confirms that this weakening convection was accompanied by a progressive, almost linear increase in the
temperature and salinity of the deepest layers to
unprecedented values by the mid-1990s, at which
time deepwater properties lay quite outside the
defined range for GSDW (compare Figure 7.3.8 c,d
with the GSDW properties as defined by Swift
et al., 1983; i.e. temperatures between 91.26 and
91.29°C and salinities between 34.889 and
34.892). Verduin and Quadfasel (1999) confirm that
this warming and salinification was not confined to
the deep layers but typified much of the water column of the Greenland Sea deeper than 1000 m.
There would be little debate that a part of
this change is due to dwindling convection. However, three supplementary mechanisms have been
advanced. First, there is the evidence, already mentioned, that reduced vertical exchange has been
replaced by increased deep horizontal exchange
through Fram Strait. Perhaps the best evidence
for this is the shift in GSDW characteristics away
from those of Greenland Sea Surface Water
(GSSW) and towards those of Eurasian Basin Deep
Water (Fig. 7.3.8a) reported by Boenisch et al.
(1997). This seems confirmed in the subtle but
steady decline in the oxygen content of GSDW as
deep temperatures have risen (Swift et al., 2001),
since this, too, is the expected net result of a weakening convective link with the surface and an
increasing exchange with the less-oxygenated deep
waters of the Arctic Ocean (Aagaard et al., 1991).
A second mechanism is thought to be the collapse
of the ‘domed’ density structure in the Greenland
Sea as a reduced wind stress curl (Jónsson, 1991)
supported a less intense cyclonic basin circulation
there (Rudels and Quadfasel, 1991; Meincke et al.,
1992). The evidence lies in the steady deepening of
intermediate and deep isopycnals in the Greenland
Sea since the early 1980s, reported by Boenisch et al.
(1997; Fig. 7.3.8b). Meincke and Rudels (1995) and
Meincke et al. (1999) combine the two mechanisms
by suggesting that the collapse of the dome might
have caused a compensating influx of Canadian
Basin Deep Water (CBDW) and Eurasian Basin
Deep Water (EBDW) into the centre of the Basin at
intermediate depths.
While accepting that cold conditions in the deep
Greenland sea were the result of deep convection,
Visbeck and Rhein (2000) point out that the subsequent deep warming has been accompanied by a
slow but significant increase of CFC concentrations
in the bottom layers that cannot be explained by
deep exchanges of ‘old’ water through Fram Strait.
They propose an alternative bottom water ventilation pathway to explain continued deep water
renewal in the absence of deep convection. By this
mechanism, ventilated intermediate water spreads
laterally to the basin margins where enhanced diapycnal mixing over the rough topography of Mohn’s
Ridge is responsible for transferring some portion of
its tracer signal and warmth to the deeper layers.
Whether, following these scenarios, strong vertical exchange in the Greenland Sea gave way to
enhanced lateral exchange with the Arctic Ocean,
or merely with its own boundaries, there seems to
be clear evidence of interbasin exchange with deep
waters of the Norwegian Sea. NSDW is composed
partly of GSDW (Swift and Koltermann, 1988)
and, until the early 1990s, had exhibited a lowamplitude and lagged version of the temperature
changes shown by GSDW itself (Østerhus and
Gammelsrod, 1999).There was also an apparent
lag with decreasing depth in the Norwegian Sea,
so that the recent abrupt warming appeared first
at 2000 m in 1987, where the greatest change in
the GSDW had been recorded, but not until 1990
at 1200 m. In the early 1990s, as GSDW production dwindled to near extinction, there are some
indications that NSDW ceased to warm further at
depth and that the connecting flow through the Jan
Mayen Channel may have temporarily reversed
(Østerhus and Gammelsrod, 1999 and personal
communication; see also Bourke et al., 1993).
Though their wider influence is certainly regulated by the gaps and passageways that form their
connections to neighbouring seas, the Nordic Seas
are potentially important as a source of change for
both the climatically sensitive Arctic Ocean and for
the northern overflows that form the headwaters of
the meridional overturning circulation (MOC). It is
worth recording, then, that over much of the water
column, the hydrographic character of these seas
during WOCE was beyond the range of our past
experience; in one way or another, these extreme
anomalies appear to have arisen through a changing balance, sense or pattern of ‘exchange’.
7.3.7 System-wide changes in the
Arctic Ocean
For some time now, we have known or supposed
that variability in the Arctic has some special role
7.3 The World during WOCE
571
Dickson et al.
temperature and salinity of the deepest layers to
unprecedented values by the mid-1990s, at which
time deepwater properties lay quite outside the
defined range for GSDW (compare Figure 7.3.8 c,d
with the GSDW properties as defined by Swift
et al., 1983; i.e. temperatures between 91.26 and
91.29°C and salinities between 34.889 and
34.892). Verduin and Quadfasel (1999) confirm that
this warming and salinification was not confined to
the deep layers but typified much of the water column of the Greenland Sea deeper than 1000 m.
There would be little debate that a part of
this change is due to dwindling convection. However, three supplementary mechanisms have been
advanced. First, there is the evidence, already mentioned, that reduced vertical exchange has been
replaced by increased deep horizontal exchange
through Fram Strait. Perhaps the best evidence
for this is the shift in GSDW characteristics away
from those of Greenland Sea Surface Water
(GSSW) and towards those of Eurasian Basin Deep
Water (Fig. 7.3.8a) reported by Boenisch et al.
(1997). This seems confirmed in the subtle but
steady decline in the oxygen content of GSDW as
deep temperatures have risen (Swift et al., 2001),
since this, too, is the expected net result of a weakening convective link with the surface and an
increasing exchange with the less-oxygenated deep
waters of the Arctic Ocean (Aagaard et al., 1991).
A second mechanism is thought to be the collapse
of the ‘domed’ density structure in the Greenland
Sea as a reduced wind stress curl (Jónsson, 1991)
supported a less intense cyclonic basin circulation
there (Rudels and Quadfasel, 1991; Meincke et al.,
1992). The evidence lies in the steady deepening of
intermediate and deep isopycnals in the Greenland
Sea since the early 1980s, reported by Boenisch et al.
(1997; Fig. 7.3.8b). Meincke and Rudels (1995) and
Meincke et al. (1999) combine the two mechanisms
by suggesting that the collapse of the dome might
have caused a compensating influx of Canadian
Basin Deep Water (CBDW) and Eurasian Basin
Deep Water (EBDW) into the centre of the Basin at
intermediate depths.
While accepting that cold conditions in the deep
Greenland sea were the result of deep convection,
Visbeck and Rhein (2000) point out that the subsequent deep warming has been accompanied by a
slow but significant increase of CFC concentrations
in the bottom layers that cannot be explained by
deep exchanges of ‘old’ water through Fram Strait.
They propose an alternative bottom water ventilation pathway to explain continued deep water
renewal in the absence of deep convection. By this
mechanism, ventilated intermediate water spreads
laterally to the basin margins where enhanced diapycnal mixing over the rough topography of Mohn’s
Ridge is responsible for transferring some portion of
its tracer signal and warmth to the deeper layers.
Whether, following these scenarios, strong vertical exchange in the Greenland Sea gave way to
enhanced lateral exchange with the Arctic Ocean,
or merely with its own boundaries, there seems to
be clear evidence of interbasin exchange with deep
waters of the Norwegian Sea. NSDW is composed
partly of GSDW (Swift and Koltermann, 1988)
and, until the early 1990s, had exhibited a lowamplitude and lagged version of the temperature
changes shown by GSDW itself (Østerhus and
Gammelsrod, 1999).There was also an apparent
lag with decreasing depth in the Norwegian Sea,
so that the recent abrupt warming appeared first
at 2000 m in 1987, where the greatest change in
the GSDW had been recorded, but not until 1990
at 1200 m. In the early 1990s, as GSDW production dwindled to near extinction, there are some
indications that NSDW ceased to warm further at
depth and that the connecting flow through the Jan
Mayen Channel may have temporarily reversed
(Østerhus and Gammelsrod, 1999 and personal
communication; see also Bourke et al., 1993).
Though their wider influence is certainly regulated by the gaps and passageways that form their
connections to neighbouring seas, the Nordic Seas
are potentially important as a source of change for
both the climatically sensitive Arctic Ocean and for
the northern overflows that form the headwaters of
the meridional overturning circulation (MOC). It is
worth recording, then, that over much of the water
column, the hydrographic character of these seas
during WOCE was beyond the range of our past
experience; in one way or another, these extreme
anomalies appear to have arisen through a changing balance, sense or pattern of ‘exchange’.
7.3.7 System-wide changes in the
Arctic Ocean
For some time now, we have known or supposed
that variability in the Arctic has some special role
7.3 The World during WOCE
571
Dickson et al.