Discriminating Salt Fingering from Turbulence-induced Microstructure. Analysis of Towed Temperature–Conductivity Chain Data

Abstract

Several techniques are explored for discriminating salt fingering from turbulence-induced microstructure in towed array measurements of temperature and conductivity. Such separation of source mechanisms for ocean mixing is critical before applying models to estimate the vertical diffusivity from microstructure measurements. The density ratio (R_ρ); the conductivity gradient spectral slope; the conductivity gradient kurtosis; and the optimal discriminator for a two-parameter (slope and kurtosis) system, the likelihood ratio, are examined. Density ratio can indicate where salt fingering is likely or not possible, but alone it cannot distinguish between turbulence-induced microstructure and salt fingering for R_ρ values that permit either. From the “control” microstructure patches previously reported by Mack, it is shown that the probability density functions (PDFs) of slope and kurtosis for salt fingering differ from the corresponding PDFs for turbulence-induced microstructure. The parameters of the likelihood ratio are derived from the nearly Gaussian PDFs of slope and log-kurtosis for these “control” patches, and the technique is then applied to a 10-km tow in the Sargasso Sea to determine the origin of the microstructure patches found within the array's 15-m vertical aperture. The two-dimensional likelihood ratio discrimination identifies turbulence at some sites where R_ρ alone would suggest salt fingering; it improves upon the individual use of either slope or kurtosis, and it minimizes the errors inherent in such discrimination approaches. It should be a useful technique for determining the relative importance of salt fingering and turbulence over large ocean regions, such as in the “Central Waters,” or during tracer release experiments.