Abstract
Direct Numerical Simulation (DNS) has been employed with success in a variety of oceanographic applications, particularly for investigating the internal dynamics of Kelvin Helmholz (KH) billows. These studies have yielded an impressive array of information regarding interactions (including pairing) of adjacent KH billows, the generation of secondary instabilities, and the evolution of local mixing efficiencies, among other topics. An alternative focus of DNS models is to mimic naturally occurring turbulence at laboratory or weakly forced geophysical scales, in order to broaden the parameter space of available measurements to enable the development of enhanced closure techniques. To do this, it is necessary to demonstrate the mean values of turbulent quantities (e.g., production, buoyancy flux, and dissipation) are consistent with measurements of similarly parameterized flows. Modeled turbulence evolution in a stratified shear flow is determined not only by the strengths of the background shear and stratification, but also by the initial and boundary conditions used to define the, and focus our analysis on two parameters: the intensity of background "noise" which is used to stimulate the turbulent field, and the length of the domain in the streamwise direction. Variation of the domain length is shown to have profound implications not only for the number of billows formed, but also for the nature of their interactions, and for the intensity of the resulting mean turbulence. Results indicate that the ratio of domain length to a theoretical KH wavelength, , where h is the thickness of the shear layer and RiB represents the bulk Richardson number, is a critical parameter in evaluating the effect of variable domain lengths. Results of simulations with varying initial noise amplitudes are used to identify the impact of initial noise on final turbulence parameters and identify ideal initialization noise levels.