The time evolution of mixing in turbulent overturns is investigated using a combination of direct numerical simulations (DNS) and microstructure profiles obtained during two field experiments. The focus is on the flux coefficient ⌫, the ratio of the turbulent buoyancy flux to the turbulent kinetic energy dissipation rate ⑀. In observational oceanography, a constant value ⌫ ϭ 0.2 is often used to infer the buoyancy flux and the turbulent diffusivity from measured ⑀. In the simulations, the value of ⌫ changes by more than an order of magnitude over the life of a turbulent overturn, suggesting that the use of a constant value for ⌫ is an oversimplification. To account for the time dependence of ⌫ in the interpretation of ocean turbulence data, a way to assess the evolutionary stage at which a given turbulent event was sampled is required. The ratio of the Ozmidov scale L O to the Thorpe scale L T is found to increase monotonically with time in the simulated flows, and therefore may provide the needed time indicator. From the DNS results, a simple parameterization of ⌫ in terms of L O / L T is found. Applied to observational data, this parameterization leads to a 50%-60% increase in median estimates of turbulent diffusivity, suggesting a potential reassessment of turbulent diffusivity in weakly and intermittently turbulent regimes such as the ocean interior.

Observations of vertical temperature microstructure at ocean station P during the mixed layer experiment (Mile) indicate that the shape of the high-frequency temperature gradient spectrum depends on the relative strengths of turbulence and stratification. For low Cox number ((dT/dz) 2) /(dT/dz) • the linear range of the Batchelor spectrum is not well approximated by observed spectra, while for high Cox number a remarkably close correspondence to the Batchelor spectrum is found. Dissipation rates calculated by the temperature gradient spectrum cutoff wave number method show a dramatic contrast in turbulence between low and high wind speed periods separated by only 3 hours, showing that the response of the mixed layer and transition zone to wind forcing is rapid. Some indication is found that the thermocline may also respond rapidly to surface forcing. E 1.0 Spectrum of heated laminar jet ß ß ß ß _ ß ß Uncorrected Spectrum o Corrected Spectrum, i,af •' ,bf4.cf 6

A 38‐day, 5990‐cast microstructure study on the equator performed during the onset of the 1991–1993 El Niño shows the effect on small‐scale activity at 140°W of an equatorial Kelvin wave. By using two ships, data were taken continuously from November 4 to December 12, 1991, near the National Oceanic and Atmospheric Administration Pacific Marine Environmental Laboratory mooring at O°N, 140°W. The ships occupied the station sequentially with a 3.5‐day overlap for intercalibration. Variability in currents was observed on tidal periods, and periods of 4 days (presumably equatorially trapped internal gravity waves), 8 days (cause unknown), 20 days (tropical instability waves), and longer (Kelvin waves). Variation in water structure occurred most prominently on the timescale of Kelvin waves. The diurnal cycle typical of that location was observed: nocturnal deepening of the surface mixed layer was accompanied by a “deep cycle,” bursts of turbulence penetrating into the stratified region below the nighttime mixed layer. During the observational period, one Kelvin wave trough and one crest passed through the site. Changes accompanying the phase change in the Kelvin wave included a reversal of the near‐surface current, a deepening of the thermocline, and a change of water mass. Changes in small‐scale activity included a tenfold decrease of the thermal dissipation rate and a fourfold decrease of the rate of heat transport downward from the mixed layer. The nighttime mixed layer deepened from 30 to 60 m. The thickness of the stratified region in which nocturnal turbulence bursts occurred, the deep cycle region, thinned from 40 to 20 m because it was confined between the bottom of the nighttime mixed layer and the low‐shear region near the core of the undercurrent. The decrease in downward heat flux observed at this passage of the downwelling Kelvin wave front could explain the rapid sea surface temperature (SST) increase seen at El Nino onsets. The magnitude of the change in vertical flux is similar to the magnitude of the change in horizontal advection. This process would produce a warmer SST much more quickly than could the advection of warm waters eastward.

No abstract

Twelve days of microstructure measurements at the equator (140°W) in November 1984 showed a surprisingly strong effect of both the daily cycle of solar heating and wind on mixing in the upper ocean. Because of limited variations in atmospheric forcing and currents during the experiment, processes in the daily mixing cycle were similar from day to day. Only the intensity of mixing varied. The lower boundary of the diurnal surface layer separated two distinct mixing regimes, the diurnal surface layer and the thermocline. Within the diurnal surface layer (which extended to 10‐ to 35‐m depth), turbulent kinetic energy dissipation rates ε varied relatively little. Although variations in surface layer depth coincided with the daily change in direction of air‐sea surface buoyancy production of turbulent kinetic energy (or simply, the surface buoyancy flux), ε was significantly greater relative to the buoyancy flux than was expected for a simple convective layer. In the thermocline below the diurnal surface layer, ε was highly intermittent; the day‐night cycle was stronger, and variability was enhanced by turbulent “bursts” of 2–3 hours duration, which may be related to internal wave breaking events. The turbulent heat flux crossing 20‐m depth was almost equal to the surface heat flux less the irradiance penetrating below 20 m. Seventy percent of the surface heat flux was transported vertically to the water below 30 m by turbulent mixing. Only a negligible amount penetrated to the core of the Equatorial Undercurrent. The gradient Richardson number Ri distinguishes between statistically different mixing environments. However, ε cannot be predicted from the value of Ri, since the intensity of mixing depends on the intensity of forcing in a way not specified by the value of Ri alone.