Worthington and Volkmann (1965) but an analysis of recent observations of nutrients and tracers
(U. Fleischmann et al., personal communication)
suggests otherwise.
Just west of 35°W the CGFZ overflow turns
northward and flows along the western margin of
the Reykjanes Ridge (Worthington and Volkmann,
1965). The transport has not been measured there
and, as was indicated in Section 5.6.1, the location
of the crossing of the overflow to the western
margin is unknown. Dickson and Brown (1994),
by virtue of the near doubling of transport in the
Denmark Srait overflow between the TTO and
Angmagssalik arrays (Table 5.6.3), argue that the
merging must occur between them, near 63.5°N.
Independent evidence for this conclusion, for
example in the form of a tracer signal, is strongly
desirable.
5.6.3.4 The Denmark Strait Overflow
Ross (1976) constructed a 36-day-long overflow
time series for a section 55 km south of the sill of
the Denmark Strait (not reproduced in Fig. 5.6.6).
This brief record showed no reversals, a mean of
2.5<0.5 Sv towards 250° (for water colder than
2°C), a standard deviation of 1.5 Sv and bursts of
1-day duration every few days. Remarkably, these
observations have not been repeated, although isolated year-long records nearby (Aagard and Malmberg, 1978) confirmed the persistence of the core
of the flow and demonstrated the absence of a seasonal signal – just as in the Faroe Bank Channel.
About 160 km downstream from this site was
the first of three arrays deployed by Dickson and
Brown (1994), which they named Dohrn Bank.
The transport record is 120 days in length, has a
mean of 4.25<0.3 Sv (for water colder than 3°C)
and a mean potential temperature of 1.22°C. The
flow shows no reversals but is energetic at periods
2–5 days (Fig. 5.6.7).
The TTO current meter site, 380 km downstream from the sill, was chosen to lie along a
hydrographic section conducted during the TTO
experiment. The transport record is 11 months in
length, has a mean of 5.7 <1.0 Sv (for water
colder than 3°C) and a mean potential temperature of 1.59°C. The large uncertainty in the mean
transport is due to loss of data, not only on a lost
mooring, which lay in the outer parts of strong
flow (see Fig. 5.6.4), but also incomplete records
from the adjacent inshore mooring. Figure 5.6.6
shows a period of very weak transport across the
1990–91 boundary, and surprisingly long period
variability (from 4 to 60 days, see Fig. 5.6.7). The
uncertainty in the mean transport makes model
validation (Section 5.6.5) an unreliable activity
here.
At a distance of about 550 km from the Denmark
Strait sill lies the Angmagssalik site at which three
year-long arrays were deployed. These have been
described in the section on re-analysis and the
value of the mean transport of 9.0 <0.5 Sv
reported. The mean potential temperature is
1.78°C and the periods of energetic variability
range from 4 to 60 days (Fig. 5.6.7). Monthly
averaged transports (not shown) have only a small
range, namely from 5 to 11 Sv, but in the last 4
months of 1988 and the first month of 1989 (days
600–750) some very low persistent values were
achieved. These were not reported by Dickson and
Brown. Before these measurements were made, a
conjecture that the overflow would show such
overall steadiness would have been met with derision. The absence of a seasonal signal was also
unexpected in light of the seasonal fluctuation seen
in the inflow of North Atlantic water hugging the
shelf edge east of the Faroe Islands (Gould et al.,
1985). However, neither the inflow on the North
side of the Faroe Islands nor the East Greenland
current have been observed sufficiently well to
exclude the possibility of seasonal compensation.
In both 1995 and 1997 moorings were deployed at
depths of 1950 m and 2325 m at the Angmagssalik
site (corresponding to distances of 50 and 80 km,
respectively, on Fig. 5.6.5). Dickson et al. (1999)
find that for the months of January and February
1997 the mean speed of the flow at the deeper
mooring is much weaker than measured at all previous months, although afterwards the speed recovers to more nearly normal values. However, from
these measurements alone, quantitative changes in
overflow transport cannot be derived.
Although the Angmagssalik and southeast
Iceland sites are the same distance downstream from
their respective sills, namely 500 km, their overflow
characteristics differ (Table 5.6.3). In the Iceland
basin the mean flow is weaker than in the Irminger
basin (3.2 Sv versus 9.0 Sv) and its variability is
less energetic (2.0 Sv
2 versus 9.0 Sv
2 ). Compared
with the transport at the sills, the flux has
increased by factors of 1.7 and 3.5, respectively,
and the core of the flow has descended 600 m
1
ᎏ
2
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
410
(U. Fleischmann et al., personal communication)
suggests otherwise.
Just west of 35°W the CGFZ overflow turns
northward and flows along the western margin of
the Reykjanes Ridge (Worthington and Volkmann,
1965). The transport has not been measured there
and, as was indicated in Section 5.6.1, the location
of the crossing of the overflow to the western
margin is unknown. Dickson and Brown (1994),
by virtue of the near doubling of transport in the
Denmark Srait overflow between the TTO and
Angmagssalik arrays (Table 5.6.3), argue that the
merging must occur between them, near 63.5°N.
Independent evidence for this conclusion, for
example in the form of a tracer signal, is strongly
desirable.
5.6.3.4 The Denmark Strait Overflow
Ross (1976) constructed a 36-day-long overflow
time series for a section 55 km south of the sill of
the Denmark Strait (not reproduced in Fig. 5.6.6).
This brief record showed no reversals, a mean of
2.5<0.5 Sv towards 250° (for water colder than
2°C), a standard deviation of 1.5 Sv and bursts of
1-day duration every few days. Remarkably, these
observations have not been repeated, although isolated year-long records nearby (Aagard and Malmberg, 1978) confirmed the persistence of the core
of the flow and demonstrated the absence of a seasonal signal – just as in the Faroe Bank Channel.
About 160 km downstream from this site was
the first of three arrays deployed by Dickson and
Brown (1994), which they named Dohrn Bank.
The transport record is 120 days in length, has a
mean of 4.25<0.3 Sv (for water colder than 3°C)
and a mean potential temperature of 1.22°C. The
flow shows no reversals but is energetic at periods
2–5 days (Fig. 5.6.7).
The TTO current meter site, 380 km downstream from the sill, was chosen to lie along a
hydrographic section conducted during the TTO
experiment. The transport record is 11 months in
length, has a mean of 5.7 <1.0 Sv (for water
colder than 3°C) and a mean potential temperature of 1.59°C. The large uncertainty in the mean
transport is due to loss of data, not only on a lost
mooring, which lay in the outer parts of strong
flow (see Fig. 5.6.4), but also incomplete records
from the adjacent inshore mooring. Figure 5.6.6
shows a period of very weak transport across the
1990–91 boundary, and surprisingly long period
variability (from 4 to 60 days, see Fig. 5.6.7). The
uncertainty in the mean transport makes model
validation (Section 5.6.5) an unreliable activity
here.
At a distance of about 550 km from the Denmark
Strait sill lies the Angmagssalik site at which three
year-long arrays were deployed. These have been
described in the section on re-analysis and the
value of the mean transport of 9.0 <0.5 Sv
reported. The mean potential temperature is
1.78°C and the periods of energetic variability
range from 4 to 60 days (Fig. 5.6.7). Monthly
averaged transports (not shown) have only a small
range, namely from 5 to 11 Sv, but in the last 4
months of 1988 and the first month of 1989 (days
600–750) some very low persistent values were
achieved. These were not reported by Dickson and
Brown. Before these measurements were made, a
conjecture that the overflow would show such
overall steadiness would have been met with derision. The absence of a seasonal signal was also
unexpected in light of the seasonal fluctuation seen
in the inflow of North Atlantic water hugging the
shelf edge east of the Faroe Islands (Gould et al.,
1985). However, neither the inflow on the North
side of the Faroe Islands nor the East Greenland
current have been observed sufficiently well to
exclude the possibility of seasonal compensation.
In both 1995 and 1997 moorings were deployed at
depths of 1950 m and 2325 m at the Angmagssalik
site (corresponding to distances of 50 and 80 km,
respectively, on Fig. 5.6.5). Dickson et al. (1999)
find that for the months of January and February
1997 the mean speed of the flow at the deeper
mooring is much weaker than measured at all previous months, although afterwards the speed recovers to more nearly normal values. However, from
these measurements alone, quantitative changes in
overflow transport cannot be derived.
Although the Angmagssalik and southeast
Iceland sites are the same distance downstream from
their respective sills, namely 500 km, their overflow
characteristics differ (Table 5.6.3). In the Iceland
basin the mean flow is weaker than in the Irminger
basin (3.2 Sv versus 9.0 Sv) and its variability is
less energetic (2.0 Sv
2 versus 9.0 Sv
2 ). Compared
with the transport at the sills, the flux has
increased by factors of 1.7 and 3.5, respectively,
and the core of the flow has descended 600 m
1
ᎏ
2
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
410