On the role of laboratory experiments in the development of numerical ocean models; a view gained from the study of submarine canyons

Among the difficulties of developing prognostication models of oceanic flows generally, and coastal flows in particular, is the absence of data sets with sufficient resolution in space and time to serve as tests for these models. It is submitted that carefully designed laboratory experiments closely coupled with the development of numerical models can aid in the development of more realistic versions of the latter.

Studies of the time-dependant and the time-mean flow of an oscillatory along-shelf current past a generic submarine canyon incised in an otherwise continuous shelf-slope system is investigated. The fluid is linearly stratified and the system is rotating. The principal laboratory experiments use particle imaging velocimetry (PIV) to monitor the velocity fields in horizontal planes at various depths. The numerical experiments employ the spectral element ocean model (SEOM) developed at Rutgers University.

Laboratory experiments for laminar flows are found to be in good qualitative agreement with the numerical models, but their quantitative agreement is found to be less than desirable in that the differences substantially exceed the differences suggested by the demonstrated errors in the laboratory runs. These difficulties are attributed to the fact that the numerical model does not satisfactorily represent the bottom Ekman layers. Results for the case that the bottom boundary layer is turbulent are presented but owing to resolution problems numerical results have not been obtained in these cases.

The study shows that net upwelling in the upper reaches of the canyon is a ubiquitous feature of canyons with the source of fluid being the deep water beyond the mouth of the canton. Conservation of mass is satisfied by this deep water flow passing through the mouth below the level of the shelf and then being transported across both flanks of the shelf above the shelf break with a net outward flow across the canyon mouth. Contrasts are made between an elevated coastal feature and a canyon. The physics of the processes are discussed with simplified conceptual models.