ADCP were collected over nearly all of the
autumn campaign along the ship track, with repetition of different transects, even when no CTD
sampling was done. The lines along which CUfrent data are available are depicted in Fig. 6. The
vertical resolution is of 2.5 m, while the horizontal one is of about 10-20 m,depending on the ship
speed. The instrument is not able to make measurements in the first 8 m near the surface and in
the layer near the bottom, corresponding to 15%
of the whole depth. The surface and bottom layer
currents have been extrapolated for each component, considering in the surface layer a constant
value equal to the first available datum, and in the
bottom layer, a logarithmic decay from the deepest datum to the zero bottom value. The accuracy
in the current measurement is of the order of one
cm/s, due also to the fact that the bottom tracking is always available and the speed of the ship
is, in this way, measured with the same accuracy
as the water velocities. The possible misalignment of the transducers with respect to the ship's
stern and eventual errors in the gyrocompass are
corrected in the first part of the processing. The
data have then been averaged horizontally every
500 m along each transect. Thereafter, the
method developed by Candela et al. (1992) was
applied to this data set in order to obtain the circulation of the Northern Adriatic. It permits to
separate three components in the signal: the
barotropic part of the tide, the "steady" and the
"residual" current. The steady component represents the time-independent part of the current
over the period considered, while the residual
one contains the remaining spatio-temporal variability. The method uses the least square procedure to fit horizontally the data to a function sum
of a non-depending time term and of a linear
combination of functions like sine and cosine of
the tidal frequency, whose coefficients are functions of the spatial co-ordinates of the point considered. As suggested by the author of the
method and already tested in a previous work
(Gacic et al. 1998), polynomials were the most
appropriate as the basis for the steady term and
for the coefficients of the tides. The residual term
presents a time variability on the order of a day
or greater, and thus may contain signals related to
wind effects or to discharge pulses of the rivers,
among which the Po plays the main role. In fact,
the river discharge is seen in the residual field
only in concomitance with a sudden variation in
the mean flow.
Hydrology of the Northern Adriatic
27
Results and Discussion
From its elongated shape, the T-5 diagram indicates a high variability mainly in temperature,
but also in salinity; in fact these parameters present values varying from 13.5 to 23°C and 30 to
38. Corresponding sigma-t variations are from
21 kglm 3 to 28.5 kglm 3 • The presence of three
main different water bodies can be observed: the
light waters, located to the left of the upper part
of the diagram are warmer (20-23°C) and less
salty (30-35). The dense waters instead, are located at the bottom and present lower values in
temperature but higher in salinity (around 14°C
and 37-38). The intermediate waters could be
recognised in the central part of the diagram,
being characterised mainly by two close branches. The application of cluster analysis provides
more information than the classical T -5 diagram
analysis, because it gives not only the characteristics of the possible water bodies presented in
the area, but also their spatial distribution. The
data have been separated into three dusters, each
one being associated with a water mass possessing distinct characteristics (Fig. 2). The first duster groups together water probably of a fluvial
origin with a temperature around 20.91±0.63°C,
and a relatively low salinity of 34.77±O.88. It is
confined principally to the surface layer along
the Italian coast in correspondence to river
mouths and lagunal zones, and tends to flow
along the coast varying in width and position as
a function of the river flow inputs and the
mesoscale variability. Due to the low fresh water
discharge in the area (Po river discharge rate
-940 m 3 /s) during the period when data were
collected, this water is restricted to an area
extending -20 km out from the coast. The second duster comprises water with a temperature
similar to that of the preceding group
(21.02±O.89 0 C), but with a higher mean salinity:
36.34±O.63. It is found in the intermediate layer
of the coastal zone and reaches the surface out to
sea. Finally. the third duster is associated with
water from the bottom layer below the pycnocline having a relatively low temperature
(15.15±1.3 0 C) and high salinity (37.42±024).
The horizontal distributions of the thermohaline properties at the surface (Fig. 3a) show the
presence of various mesoscale cyclonic and anticyclonic structures, as also a strong coastal front
separating water of fluvial origin from that of the
open sea. In the southern part of the studied
autumn campaign along the ship track, with repetition of different transects, even when no CTD
sampling was done. The lines along which CUfrent data are available are depicted in Fig. 6. The
vertical resolution is of 2.5 m, while the horizontal one is of about 10-20 m,depending on the ship
speed. The instrument is not able to make measurements in the first 8 m near the surface and in
the layer near the bottom, corresponding to 15%
of the whole depth. The surface and bottom layer
currents have been extrapolated for each component, considering in the surface layer a constant
value equal to the first available datum, and in the
bottom layer, a logarithmic decay from the deepest datum to the zero bottom value. The accuracy
in the current measurement is of the order of one
cm/s, due also to the fact that the bottom tracking is always available and the speed of the ship
is, in this way, measured with the same accuracy
as the water velocities. The possible misalignment of the transducers with respect to the ship's
stern and eventual errors in the gyrocompass are
corrected in the first part of the processing. The
data have then been averaged horizontally every
500 m along each transect. Thereafter, the
method developed by Candela et al. (1992) was
applied to this data set in order to obtain the circulation of the Northern Adriatic. It permits to
separate three components in the signal: the
barotropic part of the tide, the "steady" and the
"residual" current. The steady component represents the time-independent part of the current
over the period considered, while the residual
one contains the remaining spatio-temporal variability. The method uses the least square procedure to fit horizontally the data to a function sum
of a non-depending time term and of a linear
combination of functions like sine and cosine of
the tidal frequency, whose coefficients are functions of the spatial co-ordinates of the point considered. As suggested by the author of the
method and already tested in a previous work
(Gacic et al. 1998), polynomials were the most
appropriate as the basis for the steady term and
for the coefficients of the tides. The residual term
presents a time variability on the order of a day
or greater, and thus may contain signals related to
wind effects or to discharge pulses of the rivers,
among which the Po plays the main role. In fact,
the river discharge is seen in the residual field
only in concomitance with a sudden variation in
the mean flow.
Hydrology of the Northern Adriatic
27
Results and Discussion
From its elongated shape, the T-5 diagram indicates a high variability mainly in temperature,
but also in salinity; in fact these parameters present values varying from 13.5 to 23°C and 30 to
38. Corresponding sigma-t variations are from
21 kglm 3 to 28.5 kglm 3 • The presence of three
main different water bodies can be observed: the
light waters, located to the left of the upper part
of the diagram are warmer (20-23°C) and less
salty (30-35). The dense waters instead, are located at the bottom and present lower values in
temperature but higher in salinity (around 14°C
and 37-38). The intermediate waters could be
recognised in the central part of the diagram,
being characterised mainly by two close branches. The application of cluster analysis provides
more information than the classical T -5 diagram
analysis, because it gives not only the characteristics of the possible water bodies presented in
the area, but also their spatial distribution. The
data have been separated into three dusters, each
one being associated with a water mass possessing distinct characteristics (Fig. 2). The first duster groups together water probably of a fluvial
origin with a temperature around 20.91±0.63°C,
and a relatively low salinity of 34.77±O.88. It is
confined principally to the surface layer along
the Italian coast in correspondence to river
mouths and lagunal zones, and tends to flow
along the coast varying in width and position as
a function of the river flow inputs and the
mesoscale variability. Due to the low fresh water
discharge in the area (Po river discharge rate
-940 m 3 /s) during the period when data were
collected, this water is restricted to an area
extending -20 km out from the coast. The second duster comprises water with a temperature
similar to that of the preceding group
(21.02±O.89 0 C), but with a higher mean salinity:
36.34±O.63. It is found in the intermediate layer
of the coastal zone and reaches the surface out to
sea. Finally. the third duster is associated with
water from the bottom layer below the pycnocline having a relatively low temperature
(15.15±1.3 0 C) and high salinity (37.42±024).
The horizontal distributions of the thermohaline properties at the surface (Fig. 3a) show the
presence of various mesoscale cyclonic and anticyclonic structures, as also a strong coastal front
separating water of fluvial origin from that of the
open sea. In the southern part of the studied