expected spring-neap tidal modulation in both CSN and CSS sites, more neatly in
the latter.
Finally, the location of the interface between Atlantic and Mediterranean waters
is relevant in this study. A proxy for this interface is the surface of maximum
vertical shear of the horizontal velocity [5, 23, 24], which has been estimated from
the low-passed along-strait velocity, using a 5 cph cut-off frequency filter, and the
same moving average window of 5 vertical bins in the vertical. The time-averaged
locations of the interface so computed have been found at 110 (36) m and 153
(52) m at CSN and CSS respectively (bracketed values indicate the standard
deviation), in good agreement with those provided by Bryden et al. [5]. A good deal
of the interface variability is tidally-induced with semidiurnal oscillations of several
tens of meters centered at M 2 frequency.
General Description
The spikes in Fig. 4a, b indicate the arrival to the bottom of CS of Mediterranean
water mixed up with fresher and warmer overlying Atlantic water (the CT probes
are hardly 10 m above the seafloor). As mixing depends on the strength of the tidal
currents, the good matching between the periods of spikes appearance and spring
tide is not surprising at all. Likewise, the diurnal inequality in the internal
hydraulics of the strait driven by the important diurnal tidal currents [7, 9, 15] is
also mirrored by the spike size, which are larger every two peaks. The size is greater
in CSS (notice the different scale in Fig. 4a, b), surely as a result of the stronger
currents across this channel. More importantly, the shape of the spikes differs
noticeably from one site to the other, a fact that is better appreciated in Fig. 5.
Figure 4c and d display the footprint on the velocity series of the
high-frequency motions associated with the internal hydraulics. Again, the contour
intensity follows the spring-neap cycle, a fact that is particularly evident in the case
of CSS. The cycle is less apparent in CSN, yet identifiable, because the variability
tends to concentrate in the uppermost part of the water column, contrary to what
happens in CSS where a mid-depth maximum is weakly suggested (see also
Fig. 8b).
A quick inspection of Fig. 5, which enlarges the series inside the rectangle of
Fig. 4, confirms the above mentioned features, but it also reveals new ones of
interest. First, the temperature and salinity peaks always occur during the falling
tide that coincides with the ebb current due to the standing-wave nature of the tidal
wave in the strait [14]. The direction of the ebb current is to the east and, if strong
enough (a situation met in spring tides), reverses the deep Mediterranean layer that
momentarily will flow towards the Mediterranean Sea (Fig. 6b). A second noteworthy feature is the regular semidiurnal periodicity of the high frequency currents
in CSS (Fig. 5b), with the expected greater similarity every two semidiurnal cycles
due to the diurnal inequality. The pattern is not as clear in CSN (Fig. 5a),
although it is still detectable. The contours, however, are now found during the
200
J. García-Lafuente et al.
Précédent

- 203/610

Suivant