Boundary Conditions for a Nested Model of the Oregon Coastal Transition Zone

Flows in coastal oceans are influenced by large scale circulation and also depend highly on small scale topography and nonlinear interactions. It is not generally feasible to run a general circulation model with sufficient resolution for the small scales, nor is it desirable to run a regional model which ignores the time-varying large scale circulation. A pragmatic approach to coastal modeling is a one-way nesting of a high resolution regional model in a relatively low resolution global model, a procedure sometimes called downscaling. The large scale features of the global model are communicated to the regional nested model via boundary conditions, and, to a lesser extent, initial conditions. A serious complication, however, is that disturbances generated within the nested model must be permitted to leave the domain, and such open boundary conditions are mathematically ill posed, so a variety of ad hoc strategies have been devised. How can a modeler identify a good choice of boundary conditions for a nested model? Big brother experiments are a methodology for studying the efficacy of the nesting strategy. A high resolution model (the big brother) is run over a large domain (not necessarily global) and an identically formulated nested model (the little brother) is run over a smaller domain with boundary conditions obtained from the big brother. A series of such experiments was conducted in a domain representing the Oregon Coastal Transition Zone. Even with nearly ideal boundary conditions, the big brother and little brother models diverge due to sensitive instabilities of the coastal flows. To evaluate more realistic nesting scenarios, the boundary and initial conditions extracted from the big brother were smoothed or interpolated and the effects of these degradations gauged by comparing the little brother solution with the "true" big brother solution. Furthermore, the effects of using different model forumations, parameterizations, topography, and forcing were explored systematically. Finally, the results of these experiments were used to evaluate boundary conditions for nesting a high resolution, s-coordinate model (ROMS) in a lower resolution, hybrid (σ-z) coordinate operational global model (NCOM).