The poleward transport of heat by atmospheric motions


In this lecture we consider only these most fundamental features of the earth's climate:
the ones that don't depend on season, on the distribution of land and sea, or on the
existence of mountain ranges.  They depend only on the fact that the equator-to-pole
contrast in incoming solar energy is stronger than the  equator-to-pole contrast in
outgoing infrared radiation (Fig. 5), and on the fact that the earth is rotating.
We will consider how land-sea geometry and mountains modify this basic picture
in the following lecture.

Fig. 4 Annual-mean incoming solar radiation and outgoing terrestrial radiation as
a function of latitude.  The latitude scale has been distorted so that equal areas on
earth occupy equal widths.
 

To deduce what the climate of such a simplified 'earth like' planet would look like
we rely on experiments that have been performed with numerical models of the
atmospheric component of the climate system.  When realistic land-sea geometry,
mountains, and seasonal changes in solar heating are prescribed in these models,
they yield quite realistic simulations of the observed climate.  Therefore the model
simulations of our idealized earth-like planet should have some credibility.

When such a model atmosphere is 'turned on', starting from a state of rest (no winds,
horizontally uniform temperatures) the tropics warm and the polar regions cool in
response to the imposed latitudinal contrast in the sun's heating, which is prescribed
in a way that mimics that in the real atmosphere.   Let's assume that when the heating i
s first turned in, the atmosphere is not rotating.   The ensuing sequence of events is
depicted schematically in Fig. 5.

Fig. 5   Schematic depiction of the development of the atmospheric circulation,
starting from a state of rest.
 

The equator-to-pole contrast in incoming solar radiation cause the tropics to
warm and the high latitudes to cool.  The temperature changes are largest in
the troposphere, which extends to about 10 km.  As the tropical troposphere
warms, thermal expansion causes the air at the 10 km level to bulge upward
relative to the cooling air at the higher latitudes, as represented by the heavy
arrows in panel (a) of Fig.5.  Hence, the mass of air weighing down on the
10 km level increases in the tropics and decreases in high latitudes.  The uneven
weight of the air gives rise to a pressure gradient, which pushes the air poleward,
as depicted in panel (a) of the figure.

The poleward  flow of air in the upper troposphere causes air to 'pile up' at
the higher latitudes.  The increasing weight of the overlying air column causes
sea level pressure to rise at high latitudes, while the sea-level pressure at low
latitudes decreases by an equal amount.  The equator-to-pole contrast in sea-level
pressure drives a compensating low level flow toward equator as depicted in
panel (b) of the figure.  Together, the poleward high level flow and the low level
equatorward flow make up a giant circulation cell extending all the way from
equator to pole, with rising motion in the tropics and inking at high latitudes,
as shown on the right hand side of panel (b).  We can imagine that the equatorial
belt, where the air in the cell is rising should be wet, and the polar regions, where
air is sinking should be dry.

Now let's turn on the rotation and see what happens.  Once the atmosphere starts
to rotate, The conservation of angular momentum begins to generate a component
of the flow that circulates around the earth's axis as shown in panel (c).  To understand
how this 'zonal' flow arises we can invoke the principle of the conservation of angular
momentum.  This conservation law tells us that if we view the motions in a fixed
frame of reference (i.e., a frame of reference that is not rotating with the planet-- like
observing a merry-go-round without being on it), the product of the zonal velocity
times the radius is conserved.  Hence,  objects moving toward the axis of rotation will t
end to rotate faster and faster, and vice versa, as illustrated in fig. 6. To demonstrate
this idea on a (small) merry-go-round,  position yourself on the outer edge of it and get
it spinning as fast as you can.  Then pull yourself inward toward the center and note how
the rotation rate increases.  The larger your mass in proportion to the mass of the
merry-go-round, the stronger the effect.  To apply the principle of the conservation of
angular momentum to a rotating atmosphere we simply have to note that the poleward
moving air at the 10-km level should begin to rotate faster than the planet, and this excess
rotation will be viewed by observers on the planet as a westerly wind component.  In a
similar manner, the low level equatorward return flow should rotate more slowly than
the planet, and this rotation deficiency should be evident to an observer on the planet as
an easterly wind component.

Fig. 6  Illustration of the conservation of angular momentum as applied to a rotating
object with movable weights (denoted by the black squares).
 

We can get the same result by invoking the Coriolis force, which has the following
properties:   (1) it is directly proportional to the wind speed on the rotation planet
(i.e., it affects only moving objects)  (2) it is proportional to rotation rate (i.e., to
have a Coriolis force, the planet has to be rotating  (3) its horizontal component,
which is what we care about, is proportional to the sine of the latitude (i.e., it is zero
on the equator and increases with latitude)  (4) it is directed to the left of the wind in
the Northern Hemisphere and to the right of the wind in the Southern Hemisphere.
We can think of the Coriolis force as deflecting the poleward flow at the 10-km level
toward the right in the Northern Hemisphere (left in the Southern Hemisphere to create a
westerly wind component and the easterly flow at the earth's surface toward the left in
the Northern Hemisphere (right in the Southern Hemisphere) to create an easterly
wind component.

As the heating in the tropics and the cooling at lower latitudes continue, the
equator-to-pole temperature contrast gradually increases, and so does the strength
of the upper level westerlies.  The resulting circulation in panel (c) resembles the
observed circulation in some ways (i.e., like the Earth it has tradewinds and en
equatorial rain belt, but it has no midlatitude westerlies and the deserts are in high
latitudes rather than in the subtropics.   Does the model have it wrong?  No, we just
have to be patient.

After a few weeks of simulated time, the behavior of the model suddenly changes.
At that point, the equator-to-poletemperature gradient reaches a critical value
which is strong enough to support the development of extratropical cyclones.
They develop as barely perceptible undulations in the westerlies, but within a
few days of simulated time they intensify to the point where they begin to change
the entire character of the flow by transporting heat, water vapor and angular
momentum poleward.  The specifics of how these changes come about is beyond t
he scope of these lecture notes, but the overall effects are:   (1) the extratropical
cyclones become the primary mechanism for the poleward transport of heat in
midlatitudes.  Once they develop they prevent the further buildup of the equator-to-pole
temperature gradient.  Whenever the gradient gets too strong, the storms intensify
and weaken it.  (2) the extratropical cyclones disturb the upper level flow in such a
way as to take angular momentum out of the tropics and deposit it in midlatitudes.
This poleward transport of angular momentum is responsible for the maintenance of t
he belt of westerly winds at the earth's surface in midlatitudes (3) the extratropical
cyclones are responsible for the sinking motions in the subtropics which favor desert
conditions and for a secondary maxima in rainfall around 55 N and 55 S shown in Figs. 2 and 7.

Fig. 7  Annual-mean incoming precipitation (P) and evaporation (E) as a function of
latitude.  The latitude scale has been distorted so that equal areas on earth occupy equal widths.
 

Panel (d) of Fig. 5 depicts the atmospheric circulation as modified by the extratropical
cyclones.  Note how the equator-to-pole circulation cell in the previous panel retreats
to the tropics.  It is known as the 'Hadley cell'.  The tradewinds undergo a similar shrinkage
as a result of the development of the extratropical cyclones.  The desert belts separate the t
rades from the midlatitude westerlies.  Note that Fig. 1 and Fig. 5d are virtually identical.