Oceanic sill-overflow systems: Investigations and simulation with the Poseidon ocean general circulation model

Marginal seas transport dense water into the oceanic system by deep gravity currents through sill-overflow systems. This dense water is essential in establishing background characteristics and the density driven component of the meridional overturing of the ocean. This thesis investigates the controlling dynamics and several modeling issues related to oceanic deep gravity currents. The objectives and questions of this study center on the development of a climate scale model that adequately simulates sill-overflow deep gravity currents. These questions include the sensitivity of the model to resolution and various turbulence parameterizations when modeling these systems. A hybrid generalized vertical coordinate model is adapted to address these questions. This study uses theoretical model domains based on the Dynamics of Mixing and Entrainment group (DOME).; Current, climate scale models typically have horizontal resolutions ranging from 1/2 to 1 degree resolution. The bulk of the vertical resolution in layered models is commonly placed in density or pressure space centered around the equatorial upper ocean, well away from the regions containing the deep gravity currents investigated in this study. Models solutions from this study show large dependence on both horizontal and vertical resolution. When horizontal resolution is decreased from 5 to 20 km, entrainment of ambient water into the gravity current decreases by approximately a factor of 3. The deep gravity currents in the lower (20 km) resolution experiments are deeper, denser and have larger along-slope velocities than the higher resolution experiments (5 km). This difference is explained by the homogenization of the gravity current in lower resolution simulations. The 5 km simulations show higher magnitudes and spatial variability of momentum. This homogenization results in underestimation of the applied bottom stress in the model. With a lower bottom stress, the current does not have sufficient shearing to meet: the mixing criteria of the Richardson number dependent mixing scheme used in the model. Altering the mixing criteria does little to improve solution. This approach to solving the mixing problem will likely corrupt the solution of the upper ocean.; By adjusting the drag coefficient to compensate for the underestimation of bottom stress, lower resolution experiments showed much better agreement in entraininent, along-slope velocity and gravity current pathway with the higher resolution experiments. This approach has the advantage of only having direct impacts on regions in proximity to topography. Model simulations, using realistic topography and forcing, of the Gulf of Cadiz show much improved agreement with observations when this adjustment is applied.