the dissimilarities in extension and depth of each depression, the western one
(PB) being noticeably greater than the eastern one (LB). Therefore, the little
intuitive presence of LB to the east of CS would actually be an indirect proof of
east-going supercritical flow downhill of CS, which is a hypothesis on the focus of
the present study.
Data and Data Processing
Two twin mooring lines equipped with a Conductivity-Temperature (CT) probe and
an upward-looking Acoustic Doppler Current Profiler (ADCP) located at 10 and
12 m above the seafloor, respectively, were deployed at the thalwegs of CSN and
CSS channels in bottom depths of 306 and 310 m, respectively (Fig. 1b, c).
The ADCP observations spanned the period 9th of June to 25th of September 2013,
and operated at a sampling interval of 4 min. The instruments were configured to
sample 40 vertical bins, each one 6 m thick, so that the velocity profile does not
reach the sea surface. However, it samples the Mediterranean water layer whose
upper boundary (interface) was estimated from in situ salinity observations at a
time-averaged depth of 120 m in CSN and 134 m in CSS by Bryden et al. [5], that
is, ∼180 m above the seafloor at the mooring sites. The CT observations were
collected every 2 min, starting on the 9th June as well, but they finished earlier due
to battery run out (on 24th and 27th of August at CSN and CSS, respectively). In
both cases, the sampling rate was enough to have several samples within the time
scale of tens of minutes typical of the internal waves [15].
Figure 4a, b show sudden and remarkable spike-shaped changes on the time
series of temperature and salinity that happen quite regularly during the spring
phase of the fortnightly cycle (see Fig. 4e). These series are, therefore, good
indicators of the occurrence of intense hydraulic features over CS. Velocity series,
however, are no so straightforward indicators because the high frequency velocity
fluctuations associated with the internal waves are superposed to very strong tidal
flows that make the former appear noise-like.
In order to extract more useful indicators of the high frequency processes from
the ADCP observations, the velocity profiles have been high-pass filtered with a
filter of 2 cph cut-off frequency. The three components of the high-passed velocity
have been combined to compute the variable V HF z, t
ð Þ=
1
2 ∑
3
i = 1 u i z, t
ð Þ
2 , where
u i z, t
ð Þ is the i-th component of the velocity at instant t and depth z. Since the
variable is rather noisy in time and space, moving average operators with windows
width of 45 min in time (11 data) and 30 m in space (5 depth-levels) have been
applied to V HF z, t
ð Þ in order to smooth out the noise. These operations are equivalent to calculating the variance of the series within the established time and depth
windows. The result of this procedure can be seen in Fig. 4c, d, which display the
Asymmetric Baroclinic Response to Tidal Forcing …
199
(PB) being noticeably greater than the eastern one (LB). Therefore, the little
intuitive presence of LB to the east of CS would actually be an indirect proof of
east-going supercritical flow downhill of CS, which is a hypothesis on the focus of
the present study.
Data and Data Processing
Two twin mooring lines equipped with a Conductivity-Temperature (CT) probe and
an upward-looking Acoustic Doppler Current Profiler (ADCP) located at 10 and
12 m above the seafloor, respectively, were deployed at the thalwegs of CSN and
CSS channels in bottom depths of 306 and 310 m, respectively (Fig. 1b, c).
The ADCP observations spanned the period 9th of June to 25th of September 2013,
and operated at a sampling interval of 4 min. The instruments were configured to
sample 40 vertical bins, each one 6 m thick, so that the velocity profile does not
reach the sea surface. However, it samples the Mediterranean water layer whose
upper boundary (interface) was estimated from in situ salinity observations at a
time-averaged depth of 120 m in CSN and 134 m in CSS by Bryden et al. [5], that
is, ∼180 m above the seafloor at the mooring sites. The CT observations were
collected every 2 min, starting on the 9th June as well, but they finished earlier due
to battery run out (on 24th and 27th of August at CSN and CSS, respectively). In
both cases, the sampling rate was enough to have several samples within the time
scale of tens of minutes typical of the internal waves [15].
Figure 4a, b show sudden and remarkable spike-shaped changes on the time
series of temperature and salinity that happen quite regularly during the spring
phase of the fortnightly cycle (see Fig. 4e). These series are, therefore, good
indicators of the occurrence of intense hydraulic features over CS. Velocity series,
however, are no so straightforward indicators because the high frequency velocity
fluctuations associated with the internal waves are superposed to very strong tidal
flows that make the former appear noise-like.
In order to extract more useful indicators of the high frequency processes from
the ADCP observations, the velocity profiles have been high-pass filtered with a
filter of 2 cph cut-off frequency. The three components of the high-passed velocity
have been combined to compute the variable V HF z, t
ð Þ=
1
2 ∑
3
i = 1 u i z, t
ð Þ
2 , where
u i z, t
ð Þ is the i-th component of the velocity at instant t and depth z. Since the
variable is rather noisy in time and space, moving average operators with windows
width of 45 min in time (11 data) and 30 m in space (5 depth-levels) have been
applied to V HF z, t
ð Þ in order to smooth out the noise. These operations are equivalent to calculating the variance of the series within the established time and depth
windows. The result of this procedure can be seen in Fig. 4c, d, which display the
Asymmetric Baroclinic Response to Tidal Forcing …
199