface ocean. In situ monitoring is a most important aspect of gaining ocean information, however it is also extremely expensive. Given the remarkable improvement in computing power and efficiency over the last couple of decades, high resolution ocean models can be run quite affordably and become a critical source of information.

Above is an animation of salinity and currents at a depth of 60m simulated in ROMS for the Benguela upwelling system.

Ocean models can be complex or simple. Complex models generally aim to be as realistic as possible and therefore require detailed measurements to drive them, for example: accurate measurements of the sea floor variations, wind field and other atmospheric variables, river discharge rates, tides etc. A large volume of ‘data’ (computer output) is produced by complex models, which can be helpful in contextualizing spatial and temporal variations of oceanographic features. These complex models can also be simplified to simulate the fluid dynamics in idealized scenarios that are specifically designed to understand theoretical processes that govern observed oceanographic features. The simplicity of such models allows for robust conclusions and is thus a powerful method of understanding ocean dynamics. Numerical models, both complex and simple, are invaluable tools for a greater understanding of ocean processes.