Role of wave breaking in the air–sea interaction - Breaking and Dissipation of Ocean Surface Waves - page 395

icant (outside the tropical and polar areas). For example, the correlation coefficient between
simulations and data (World Ocean Atlas, Levitus , 1982 ), for globally averaged values of
SST at 35 ◦ N, increased from 68% to 93%.
In this non-breaking wave-mixing context, the laboratory experiment of Dai et al. ( 2010 )
deserves a special mention. While the scale of the small wave tank is incomparable to the
open-sea and global-ocean simulations described above, the idea of the experiment is very
convincing and highlights the wave-induced mixing most prominently.
The experiments were carried out in a wave tank at the First Institute of Oceanography,
Qingdao, China. The wave tank had refrigeration tubes installed under the tank's bottom
in order to cool the water from below and create a stable stratification. A vertical array
of temperature sensors was deployed to record the water-temperature variations automati-
cally. Nine sensors were vertically spaced 1 0 cm apart, from 4 cm through 12 cm below the

water surface. The temperature difference between the upper-most and lower-most sensors
was set to 5 ◦ C.
Then, the refrigeration was switched off and the molecular diffusion was allowed to
take its course. The start time for observations and numerical simulations was chosen as
the moment when the ice at the bottom just disappeared. The slow destratification process
took about 20.5 hours for the temperature difference to reduce from 5

.

6 ◦ C.
In the next set of experiments, gentle non-breaking surface waves were generated mechan-
ically once the ice melted. Measurements were conducted at the centre of the tank, away
from the boundary layers, in the absence of wind, internal waves, shear currents and other
potential sources of turbulence. In all cases the destratification time decreased by some
two orders of magnitude. For example, for waves with amplitude a

0 ◦ Cto0 . .

= 1 cm and wave-

length 08, it took only 25 minutes to reach the

same temperature difference which the molecular viscosity achieved over 20 hours. A one-
dimensional diffusion model, with and without wave-turbulence diffusion

λ = 75 cm, i.e. steepness of = ak = 0 .

ν wave , was able
to describe both the molecular-diffusion observation and wave-mixing outcomes.
The effect of the wave-induced turbulence is apparent in the experiment of Dai et al.
( 2010 ), but interesting further questions arise with respect to the role of stratification in
such mixing. The stable stratification is known to suppress the turbulence, at least to some
extent (see, for example, discussions in Section 9.1.2 with respect to the atmospheric tur-
bulence). This role appears to be small in the described experiment, where numerical
simulations were able to reproduce the experiment by accounting for ν wave and without

considering the stratification effects, but such a role needs further examination as it has
significant implications for the wave-induced mixing through the thermocline. If the wave
orbital motions are such that corresponding wave Reynolds numbers (7.70) are above the
critical value (7.67) within the thermocline, will the turbulence be generated and effectively
facilitate the mixed-layer deepening? The answer is essential for modelling upper-ocean
mixing and air-sea interactions at all temporal scales, from hours (e.g. hurricanes) to years
(i.e. climate). For example, for hurricanes with their asymmetric structure of the wave
fields, asymmetric mixing has been indicated both in measurements and in modelling (e.g.
Fedorov et al. , 1979 ; Pudov et al. , 1979 ; Price , 1981 ).