Geography Reference
In-Depth Information
circulation is required to balance the acceleration due to the momentum flux con-
vergence, which would otherwise increase the vertical shear of the mean zonal
wind and destroy the thermal wind balance.
Thus, the combined eddy heat flux and the eddy momentum flux distributions
tend to drive mean meridional cells in each hemisphere with rising motion pole-
ward of 45
◦
and sinking motion equatorward of 45
◦
. This eddy forcing more than
compensates the direct diabatic drive at midlatitudes and is responsible for the
observed thermally indirect
Ferrel cell
.
The resultant observed climatology of the Eulerian mean meridional circulation
is shown in Fig. 10.7. It consists primarily of tropical Hadley cells driven by
diabatic heating and eddy-driven midlatitude Ferrel cells. There are also minor
thermally direct cells at polar latitudes. The meridional circulation in the winter
is much stronger than that in the summer, especially in the Northern Hemisphere.
This reflects the seasonal variation both in the diabatic and in the eddy flux forcing
terms in (10.15).
The zonal momentum balance in the upper troposphere in tropical and midlat-
itude cells is maintained by the balance between the Coriolis force caused by the
mean meridional drift and the eddy momentum flux convergence. The heat balance
is maintained by rising motion (adiabatic cooling) balancing the diabatic heating
in the tropics and the eddy heat flux convergence at high latitudes, and subsidence
(adiabatic warming) balancing the eddy heat flux divergence in the subtropics.
Because of the appearance of eddy flux terms in both mean momentum and
thermodynamic energy equations, and the near cancellation of eddy and mean
flow processes, it is rather inefficient to attempt to diagnose the net eddy forcing of
the mean flow from the conventional Eulerian mean. It can be shown that similar
Streamfunction (units: 10
2
kg m
−
1
s
−
1
) for the observed Eulerian mean meridional circu-
lation for Northern Hemisphere winter, based on the data of Schubert et al. (1990).
Fig. 10.7
