Due to the difficulty in carrying out ocean observations, numerical modeling is an invaluable tool for studying the ocean to changes in the background forcing. Mesoscale eddies are known to impact the mean state of the ocean, such as affecting material distribution, heat and momentum transport via the global circulation, though the high cost of eddy-resolving models makes coarse resolution models which require parameterization of such processes necessary. However, existing theories of parameterization generally do not consider the influence of the bottom bathymetry, and research have shown that mesoscale eddies over continental slopes possess dynamics that is different from those in the open ocean, with implications on the exchange between continental shelves and open ocean, affecting t...[ Read more ]

Due to the difficulty in carrying out ocean observations, numerical modeling is an invaluable tool for studying the ocean to changes in the background forcing. Mesoscale eddies are known to impact the mean state of the ocean, such as affecting material distribution, heat and momentum transport via the global circulation, though the high cost of eddy-resolving models makes coarse resolution models which require parameterization of such processes necessary. However, existing theories of parameterization generally do not consider the influence of the bottom bathymetry, and research have shown that mesoscale eddies over continental slopes possess dynamics that is different from those in the open ocean, with implications on the exchange between continental shelves and open ocean, affecting the path and rate of transportation of nutrients and biota. Here, we consider two different parameterization schemes, the GM and the GEOMETRIC scheme, in non-eddy-resolving ocean circulation model to explore the consequences of parameterization in models in the presence of bottom continental slopes. It is found that a naive application of the parameterization schemes in the presence of slopes performs even worse than the one without parameterization, attributed to the importance of how to represent the sharp transition of the eddy-induced velocity coefficient κ_gm around the slope regions. Two types of tapering functions acting on the eddy-induced velocity coefficient κ_gm are considered, one based on the bathymetry depth, and the other inspired from recent diagnostic works and based on the Burger number. Implications on implied numerical modeling and projected biogeochemical responses arising from the present work are discussed.