181
surface salinity, and makes surface advective freshening in the convective
region more efficient. The periodic strengthening and weakening of convection caused by subsurface advective warming and surface freshening in
the subpolar region show up as interdecadal oscillations in the model.
This mechanism focuses on the instability induced by subsurface poleward heat transport through the northeastward currents associated with
deep water formation. Because the slow vertical heat transport is small,
enough heat accumulates below the surface and leads to instability, a way to
release heat, and this happens through convection in the model. An examination of water column T-S properties in regions where largest variations
occur show that the magnitude of temperature difference between surface
and subsurface increases over the whole oscillation period while salinity
differences quickly saturates after convection weakens or stops. This indicates that the oscillation is driven by a thermal instability since short time
scale haline process is unlikely the dominant process for the oscillation. In
order for an oscillation to occur, there has to be a process acting to restore
the system back. In the model, surface freshening process acts to weaken
or suppress convective process so that the heat can once again accumulate
in the subsurface.
What determines the period of these oscillations? For the advective
mechanism, Weaver and Sarachik (1991b) interpreted the period of the
oscillation as the time that an anomaly takes to travel a certain distance.
With both the advective and convective mechanisms acting, the period of
the oscillations can be thought as the time between two consecutive strong
phases of convection. This time scale is related to the rate of subsurface
warming and surface freshening. If everything else remains the same, it
would take less time for the water column to reach a convective critical
state when the rate of subsurface warming is larger, or the rate of surface
freshening is larger. Since surface freshening is a relatively short time
scale process, the oscillation period is mainly determined by the heating
rate. In the model, the dominant subsurface warming is due to horizontal
advection, which determines an interdecadal time scale (the mechanism is
summarized in Fig. 8).
One would still like to put these mechanisms into some conceptual
framework. We lander (1982, 1986) proposed simple flip-flop and loop thermohaline oscillation models. The basic idea of flip-flop oscillations is that
the system is unable to reach an equilibrium, so it has to oscillate between
equilibrium states. The flip-flop model analyzed by Welander (1982) gives
surface salinity, and makes surface advective freshening in the convective
region more efficient. The periodic strengthening and weakening of convection caused by subsurface advective warming and surface freshening in
the subpolar region show up as interdecadal oscillations in the model.
This mechanism focuses on the instability induced by subsurface poleward heat transport through the northeastward currents associated with
deep water formation. Because the slow vertical heat transport is small,
enough heat accumulates below the surface and leads to instability, a way to
release heat, and this happens through convection in the model. An examination of water column T-S properties in regions where largest variations
occur show that the magnitude of temperature difference between surface
and subsurface increases over the whole oscillation period while salinity
differences quickly saturates after convection weakens or stops. This indicates that the oscillation is driven by a thermal instability since short time
scale haline process is unlikely the dominant process for the oscillation. In
order for an oscillation to occur, there has to be a process acting to restore
the system back. In the model, surface freshening process acts to weaken
or suppress convective process so that the heat can once again accumulate
in the subsurface.
What determines the period of these oscillations? For the advective
mechanism, Weaver and Sarachik (1991b) interpreted the period of the
oscillation as the time that an anomaly takes to travel a certain distance.
With both the advective and convective mechanisms acting, the period of
the oscillations can be thought as the time between two consecutive strong
phases of convection. This time scale is related to the rate of subsurface
warming and surface freshening. If everything else remains the same, it
would take less time for the water column to reach a convective critical
state when the rate of subsurface warming is larger, or the rate of surface
freshening is larger. Since surface freshening is a relatively short time
scale process, the oscillation period is mainly determined by the heating
rate. In the model, the dominant subsurface warming is due to horizontal
advection, which determines an interdecadal time scale (the mechanism is
summarized in Fig. 8).
One would still like to put these mechanisms into some conceptual
framework. We lander (1982, 1986) proposed simple flip-flop and loop thermohaline oscillation models. The basic idea of flip-flop oscillations is that
the system is unable to reach an equilibrium, so it has to oscillate between
equilibrium states. The flip-flop model analyzed by Welander (1982) gives