80
D. Seidov . B. J. Haupt
tion and water motion inferred using two specially designed models; a threedimensional (3-D) sediment transport model and a semi-Lagrangian water
transport model which are used as add-ons to traditional ocean general circulation
models (OGCMs).
In the following sections we give an overview of the setup of past ocean seasurface boundary conditions, a brief description of the numerical models employed, a short review of the results of our recent regional modeling of the NA
circulation and sedimentation, and a display of some global conveyor simulations at two time slices, these being at and after the last glacial maximum.
2
Time Slices and Data
The global, and even regional basin-scale numerical modeling of ocean circulation necessitates a regular-grid coverage of the sea-surface with hydrological
data comprising sea-surface boundary conditions - a requirement not easily
met in paleoceanographic modeling. In fact only the glacial-to-interglacial cycle
of the last 20 000 years has the sea-surface data coverage that might be seriously
considered as suitable for ocean circulation simulations based on proxy data.
Moreover, only the last glacial maximum (LGM) has a global sea-surface temperature (SST) array compiled by CLIMAP (1981). This data set has recently
received much criticism because it is thought that the tropics are to warm
(Guilders on et al. 1994; Beck et al. 1997; Webb et al. 1997). However, this data
set remains the only global SST compilation currently available.
Thermohaline circulation cannot be properly modeled without knowing salinity distribution. Hence, sea-surface salinity (or equivalent freshwater fluxes
across the sea-surface needed to maintain the observed salinity), should be
known along with the SST to provide complete thermohaline sea-surface
boundary conditions for computing ocean circulation. Moreover, freshwater
discharges in the high latitudes of the North Atlantic may have enough power to
profound by affect the circulation in this basin and perhaps worldwide. The
freshwater fluxes are therefore thought to have beena major cause of past ocean
climate changes linked to the global salinity conveyor belt operation.
There was a substantial increase of salinity of the World Ocean at the LGM due
to large amount of freshwater deposited in the continental and shelf ice sheets.
Some authors assume values as high as 1 psu (Duplessy et al.1988; Fichefet et al.
1994). The removal of the freshwater stored in the ice sheets may be considered
as more or less a global feature over large areas because freshwater was mainly
evaporated from the ocean. Therefore the glacial increase of sea-surface salinity
(SSS) may be assumed, although to the first approximation only, as a uniform increase. On the contrary, the deglacial freshwater discharges were tied to high latitudes, mostly in the North Atlantic and perhaps also in the Southern Ocean (as
there are many indications of an interhemispheric synchrony of deglaciations
(e.g., Bard et al. 1997), and therefore cannot be inserted into a model as a homogeneous salinity decrease everywhere. Furthermore, the distribution and inten-
D. Seidov . B. J. Haupt
tion and water motion inferred using two specially designed models; a threedimensional (3-D) sediment transport model and a semi-Lagrangian water
transport model which are used as add-ons to traditional ocean general circulation
models (OGCMs).
In the following sections we give an overview of the setup of past ocean seasurface boundary conditions, a brief description of the numerical models employed, a short review of the results of our recent regional modeling of the NA
circulation and sedimentation, and a display of some global conveyor simulations at two time slices, these being at and after the last glacial maximum.
2
Time Slices and Data
The global, and even regional basin-scale numerical modeling of ocean circulation necessitates a regular-grid coverage of the sea-surface with hydrological
data comprising sea-surface boundary conditions - a requirement not easily
met in paleoceanographic modeling. In fact only the glacial-to-interglacial cycle
of the last 20 000 years has the sea-surface data coverage that might be seriously
considered as suitable for ocean circulation simulations based on proxy data.
Moreover, only the last glacial maximum (LGM) has a global sea-surface temperature (SST) array compiled by CLIMAP (1981). This data set has recently
received much criticism because it is thought that the tropics are to warm
(Guilders on et al. 1994; Beck et al. 1997; Webb et al. 1997). However, this data
set remains the only global SST compilation currently available.
Thermohaline circulation cannot be properly modeled without knowing salinity distribution. Hence, sea-surface salinity (or equivalent freshwater fluxes
across the sea-surface needed to maintain the observed salinity), should be
known along with the SST to provide complete thermohaline sea-surface
boundary conditions for computing ocean circulation. Moreover, freshwater
discharges in the high latitudes of the North Atlantic may have enough power to
profound by affect the circulation in this basin and perhaps worldwide. The
freshwater fluxes are therefore thought to have beena major cause of past ocean
climate changes linked to the global salinity conveyor belt operation.
There was a substantial increase of salinity of the World Ocean at the LGM due
to large amount of freshwater deposited in the continental and shelf ice sheets.
Some authors assume values as high as 1 psu (Duplessy et al.1988; Fichefet et al.
1994). The removal of the freshwater stored in the ice sheets may be considered
as more or less a global feature over large areas because freshwater was mainly
evaporated from the ocean. Therefore the glacial increase of sea-surface salinity
(SSS) may be assumed, although to the first approximation only, as a uniform increase. On the contrary, the deglacial freshwater discharges were tied to high latitudes, mostly in the North Atlantic and perhaps also in the Southern Ocean (as
there are many indications of an interhemispheric synchrony of deglaciations
(e.g., Bard et al. 1997), and therefore cannot be inserted into a model as a homogeneous salinity decrease everywhere. Furthermore, the distribution and inten-