432
last years several oceanographers have studied the changes in transatlantic hydrographic
sections over periods of the order of one decade (Levitus 1989a,b,c,1990, Greatbatch et
al. 1991, Greatbatch and Xu 1993, Parrilla et al. 1994). Using hydrographic data for the
pentads 1955-1959 and 1979-1974, Levitus (1989a,b,c) found that the structure of the
thermocline has experienced relatively large changes, ego potential density surfaces have
risen as much as 50 m in the N ASG and deepened as much as 100 m in the subartic
gyre, and Greatbatch et al. (1991) found a decrease of some 30 Sv in Gulf Stream transport. Greatbatch and Xu (1993) have found a substantial decrease in northward heat
transport during these pentads as compared with calculations using mean climatological
density data (l.0 PW and 0.8 PW as compared with l.2 PW, but with 0.3 PW error
bars), caused by increased southward velocities in the top 1000 m. Decadal variations in
heat transport have been modeled by Weaver and Sarachik (1991), Weaver at al. (1994),
Weisse et al. (1994). It is not yet clear whether these variations are the result of poor
temporal resolution in our data and errors in estimating the velocity fields, or if they are
real fluctuations of the coupled ocean-atmosphere systems.
Parrilla et al. (1994) compared three hydrographic sections taken along 24°N during
1957, 1981 and 1992, and found substantial warming in the permanent thermocline (approximately between 600 m and 2500 m), with a maximum warming rate of the order
of 1°C per century at about 1200 m. They point out that this warming rate is of the
order predicted in surface layers, by numerical models which study the response of the
ocean to changes in CO2 concentration (Mikolajewicz et al. 1990, Manabe et al. 1991).
The warming observed between 1957 and 1992 corresponds to an average rate of heat
change of 1 W 1m 2 during 35 years. They also found changes in salinity that essentially
maintained the T-S relationships, resulting in a deepening of the isopycnals by some 50
m in the 1000-1500 m range; these changes, however, are approximately uniform across
the section and produce small changes in meridional heat transport. The warming of the
thermocline layers may possibly be the result of heat that escapes, through subduction,
from its seasonal storage as WWM in the eastern NASG (Church et al. 1991, Bindoff and
McDougall 1994). The time scale of these changes agrees with the time scale of recirculation within the thermocline layers of the NASG, of order 10 years. The buoyancy flux,
which together with Ekman pumping drives the thermocline circulation, is very small
compared with the other terms involved in the heat balance, hence it is very difficult
to measure. Despite its relative smallness, however, it is likely to be a very important
mechanism for net annual heat retention and long-term climate control.
We may conclude that WWM formation is caused by three mechanisms of cold water
input to the surface layers of the eastern boundary, where net heat gain is large. The main
mechanisms are southward advection by the AC-CC and permanent coastal upwelling between Cape Vert and Cape Bojador. Heat gain associated to vertical advection in cyclonic
last years several oceanographers have studied the changes in transatlantic hydrographic
sections over periods of the order of one decade (Levitus 1989a,b,c,1990, Greatbatch et
al. 1991, Greatbatch and Xu 1993, Parrilla et al. 1994). Using hydrographic data for the
pentads 1955-1959 and 1979-1974, Levitus (1989a,b,c) found that the structure of the
thermocline has experienced relatively large changes, ego potential density surfaces have
risen as much as 50 m in the N ASG and deepened as much as 100 m in the subartic
gyre, and Greatbatch et al. (1991) found a decrease of some 30 Sv in Gulf Stream transport. Greatbatch and Xu (1993) have found a substantial decrease in northward heat
transport during these pentads as compared with calculations using mean climatological
density data (l.0 PW and 0.8 PW as compared with l.2 PW, but with 0.3 PW error
bars), caused by increased southward velocities in the top 1000 m. Decadal variations in
heat transport have been modeled by Weaver and Sarachik (1991), Weaver at al. (1994),
Weisse et al. (1994). It is not yet clear whether these variations are the result of poor
temporal resolution in our data and errors in estimating the velocity fields, or if they are
real fluctuations of the coupled ocean-atmosphere systems.
Parrilla et al. (1994) compared three hydrographic sections taken along 24°N during
1957, 1981 and 1992, and found substantial warming in the permanent thermocline (approximately between 600 m and 2500 m), with a maximum warming rate of the order
of 1°C per century at about 1200 m. They point out that this warming rate is of the
order predicted in surface layers, by numerical models which study the response of the
ocean to changes in CO2 concentration (Mikolajewicz et al. 1990, Manabe et al. 1991).
The warming observed between 1957 and 1992 corresponds to an average rate of heat
change of 1 W 1m 2 during 35 years. They also found changes in salinity that essentially
maintained the T-S relationships, resulting in a deepening of the isopycnals by some 50
m in the 1000-1500 m range; these changes, however, are approximately uniform across
the section and produce small changes in meridional heat transport. The warming of the
thermocline layers may possibly be the result of heat that escapes, through subduction,
from its seasonal storage as WWM in the eastern NASG (Church et al. 1991, Bindoff and
McDougall 1994). The time scale of these changes agrees with the time scale of recirculation within the thermocline layers of the NASG, of order 10 years. The buoyancy flux,
which together with Ekman pumping drives the thermocline circulation, is very small
compared with the other terms involved in the heat balance, hence it is very difficult
to measure. Despite its relative smallness, however, it is likely to be a very important
mechanism for net annual heat retention and long-term climate control.
We may conclude that WWM formation is caused by three mechanisms of cold water
input to the surface layers of the eastern boundary, where net heat gain is large. The main
mechanisms are southward advection by the AC-CC and permanent coastal upwelling between Cape Vert and Cape Bojador. Heat gain associated to vertical advection in cyclonic