On the use of laboratory observations to validate numerical models

In order to develop data sets which can be used as benchmarks for coastal circulation numerical models, laboratory experiments are being conducted in a cylindrical tank in which a continuous continental shelf model, interrupted only by a single smooth canyon, is placed along the periphery of the test cell. Prior to experimentation, the tank is filled with a linearly stratified fluid and the tank is then slowly brought up to solid body rotation with Coriolis parameter f. To initiate the experiments, the turntable rotation rate is then either (i) impulsively changed by ±∆ω or (ίi) modulated sinusoidally about the background rotation rate f∕2. The objectives of the experiments are to (i) observe and better understand the motion field in the vicinity of a submarine canyon and (ii) provide benchmarks for the development of numerical coastal circulation models. The present communication reports on some preliminary laboratory and numerical experiments directed toward these goals.