responds to the tidally-induced fluctuations of the total flow, which modify its
position and can even change its sign. Actually, during the ebb tide the whole water
column may be flowing eastwards, particularly in spring tides (Fig. 6), which
reduces the interface slope due to the cross-strait geostrophy. At the same time, the
interface sinks to make room for the large volume of Atlantic water displaced
towards the Mediterranean during this tidal cycle. Both processes act jointly to
bring mixed water deeper in CSN than in CSS and explain why it reaches the CSN
probe earlier and quits it later (labels (2) and (3), respectively, in Fig. 6). It is only
by the end of the ebb tide, when the interface is at its maximum depth, that CSS
probe registers the mixed water.
This description would suggest a gradual change rather than the pronounced
temperature and salinity peaks that echo the arrival of this water at CSS, profusely
shown in the figures of the paper. The drawback is overcome by letting the internal
hydraulics act out. The peaks at this site have been related in section “Ebb Tide” to
the sudden arrival of the internal bore released when the east-of-CS hydraulic jump
decays, a circumstance that happens by the end of the ebb tide when the thickness
of the Atlantic layer is maximum. If this jump is formed over the eastern rim of the
sill elevation (see Fig. 1b), the bore will take a little longer to get to CSN than to
CSS, since the former is ∼2.5 km farther from the rim than the latter. Converted
into time, it means 40–45 min for a typical first baroclinic mode speed of 1 ms
−1 ,
CSS leading CSN in satisfactory agreement with the delay inferred from
cross-correlating the series of temperature (or salinity) at both sites discussed in
section “Ebb Tide”. The former analysis relies on a cross-strait orientation of the
jump and on its simultaneous decay throughout its extension, which appear as
reasonable assumptions. Notice that the interpretation of temperature and salinity
peaks at CSS in terms of the passing of internal bores is further supported by the
short-living deep current reversals observed in CSS (labels (4) in Fig. 6) that
accompanies the occurrence of the spikes.
The considerably more studied hydraulic jump formed westwards of CS during
the flood tide [10, 11, 13] leaves clearer signatures at CSS than at CSN, but only in
the velocity profiles and not in the temperature or salinity series. The profiles reveal
high-frequency fluctuations that shoal along with the interface in CSS (Fig. 8b).
This uplift of the interface is a distinctive upstream response of hydraulically
controlled flows, in which the interface upstream of the control section (the sill crest
in this case) must shoal in order to accommodate larger (tidal) volume transport
(e.g., [28]; Sect. 1.8). Such a response during the flood tide is therefore indicative of
the recovery of hydraulic control over the sill that prevents the propagation of
internal disturbances from the Atlantic to the Mediterranean.
The enhancement of the high-frequency velocity fluctuations at CSS can be due
to either the presence of shear instabilities or to the trace of a transient upstream
internal hydraulic jump evident in the high-resolution non-hydrostatic numerical
model of Sánchez-Garrido et al. [13] (see Fig. 2a). We put forward the latter as their
origin. This secondary jump would be associated with a second control section
arising east of the sill as a result of the shoaling of the interface there and the
concomitant reduction of the phase speed of internal disturbances. This control
Asymmetric Baroclinic Response to Tidal Forcing …
207
position and can even change its sign. Actually, during the ebb tide the whole water
column may be flowing eastwards, particularly in spring tides (Fig. 6), which
reduces the interface slope due to the cross-strait geostrophy. At the same time, the
interface sinks to make room for the large volume of Atlantic water displaced
towards the Mediterranean during this tidal cycle. Both processes act jointly to
bring mixed water deeper in CSN than in CSS and explain why it reaches the CSN
probe earlier and quits it later (labels (2) and (3), respectively, in Fig. 6). It is only
by the end of the ebb tide, when the interface is at its maximum depth, that CSS
probe registers the mixed water.
This description would suggest a gradual change rather than the pronounced
temperature and salinity peaks that echo the arrival of this water at CSS, profusely
shown in the figures of the paper. The drawback is overcome by letting the internal
hydraulics act out. The peaks at this site have been related in section “Ebb Tide” to
the sudden arrival of the internal bore released when the east-of-CS hydraulic jump
decays, a circumstance that happens by the end of the ebb tide when the thickness
of the Atlantic layer is maximum. If this jump is formed over the eastern rim of the
sill elevation (see Fig. 1b), the bore will take a little longer to get to CSN than to
CSS, since the former is ∼2.5 km farther from the rim than the latter. Converted
into time, it means 40–45 min for a typical first baroclinic mode speed of 1 ms
−1 ,
CSS leading CSN in satisfactory agreement with the delay inferred from
cross-correlating the series of temperature (or salinity) at both sites discussed in
section “Ebb Tide”. The former analysis relies on a cross-strait orientation of the
jump and on its simultaneous decay throughout its extension, which appear as
reasonable assumptions. Notice that the interpretation of temperature and salinity
peaks at CSS in terms of the passing of internal bores is further supported by the
short-living deep current reversals observed in CSS (labels (4) in Fig. 6) that
accompanies the occurrence of the spikes.
The considerably more studied hydraulic jump formed westwards of CS during
the flood tide [10, 11, 13] leaves clearer signatures at CSS than at CSN, but only in
the velocity profiles and not in the temperature or salinity series. The profiles reveal
high-frequency fluctuations that shoal along with the interface in CSS (Fig. 8b).
This uplift of the interface is a distinctive upstream response of hydraulically
controlled flows, in which the interface upstream of the control section (the sill crest
in this case) must shoal in order to accommodate larger (tidal) volume transport
(e.g., [28]; Sect. 1.8). Such a response during the flood tide is therefore indicative of
the recovery of hydraulic control over the sill that prevents the propagation of
internal disturbances from the Atlantic to the Mediterranean.
The enhancement of the high-frequency velocity fluctuations at CSS can be due
to either the presence of shear instabilities or to the trace of a transient upstream
internal hydraulic jump evident in the high-resolution non-hydrostatic numerical
model of Sánchez-Garrido et al. [13] (see Fig. 2a). We put forward the latter as their
origin. This secondary jump would be associated with a second control section
arising east of the sill as a result of the shoaling of the interface there and the
concomitant reduction of the phase speed of internal disturbances. This control
Asymmetric Baroclinic Response to Tidal Forcing …
207