1University of British Columbia, Vancouver, British Columbia, Canada
2Institute of Ocean Sciences, Sidney, British Columbia, Canada
Treatment of boundary conditions is a most important issue for every fluid dynamics model. Global ocean models (GOM) usually rely on simple ``sidewall'' boundary conditions for passive scalars (eg. chemical tracers). The purpose of this study is to better understand mechanisms of passive scalar transport between the coastal zone and the open ocean in order to improve GOM results.
The coastal zone topography is characterized by two features: the shallow, gently sloped continental shelf and a steep slope which connects to the deep ocean. Small scale variability along the shore has been included on the slope. Strong thermodynamic and biotic conversions tend to occur over the shallow shelf region, with products of these conversions exchanged with the deep ocean (production in part depending upon deep ocean exchange). We focus on possible exchange mechanisms.
We use a finite difference, two-layer, primitive equation model with periodic boundaries in the along-shore direction. Special attention has been given to the coast boundary condition. The transport of the scalar is computed with a second-order, Crowley-based advection \forcenl scheme. A time-splitting approach is used and the Flux Corrected Transport algorithm has been extended to treat transport in layers with dynamically evolving thickness while assuring monotonicity of advection.
An upwelling event (sliding of the uppermost layer of the ocean offshore and deep cold water rising near the coast) has been considered. The model successfully reproduces the main features of an upwelling circulation. The surface elevation appears slightly tilted in the cross-shelf direction. The interface between the two layers clearly shows the deep water rising. Meanders develop at the shelf edge. The mean cross-shelf transport is well approximated by the Ekman transport value $\tau / \rho f$. The cross-shelf advection of scalar is strongly correlated with the along-shore anomalies of the bottom topography.
Future work will use a zero-mean stochastic wind forcing to excite a statistically stationary eddy field. We will estimate a cross-shelf diffusion coefficient of the passive scalar, with particular attention to relations between cross-shelf transport and topography.