AIR SEA INTERA(”II0NS
395
(Rcsch, 1968). the recording of surface motions required the use of sensitive
capacitance wave gauges (Ramamonjiarisoa. 1971). and the determination
of mean values and turbulent fluctuations of humidity raises difficult
problems (Coantic and Leducq, 1969). While instrument research is still
going on concerning this last point. reasonably accurate determinations
have becn obtained using an automatic dew-point instrument for the mean
values. and a small microwave refractometcr (supplied by J. Lane, of
R.S.R.S.. and operated by R. Bolgiano and Z. Warhaft) for the turbulent
fluctuations.
After conditioning and visual control, data are recorded and analyzed
using either analogue statistical processing equipment or a high performance digital data acquisition system allowing direct preprocessing and
control of information, and the editing of digital magnetic tapes for detailed
processing by a computing center.
4. RES~JLTS CONCERNING WIND-GENERATED WAVES
An experimental study of the statistical properties of the wavy interface
has been undertaken as a first step in a more general investigation of the
wave-generating mechanism. Some typical results will now be presented and
discussed.
The frequency spectrum of the wave profile has been determined in the
small (Ramamonjiarisoa, 1971) as well as in the large facility (Ramamonjiarisoa, 1973). Results have been analysed in terms of various theoretical
findings. For frequencies lower than the dominant wave’s frequency no, they
are found to obey Kitaigorodskii’s (1961) similarity law:
log(F(o) * 4 g * ) = a + h ( u + o / g ) .
From these, and other experimental data sources, the constants seem to
obey Mitsuyasu’s (1968) relations:
u _Y -8.78;
h 2 1 . 7 1 ( g X / ~ + ~ ) ~ . ~ ~ ~ .
Phillips’ ( 1958) equilibrium range has been effectively observed for frequencies between approximately 2no and 10 Hz. In the intermediate zone, i.e.,
between no and 2no. the usual overshoot and undershoot effects noted by
other investigators are displayed.
An interesting property appears when wave spectra from different sources
are plotted (see Fig. 3) as F(n)/F(no) as a function of n/no. It is clear that the
relative width of the spectral peak increases when the fetch increases. The
reason for this kind of “spectral ray widening effect” can be attributed to
various linear or nonlinear modulation processes inside the liquid phase.
Another explanation could lie in the turbulent wind’s variability. As recently
395
(Rcsch, 1968). the recording of surface motions required the use of sensitive
capacitance wave gauges (Ramamonjiarisoa. 1971). and the determination
of mean values and turbulent fluctuations of humidity raises difficult
problems (Coantic and Leducq, 1969). While instrument research is still
going on concerning this last point. reasonably accurate determinations
have becn obtained using an automatic dew-point instrument for the mean
values. and a small microwave refractometcr (supplied by J. Lane, of
R.S.R.S.. and operated by R. Bolgiano and Z. Warhaft) for the turbulent
fluctuations.
After conditioning and visual control, data are recorded and analyzed
using either analogue statistical processing equipment or a high performance digital data acquisition system allowing direct preprocessing and
control of information, and the editing of digital magnetic tapes for detailed
processing by a computing center.
4. RES~JLTS CONCERNING WIND-GENERATED WAVES
An experimental study of the statistical properties of the wavy interface
has been undertaken as a first step in a more general investigation of the
wave-generating mechanism. Some typical results will now be presented and
discussed.
The frequency spectrum of the wave profile has been determined in the
small (Ramamonjiarisoa, 1971) as well as in the large facility (Ramamonjiarisoa, 1973). Results have been analysed in terms of various theoretical
findings. For frequencies lower than the dominant wave’s frequency no, they
are found to obey Kitaigorodskii’s (1961) similarity law:
log(F(o) * 4 g * ) = a + h ( u + o / g ) .
From these, and other experimental data sources, the constants seem to
obey Mitsuyasu’s (1968) relations:
u _Y -8.78;
h 2 1 . 7 1 ( g X / ~ + ~ ) ~ . ~ ~ ~ .
Phillips’ ( 1958) equilibrium range has been effectively observed for frequencies between approximately 2no and 10 Hz. In the intermediate zone, i.e.,
between no and 2no. the usual overshoot and undershoot effects noted by
other investigators are displayed.
An interesting property appears when wave spectra from different sources
are plotted (see Fig. 3) as F(n)/F(no) as a function of n/no. It is clear that the
relative width of the spectral peak increases when the fetch increases. The
reason for this kind of “spectral ray widening effect” can be attributed to
various linear or nonlinear modulation processes inside the liquid phase.
Another explanation could lie in the turbulent wind’s variability. As recently