These high frequency temperature and salinity oscillations appear to be correlated to
the perturbations of the current, which are visible at shallower depths in CSN (labels
(2)–(3) in Fig. 6, see also Fig. 5a). The correlation suggests certain proximity of the
instruments to the deeper part of the interfacial layer, whose small-scale vertical
oscillations would leave correlated high-frequency signatures in all series. This period
comes to an end around low water, when the ebb tide ceases and the flood cycle begins.
Salinity and temperature series in CSN show another sudden change that brings the
observed values back to the expected for Mediterranean waters (labels (3) in Fig. 6).
This change is greater than the previous one (cf. labels (2) in Fig. 6), it happens more
regularly and it is usually accompanied by a pulse of current that affects the whole
water column in the variance series (see labels (1) in Fig. 5a and also labels (4) in
Fig. 8a later on). We reason that these features are caused by the westward progression
of an internal bore and the tied in internal wave packet that follows the release of the
hydraulic jump formed east-of-CS.
The situation in CSS appears somewhat different. Here, the deepening of the
interface during the ebb cycle does not reach the CT probe for as long as in CSN,
and salinity and temperature fluctuations just consist of a single peak up to four
times greater than those observed in CSN (Fig. 6a). It requires larger fraction of
Atlantic water in the mixed water sampled by the CSS probe than by the CSN
probe, which can only be explained if the interface goes deeper in the former site,
although this deep excursion lasts shorter in the latter.
The peak at CSS occurs during the last hour of the ebb cycle and shortly before
than the already mentioned second peak in CSN is detected (label (3) in Fig. 6).
Actually, a lagged cross-correlation between the salinity (or temperature) series at
both sites indicates that the second peak in CSN lags the sharp peak in CSS
by ∼ 50 min. Despite being so noticeable in temperature or salinity, the process
that causes these fluctuations in CSS does not leave such a clear signature in the
series of high-frequency velocity as it does in CSN (see labels (1) in Fig. 5a).
Instead of affecting the whole water column, as it is the case in CSN, peaks in CSS
only leave a weak visible signature in the lower layer velocity consisting of short
reversals (less than half an hour, see labels (4) in Fig. 6b) of the total current, which
is about starting flowing westward. And the reversal does not even happen on a
regular basis, but rather occasionally.
Flood Tide
The flood tide cycle begins after the occurrence of the sharp temperature and
salinity spikes, which signpost the end of the ebb. During this cycle and under
spring tide conditions, the east-going flow in CS reverses and the whole water
column moves westwards (Fig. 6b). Following the progressive acceleration of the
Mediterranean water flow, the interfacial layer rises up to even a few meters from
the free surface. In fact, it is not unusual to detect cold sea surface signatures in the
eastern approaches of CS in remote sensed images during the flood tide (see Fig. 7,
Asymmetric Baroclinic Response to Tidal Forcing …
203
the perturbations of the current, which are visible at shallower depths in CSN (labels
(2)–(3) in Fig. 6, see also Fig. 5a). The correlation suggests certain proximity of the
instruments to the deeper part of the interfacial layer, whose small-scale vertical
oscillations would leave correlated high-frequency signatures in all series. This period
comes to an end around low water, when the ebb tide ceases and the flood cycle begins.
Salinity and temperature series in CSN show another sudden change that brings the
observed values back to the expected for Mediterranean waters (labels (3) in Fig. 6).
This change is greater than the previous one (cf. labels (2) in Fig. 6), it happens more
regularly and it is usually accompanied by a pulse of current that affects the whole
water column in the variance series (see labels (1) in Fig. 5a and also labels (4) in
Fig. 8a later on). We reason that these features are caused by the westward progression
of an internal bore and the tied in internal wave packet that follows the release of the
hydraulic jump formed east-of-CS.
The situation in CSS appears somewhat different. Here, the deepening of the
interface during the ebb cycle does not reach the CT probe for as long as in CSN,
and salinity and temperature fluctuations just consist of a single peak up to four
times greater than those observed in CSN (Fig. 6a). It requires larger fraction of
Atlantic water in the mixed water sampled by the CSS probe than by the CSN
probe, which can only be explained if the interface goes deeper in the former site,
although this deep excursion lasts shorter in the latter.
The peak at CSS occurs during the last hour of the ebb cycle and shortly before
than the already mentioned second peak in CSN is detected (label (3) in Fig. 6).
Actually, a lagged cross-correlation between the salinity (or temperature) series at
both sites indicates that the second peak in CSN lags the sharp peak in CSS
by ∼ 50 min. Despite being so noticeable in temperature or salinity, the process
that causes these fluctuations in CSS does not leave such a clear signature in the
series of high-frequency velocity as it does in CSN (see labels (1) in Fig. 5a).
Instead of affecting the whole water column, as it is the case in CSN, peaks in CSS
only leave a weak visible signature in the lower layer velocity consisting of short
reversals (less than half an hour, see labels (4) in Fig. 6b) of the total current, which
is about starting flowing westward. And the reversal does not even happen on a
regular basis, but rather occasionally.
Flood Tide
The flood tide cycle begins after the occurrence of the sharp temperature and
salinity spikes, which signpost the end of the ebb. During this cycle and under
spring tide conditions, the east-going flow in CS reverses and the whole water
column moves westwards (Fig. 6b). Following the progressive acceleration of the
Mediterranean water flow, the interfacial layer rises up to even a few meters from
the free surface. In fact, it is not unusual to detect cold sea surface signatures in the
eastern approaches of CS in remote sensed images during the flood tide (see Fig. 7,
Asymmetric Baroclinic Response to Tidal Forcing …
203