In the ocean partially covered by sea ice, nearly uniform atmosphericcooling can induce much more intense cooling through an open waterarea (polynya or lead), and hence, more rapid brine rejection from iceformation than the ice-covered portion. This non-uniform negativebuoyancy flux produces a unique situation of convection: an area ofconvection is comparable with a convective plume. To investigaterelationship between effects on the dense water formation ofconvection and external parameters (the horizontal length scale of thebuoyancy forcing, the strength of the cooling, or rotation) in thissituation, numerical experiments of convection driven by horizontallyrestricted forcing (isolated convection) are carried out.
A three-dimensional ocean model is developed without a hydrostaticassumption. The model domain is a doubly periodic box 2560m square by200m deep. The spatial resolution is 10m in both the horizontal andvertical direction. The buoyancy forcing region has a disc shape andis set at the center of the model domain. In this region, buoyancyloss is introduced by subtracting a decrement of density from top gridthroughout the model integration. The initial density profile isneutral. Many experiments were conducted for the different sets ofexternal parameters.
From results of these experiments, the two dynamical regimes describedin Jacobs and Ivey (1999): the baroclinically unstable convection andthe baroclinically stable convection is confirmed. These two regimesare distinguished by the Rossby numberRoR=(B0/f3R2)¼.Here B0 is the buoyancy flux, f is theCoriolis parameter, R is the radius of the forcing disc andH is the total ocean depth. In the baroclinically stableconvection, it is found that the shift of the chimney is important,and the density scale of newly formed water has stronger dependencyagainst the radius of the forcing disc than the previous studies.
In the case where RoR is high and R/H is low,multiple convective plume dose not fully developed. And, the densityscale has nearly no dependency against the radius of the forcingdisc. That is considered as a new regime. The change of thedependency of density scale is explained by scaling analysis. The termbalance in the vertical plain vorticity equation is crucial to thischange.
The information about the density of newly formed water is useful forparameterizing dense water formation under the ice cover in anumerical model with a large (a few tens of km or larger) grid size. |
|
|