margins an enhanced diapycnic mixing in the
isopycnal model. But results from these upgrades
are not yet available.
One important result from several studies of the
North Atlantic has been the sensitivity of the
resulting meridional overturning to the details of
the overflow process (reviewed in Willebrand
et al., 2000; see also Böning and Semtner, Chapter
2.2). They show that the downstream development
of the overflow affects both the strength and structure of the meridional overturning. Realism in
modelling the overflow is thus a prerequisite for
credible climate modelling.
5.6.7 Overflow variability
Because the current meter records cover only a few
years, we must appeal to hydrographic observations for estimates of decadal variability. Bacon
(1998) has assembled and analysed section data
for the Cape Farewell data for the period 1955
to the present. Assuming the velocity field is
geostrophic and is zero at 1000 m (‘as other
authors have done’) the baroclinic transport for
densities 0 927.8 kg m
93 has been calculated. The
range of values is from about 2.5 to 8.5 Sv southward and is shown to be predominantly of low
frequency (groups of closely spaced sections have
small variance compared with that of the 42-year
record). Bacon argues that while the Clarke value
of 13 Sv measured in 1978 (discussed in Section
5.6.3) contains a barotropic component of 5 Sv,
his own observations made in 1991, for which he
estimated a combined barotropic and baroclinic
estimate of 6 Sv, is quite inconsistent with such a
large barotropic value. He concludes that on
decadal time scales the overflow is variable, with
low values in the 1950s and 1960s, high in the
1970s and 1980s and low in the 1990s. Because
the causes of the barotropic velocity components
(those hidden to hydrography) are unknown,
Bacon’s assertions are not universally accepted and
their confirmation will perhaps only be demonstrated by a repetition of some of the transport
measurements described in Section 5.6.3, or by
coupled or ocean models with credible time-dependent overflow components.
On very much longer time scales fluctuations in
the overflow has been investigated using palaeographic data. The sedimentary drifts that are a
conspicuous feature of the Iceland basin (three are
sketched in Fig. 5.6.4) have been utilized for this
purpose (Bianchi and McCave, 1999). These
authors have constructed a time series of sediment
grain size on the Gardar drift at 56°22ЈN and
27°49ЈW from the termination of the last glaciation (13 000–11 500 years BP) to the present. They
argue that the sortable silt size (in the range
10–63 m diameter) derived from a core is a qualitative measure of the near-bottom current speed
and, on the assumption that the Iceland–Scotland
overflow path remains close to the site sampled
throughout their record, derive variations in the
overflow strength. They find fluctuations that correspond to the ‘little ice age’ (ADϳ1550) when the
flow slowed, and ‘the medieval warm period’
(ADϳ900–1250) when the flow speed intensified,
and find the principal variability at a period of
1500 years. The information palaeographic data
provides cannot be obtained by any other method
and the verification of the conclusions derived
from it can probably only be established and
quantified by climate theory or climate modelling.
The concept of multiple equilibria for the global
thermohaline circulation was first proposed by
Broecker et al. (1985a) although Stommel (1961)
had demonstrated such behaviour in a simple
resevoir model. It was argued that sufficient freshening of the waters in the Nordic seas would prevent the production of dense water and shut off
the saline upper inflow, leaving dense water production to the Antarctic region and/or the Pacific.
Such a phenomenon occurred during the last
glaciation when the overflows were believed
shifted south, weakened or entirely absent. Such a
change in the nature and strength of the meridional overturning in the North Atlantic appears to
be confirmed in the pioneer modelling work of
Bryan (1986) and Manabe and Stouffer (1988).
5.6.8 What have we learnt in WOCE?
In project WOCE the sustained experimental study
of the northern North Atlantic (the subpolar gyre)
was in 1997–98 and most of the in-situ observations, hydrography and float tracks, have yet to be
reported. Via inverse analysis or via model assimilation the strength of the overturning circulation
and in particular the contribution from the overflow may thus be re-assesed. Much of the science
on which this review was based thus predated the
WOCE programme and in particular on the work
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
416
isopycnal model. But results from these upgrades
are not yet available.
One important result from several studies of the
North Atlantic has been the sensitivity of the
resulting meridional overturning to the details of
the overflow process (reviewed in Willebrand
et al., 2000; see also Böning and Semtner, Chapter
2.2). They show that the downstream development
of the overflow affects both the strength and structure of the meridional overturning. Realism in
modelling the overflow is thus a prerequisite for
credible climate modelling.
5.6.7 Overflow variability
Because the current meter records cover only a few
years, we must appeal to hydrographic observations for estimates of decadal variability. Bacon
(1998) has assembled and analysed section data
for the Cape Farewell data for the period 1955
to the present. Assuming the velocity field is
geostrophic and is zero at 1000 m (‘as other
authors have done’) the baroclinic transport for
densities 0 927.8 kg m
93 has been calculated. The
range of values is from about 2.5 to 8.5 Sv southward and is shown to be predominantly of low
frequency (groups of closely spaced sections have
small variance compared with that of the 42-year
record). Bacon argues that while the Clarke value
of 13 Sv measured in 1978 (discussed in Section
5.6.3) contains a barotropic component of 5 Sv,
his own observations made in 1991, for which he
estimated a combined barotropic and baroclinic
estimate of 6 Sv, is quite inconsistent with such a
large barotropic value. He concludes that on
decadal time scales the overflow is variable, with
low values in the 1950s and 1960s, high in the
1970s and 1980s and low in the 1990s. Because
the causes of the barotropic velocity components
(those hidden to hydrography) are unknown,
Bacon’s assertions are not universally accepted and
their confirmation will perhaps only be demonstrated by a repetition of some of the transport
measurements described in Section 5.6.3, or by
coupled or ocean models with credible time-dependent overflow components.
On very much longer time scales fluctuations in
the overflow has been investigated using palaeographic data. The sedimentary drifts that are a
conspicuous feature of the Iceland basin (three are
sketched in Fig. 5.6.4) have been utilized for this
purpose (Bianchi and McCave, 1999). These
authors have constructed a time series of sediment
grain size on the Gardar drift at 56°22ЈN and
27°49ЈW from the termination of the last glaciation (13 000–11 500 years BP) to the present. They
argue that the sortable silt size (in the range
10–63 m diameter) derived from a core is a qualitative measure of the near-bottom current speed
and, on the assumption that the Iceland–Scotland
overflow path remains close to the site sampled
throughout their record, derive variations in the
overflow strength. They find fluctuations that correspond to the ‘little ice age’ (ADϳ1550) when the
flow slowed, and ‘the medieval warm period’
(ADϳ900–1250) when the flow speed intensified,
and find the principal variability at a period of
1500 years. The information palaeographic data
provides cannot be obtained by any other method
and the verification of the conclusions derived
from it can probably only be established and
quantified by climate theory or climate modelling.
The concept of multiple equilibria for the global
thermohaline circulation was first proposed by
Broecker et al. (1985a) although Stommel (1961)
had demonstrated such behaviour in a simple
resevoir model. It was argued that sufficient freshening of the waters in the Nordic seas would prevent the production of dense water and shut off
the saline upper inflow, leaving dense water production to the Antarctic region and/or the Pacific.
Such a phenomenon occurred during the last
glaciation when the overflows were believed
shifted south, weakened or entirely absent. Such a
change in the nature and strength of the meridional overturning in the North Atlantic appears to
be confirmed in the pioneer modelling work of
Bryan (1986) and Manabe and Stouffer (1988).
5.6.8 What have we learnt in WOCE?
In project WOCE the sustained experimental study
of the northern North Atlantic (the subpolar gyre)
was in 1997–98 and most of the in-situ observations, hydrography and float tracks, have yet to be
reported. Via inverse analysis or via model assimilation the strength of the overturning circulation
and in particular the contribution from the overflow may thus be re-assesed. Much of the science
on which this review was based thus predated the
WOCE programme and in particular on the work
SECTION 5 FORMATION AND TRANSPORT OF WATER MASSES
416