During the last few years, I have developed a new type of neutrally buoyant float (see image on right) designed to be used in energetic turbulent flows such as those found in the top and bottom boundary layers of the ocean. A combination of accurate ballasting, a compressibility matched to that of seawater and high drag is used to make these floats follow the motion of water parcels accurately. The floats measure their depth and are acoustically tracked in the horizontal and thus produce measurements of vertical and horizontal velocity. They measure the temperature and vertical vorticity (from spin rate) of the water surrounding them. We have made about 250 deployments of these floats, in a wide variety of turbulent coastal and open ocean environments.
A major task over the past few years has been to evaluate the accuracy of the floats and understand their behavior in homogeneous turbulent flows. A paper describing the floats, their accuracy, and some of the analysis methods has been published in the Journal of Atmospheric and Oceanic Technology. A second paper, describing in detail the Lagrangian spectra of vertical velocity and vorticity in high Reynolds number turbulence and explaining it using concepts from homogeneous turbulence theory has been published in theJournal of Fluid Mechanics. We find good agreement between the theory and our observations of acceleration spectrum. The vorticity spectra agree only at low frequency; we conclude that we do not understand the what controls the high frequency fluctuations in float rotation rate. Our major result is: the large-eddy-frequency and turbulent kinetic energy dissipation rate of a turbulent flow can be determined by Lagrangian floats as long as the overturning scale of the flow is larger than the float.
Our most exciting new work has been to use the Lagrangian floats to help understand turbulence in a density stratified environment. Here, both internal waves and turbulence can exist and it is often difficult to separate them. We find that internal waves can be separated from turbulence based on Lagrangian frequency as shown in Fig. 1. Here, average spectra of vertical (W) and horizontal (U,V) velocity measured by floats from strongly mixing but stratified environments (mostly in Knight Inlet) are plotted vrs frequency. For w > N (right of the dashed vertical line) the data are consistent with turbulence: the velocity is nearly isotropic since the three components have similar energy, and the spectral slope is -2, as we have found for the inertial subrange of unstratified turbulence. For w < N (left of the dashed vertical line) the data are consistent with internal waves: the velocity is highly anisotropic, there is more horizontal kinetic energy than vertical, and the ratio of vertical to horizontal energy varies with frequency in agreement with the usual consistancy relation for internal waves. Furthermore, the W spectrum is white, while the (U,V) spectra have a -2 slope as found in the Garrett-Munk spectrum.
The form of these spectra imply a new parameterization for stratified turbulence relating the energy in the waves and the turbulent mixing rate. For sufficiently energetic flows we find the rate of kinetic energy dissipation, e, is proportional to the mean vertical kinetic energy in the waves and the stratification N. This should apply for internal wave fields about 10-20 times more energetic than the open ocean Garrett-Munk levels. Such mixing rates are not uncommon in the Littoral zone.
RESULTS
1.. Lagrangian velocity spectra in stratified flows can separate the energy of internal waves (frequencies less than N) from that of turbulence (frequencies greater than N).
2. All of the anisotropy in stratified turbulence appears to be due to internal waves, the turbulence remains isotropic.
3. The matching of these two regimes at N implies a new parameterization of internal wave driven mixing in which the kinetic energy dissipation rate is proportional to the internal wave vertical kinetic energy times N.
4. This parameterization should be appropriate for energetic shallow water regimes.
RELATED PROJECTS
These floats are close relatives of those used to study deep convection in the Labrador Sea. Much of the data used here came from measurements in Knight Inlet as part of an ONR322PO study of solibores. Mixing processes in these various environments are similar in many ways and we learn the most by comparing and contrasting them.
REFERENCES
D'Asaro, E. A., D. M. Farmer, J. T. Osse, G. T. Dairiki, 1996, A Lagrangian Float, J. Oceanic and Atmos. Tech., 13, 6, 1230-1246
Lien, R.C., E.A. D'Asaro, G.T. Dairiki, 1998, Lagrangian Frequency Spectra of Vertical Velocity and Vorticity in High Reynolds' Number Oceanic Turbulence, J. Fluid Mechanics,Volume 362, pp 177-198 ,10 May 1998
Web site for Knight Inlet measurements http://pinger.ios.bc.ca:80/cruises/knight95/start.html