189
is decadal according to baroclinic geostrophic adjustment scaling.
In classifying deep-decoupled millennial oscillations, one can use the
concepts of low order dynamic systems. Winton (1993) has constructed a
three box model to mimic these oscillations. For a weak fresh water input,
the model is at an equilibrium resembling the normal thermohaline circulation with steady convection at high latitudes. When the fresh water input
is large, a permanent halo cline prevents high latitude convection, and the
model is at an equilibrium corresponding a state of collapsed thermohaline
circulation. For an intermediate fresh water flux, the model oscillates, but
unable to settle into either of these two equlibria. The model also has
an important characteristics for the oscillations: the adjustment time for
the high/deep salinity gradient is shorter than that for the temperature.
As shown in the model, the period of these oscillations is determined by
vertical diffusive time, which is millennial.
Although the internal oscillations of the sector models are encouraging,
three limitations should be borne in mind. Two of these have to do with
the way the oscillations work - in particular with the way the halo cline is
broken down. This is accomplished by the mixing of heat that has entered
the basin diffusively during the deep-decoupled phase up to high-latitudes
beneath the halo cline as discussed above. Thus the oscillatory nature of
this model is enhanced by 1) the inhibition of alternative deepwater sources
that might import heat making it unavailable to the halo cline covered
region, and 2) a large value of the vertical diffusivity to quickly provide
heat before the halo cline becomes too strong to be broken down. The role of
the vertical diffusivity is particularly troublesome because large diffusivities
also favor thermally direct circulations requiring a larger freshwater forcing
to induce oscillations. The vertical diffusivity of 0.5 cm 2 / sec used here is
several times larger than the value thought to be appropriate for today's
ocean.
The third limitation has to do with the way the oscillations are induced: for a given vertical diffusivity and wind stress this is accomplished
by increasing the freshwater forcing. This would seem to call upon the
cold glacial climate to have a larger atmospheric and riverine meridional
transport of water than the warm interglacial climates that exhibit steady
thermohaline circulations. Equilibrium climate simulations with various
levels of CO2 by Manabe and Bryan (1985) indicate that the opposite is
true: The cold, low CO2 climates are accompanied by reduced mid-latitude
precipitation minus evaporation. By itself this would favor increased ther-
is decadal according to baroclinic geostrophic adjustment scaling.
In classifying deep-decoupled millennial oscillations, one can use the
concepts of low order dynamic systems. Winton (1993) has constructed a
three box model to mimic these oscillations. For a weak fresh water input,
the model is at an equilibrium resembling the normal thermohaline circulation with steady convection at high latitudes. When the fresh water input
is large, a permanent halo cline prevents high latitude convection, and the
model is at an equilibrium corresponding a state of collapsed thermohaline
circulation. For an intermediate fresh water flux, the model oscillates, but
unable to settle into either of these two equlibria. The model also has
an important characteristics for the oscillations: the adjustment time for
the high/deep salinity gradient is shorter than that for the temperature.
As shown in the model, the period of these oscillations is determined by
vertical diffusive time, which is millennial.
Although the internal oscillations of the sector models are encouraging,
three limitations should be borne in mind. Two of these have to do with
the way the oscillations work - in particular with the way the halo cline is
broken down. This is accomplished by the mixing of heat that has entered
the basin diffusively during the deep-decoupled phase up to high-latitudes
beneath the halo cline as discussed above. Thus the oscillatory nature of
this model is enhanced by 1) the inhibition of alternative deepwater sources
that might import heat making it unavailable to the halo cline covered
region, and 2) a large value of the vertical diffusivity to quickly provide
heat before the halo cline becomes too strong to be broken down. The role of
the vertical diffusivity is particularly troublesome because large diffusivities
also favor thermally direct circulations requiring a larger freshwater forcing
to induce oscillations. The vertical diffusivity of 0.5 cm 2 / sec used here is
several times larger than the value thought to be appropriate for today's
ocean.
The third limitation has to do with the way the oscillations are induced: for a given vertical diffusivity and wind stress this is accomplished
by increasing the freshwater forcing. This would seem to call upon the
cold glacial climate to have a larger atmospheric and riverine meridional
transport of water than the warm interglacial climates that exhibit steady
thermohaline circulations. Equilibrium climate simulations with various
levels of CO2 by Manabe and Bryan (1985) indicate that the opposite is
true: The cold, low CO2 climates are accompanied by reduced mid-latitude
precipitation minus evaporation. By itself this would favor increased ther-