Key points are not available for this paper at this time.
The dissipation rate of kinetic energy, ε, was estimated from adjacent and simultaneous measurements with a submarine and a vertical profiler. The submarine cycled up and down through the water column measuring both a vertical and a horizontal component of the turbulent velocity. The free-fall profiler measured two perpendicular, horizontal components, The mean value of the dissipations from the two systems, between 50 and 120 m depth along a 25 km transect, differed by a factor of 1.8. This difference was statistically significant and led us to examine the statistical distribution of the dissipation estimates. For each profile from the submarine and the vertical profiler, the probability density function of the dissipation estimates is bimodal and well represented by a mixture of two lognormal distribution. This division of the data is a combination of an active mode and a relatively quiescent mode. The mean values of ε for the more dissipative modes are comparable for the submarine and the vertical profiler, often exceeding 20ρνN2. The mean rates of the quiescent mode are not all the same. Those from the vertical velocity components are one decade smaller than the other three estimates which are all based on horizontal velocity components. Thus, this lower mode of the turbulence is anisotropic. All the means from the lower mode are smaller than 20ρνN2. The effect of anisotropy on the dissipation estimates is not sufficient to explain the differences in average dissipation between the two platforms. The difference was due to heterogeneity of the turbulence, on kilometer scales, which was revealed by the denser spatial sampling of the submarine but not resolved by the vertical profiler.
Yamazaki et al. (Thu,) studied this question.