Laboratory experiments on diapycnal mixing in stratified fluids

Our turbulent lengthscale, velocity and diffusivity scalings are compared with data from other numerical, laboratory and field experiments. Comparison is made with reference to the turbulence intensity measure /νN. We showed that our turbulent lengthscale and velocity results are consistent with measurements from the experiments considered, and that the rms turbulent lengthscale Lt is independent of the rate of dissipation of turbulent kinetic energy when /νN > 300. A diffusivity modelled in terms of an advective buoyancy flux, b, is found to reproduce our direct measurements of Kρ in the experiments considered when /νN 2 > 300 . It is shown that modelling Kρ as 0.2 /N 2 is a poor parameterisation of the advective buoyancy flux model in all the experiments considered, and that at large /νN , this parameterisation can over-predict the true Kρ by two orders of magnitude. This overprediction is discussed in terms of a mixing efficiency and it is shown that in the experiments considered the mixing efficiency decreases rapidly with increasing /νN. Finally, the application of our diffusivity scaling to other geophysical flows is discussed, and it is shown that a necessary requirement for the use of this scaling is that Lt is independent of the rate of dissipation of turbulent kinetic energy.