Geoscience Reference
In-Depth Information
Furthermore, based on comparisons of the model and experiment,
Babanin & Makin
(
2008
) revealed that the most distant outliers in their dependences of
C
D
on
U
10
and
U
10
/
c
p
were field cases with high wind gustiness. Once these obvious outliers were removed,
some crude dependences emerged. They are in approximate agreement with some other
known field parameterisations. The remaining scatter, however, appeared to be brought
about by causes other than the wind instabilities only and does not asymptote to cases of
the zero trend and minimal gustiness.
One of such causes can be wave breaking. As discussed throughout the topic, breaking
is only indirectly related to the wind, unless the wind is very strong (see
Section 9.1.3
), and
for the same wind speed there can be different breaking rates (see, for example
Figure 5.20
). If breaking happens, however, it does affect the momentum and energy fluxes
(
Kudryavtsev & Makin
,
2002
;
Babanin
et al.
,
2007b
,
Section 8.3
). Therefore, correlation
of the drag coefficient
C
D
with wave breaking can be expected.
In
Figure 9.3
,
C
D
dependences are plotted for records used by
Babanin &Makin
(
2008
),
but only those in their Table 1 for which the dominant breaking rates
b
T
(2.3)
were known
and they were
b
T
≥
02. As in
Section 5.3.2
(see
Figure 5.30
), this is done to avoid bias
and scatter due to zero-breaking contributions, which are not meant to belong to these
dependences anyway, when the rates are low.
The left panel is a regular
C
D
versus
U
10
plot. It is instructive to note that there are no
wind-speed values less than 7
0
.
s. This, again, points to the change in the wave-system
dynamics at such wind speed, due to the absence of breaking, at least at Lake George.
In
Section 5.3.4
, this change was discussed with respect to the dissipation in the water
column beneath the waves (see parameterisation
(5.74)
), and now it is apparent with respect
to sea drag and the air-side boundary layer.
The right panel of
Figure 9.3
is a plot of
C
D
versus
b
T
. The correlation of 0.8 is lower
than that of 0.9 for the wind, but it is still quite high in every regard and therefore the depen-
dence is significant. This is particularly essential if we notice that the breaking as such is
not an independent player in WBL, it only alters the drag imposed by the wind-generated
waves when those are breaking. Such a large correlation of
C
D
with the breaking rates of
dominant waves points out that these alterations, which underlie any existing drag-versus-
wind dependence, need attention, examination and parameterisation to reduce the scatter.
Indeed, as we know very well, wave breaking is distributed across all wave scales, and
regardless of whether they produce whitecapping or not the breaking events are likely to
introduce the flow separations at respective scales and associated changes to the momen-
tum fluxes. Therefore, at the very least, the link between sea drag and breaking is not as
straightforward as a correlation between
C
D
and dominant breaking rates
b
T
. Whitecap-
ping, when present, certainly changes the roughness of the surface locally, and its patches
would affect the sea drag too, particularly if the whitecap coverage area is large. And the
extent of the whitecapping plumes relates to the breaking severity which is not directly
identified by the breaking-probability parameter in the right subplot of
Figure 9.3
either
(see
Section 3.1
). Also, in the turbulent wake of breaking waves, there is an active dis-
sipation of short waves (see
Section 7.3.4
). Since those are the main contributors to the
.
5m
/
Search WWH ::
Custom Search