Experimental study of the effects of grid configuration on the dynamical evolution of decaying turbulence in a stably stratified fluid

Abstract

A recent theoretical study by Gargett (1988, J. Geophys. Res., 93(C5): 5021–5036) of turbulent mixing in the stably stratified atmosphere and oceans has yielded reinterpretations of basic experimental observations at strong variance with earlier assessments which have been widely used in interpreting geophysical data. These differences in fundamental understanding erect a conceptual barrier to progress in understanding atmospheric and oceanic fine-scale turbulent mixing which must be overcome to make further progress. Definitive experimental tests of some of these new conjectures require comprehensive experiments in which the turbulence and internal wave generating mechanisms are systematically varied to include geometrical disturbance configurations not previously studied. We report here the results of a comprehensive series of experiments to investigate the effect of the configuration of the turbulence generating grid on the dynamics of stably stratified turbulence. The present measurements were made in a heat stratified wind-tunnel. Biplanar grids as well as parallel rods placed vertically or horizontally were used to produce the turbulence. The buoyancy frequency used was 2.1 rad s−1, resulting in rod Froude numbers larger than 100 and mesh Froude numbers larger than 20. Two velocity components and temperature were measured simultaneously, and spectra, length scales and heat fluxes were calculated. The measurements show that the normalized buoyancy flux is independent of the grid configuration, except very close to the grid in the region where the wakes from individual grid rods have not fully merged. The decay of velocity fluctuations and length-scale ratios show no significant internal waves generated by the horizontal rods. The lee-wave theory put forward by Gargett to reinterpret earlier stratified turbulence measurements is inconsistent with these results. The shape of the co-coherence between vertical velocity and fluctuating temperature is sensitive to grid configuration up to x/M of 30. Initially, the turbulence generated by horizontal rods shows stronger mixing at the Kármán vortex shedding frequency and larger co-coherence over a broad frequency interval lower than the shedding frequency. Estimates of the isotropy of the turbulence vs. downstream location are also contradictory to the hypothesis that lee-waves govern the turbulence dynamics. Based on calculated values of r.m.s. velocity fluctuations, strain rates and analytic relations between the spectra, it is confirmed that the turbulence is closest to isotropy close to the grids but becomes progressively more anisotropic as the buoyancy forces quench turbulent mixing farther downstream, irrespective of grid configuration. This is diametrically opposite to Gargett's interpretation. Off-centerline measurements, addressing another conjecture of Gargett, show the same qualitative results as on-centerline measurements reported earlier.