Overview of Tropical Meteorology and Climate

Variability

Contents
1 Introduction, Purpose, Scope and Definitions 1

2 Principal Governing Phenomena 2

3 Tropical Variability 9

4 MJO 10

5 Tropical Cyclones 18

6 ENSO 24

7 Decadal Variability 29

8 Monsoons 31

1 Introduction, Purpose, Scope and Definitions

The purpose of this note is to introduce the reader to the main features and
characteristics of the weather and climate in the tropical regions. Here we
define the tropics as more-or-less the region bounded by the Tropic of Cancer
to the north and Tropic of Capricorn to the south. This discussion will refer
mostly to atmospheric processes with some relevant points about the land
and ocean. As before, little mathematical detail will be provided (see refer-
ences), more emphasis is given to physical mechanisms and phenomonology.
Whereas there is no sharp delineation between the tropical region and
extra-tropical when considering thermodynamical processes, certain reason-
able approximations are often made in text books, such as the relative role

1
of the Corriolis effect and diabatic mechanisms, the reader should be aware
that the reality is rather more complex than such simple descriptions as given
here. I shall often make comparative statements to the mid-latitudes.
With a globe (or world altas in hand) it is instructive to familiarize one-
self with the distribution of land and ocean and the characteristics thereof.
Pay particular attention to the distribution of mountain ranges (Andes, Hi-
malaya), the deserts (Atacama, Sahara, Arabian, Kalahari, Gobi and Aus-
tralian). Also the ocean currents ( see for example my earlier post at
http://asl.umbc.edu/pub/chepplew/OceanTD_intro.html ), note the de-
tail of the zonal flow and the actual surface temperatures.

2 Principal Governing Phenomena

Starting with the obvious, as anyone who has spent time in the tropics
will know - the length of the day does not vary much throughout the year,
the (noon) solar zenith angle is close to 90-deg, and the amplitude of the
seasonal temperature cycle is generally (comparatively) small, although there
are important regional factors and micro-climates that make exceptions to
this observation. See figure 1 . Note, however that depending on location,
altitude etc, the amplitude of the diurnal temperature changes can be large,
compare for example a desert and a tropical rain forest climate).

Figure 1: Global map annual temperature cycle

2
 Next, based on a zonal average (key point) there is a net excess of heating
into the tropical ocean and atmosphere, which together with a net deficit
of heating at higher latitudes results in transport of heat from tropics to
the extra-tropics, see figure 2. A portion of the excess radiative heating
goes to warming the surface (land and ocean) and evaporation. This leads
(in general) to an abundance of atmospheric moisture (excluding the sub-
tropical high pressure belts), and so latent heat plays a significant part in
the energy cycles in the tropical atmosphere.

Figure 2: Zonal Average radiation budget

An important consequence of the excess heating at the tropics is the for-

mation of an overturning circulation which tends to form two cells (Hadley
cells) north and south of a relatively narrow central region of ascent charac-
terised by excess precipitation (the inter-tropical convergence zone, ITCZ).
The release of latent heat strengthens the uplift, and the circulation tightens
the convergence zone. The descending arms of the Hadley cells are associ-
ated with the sub-tropical high pressure belt and excess evaporation. These
are illustrated in figure 3 and figure 4 .
The tropical tropospheric winds tend to be dominated by vertical motion

3
 Figure 3: Global mean annual evaporation minus precipitation

and rather weak (on average) zonal winds of an easterly nature. See figure
5. Whereas in the mid-latitudes, a relatively strong pressure gradient can be
sustained with strong (geostrophic) winds, in the Tropics, the Coriolis effect
is very small, so when there is a pressure gradient the air moves more directly
from high to low pressure. Whereas this statement is correct, there can be
strong (horizontal) winds in the tropics associated with strong convective
systems (squalls and MCS) and the hurricanes (cyclones, typhoons), see
next section.
An important concept to appreciate and which is a consequence of the
average horizontal motions, is that (on average), since the zonal air flow
is slower than the surface of the Earth, momentum is transferred from the
Earth to the atmosphere in the Tropics, and the reverse true in mid-latitudes.
See figure 6.
Concerning the oceans, the structure of the Pacific, Atlantic and Indian
are of interest. In this short overview, we observe the typical equatorial
thermal cross-section of the Pacific as shown in figure 7. Note the surface
gradient from east to west near the surface, and the near-zero gradient below
about 300 m depth. The region of warmest surface waters tend to coincide
with the higher annual average precipitation, near Indonesia.
There are important details about the ocean circulation, depth of the
mixed layer and upwelling which are beyond the scope of this note. However,
roughly speaking the assymetry (warm pool in the west and cooler in the

4
 Figure 4: Global mean net diabatic heating

east Pacific) tends to drive a zonal circulation (ascending over the warm pool
descending over the east) called the Walker circulation, see figure 8. This
has important feedback effects which can cause the surface thermal structure
to change and lead to the ENSO, see below.
There are many other salient details of the atmosphere that are charac-
teristic of the tropical region and influence the thermodynamics and weather
patterns, which I shalln’t go into here.

5
Figure 5: Zonal wind cross section

6
 Figure 6: angular momentum

Figure 7: Pacific ocean thermal section

7
Figure 8: ENSO walker

8
3 Tropical Variability
Of particular interest to stratospheric processes, and only to mention here,
if one considers a slice at constant pressure through figure 5 at near 100
mb (and things get even more interesting at higher altitudes), one sees an
oscillation in the strength of the sub-tropical jets and the strength of the
tropical easterlies (a reversal), with a nearly 6-month period. This is the
quasi-biennial oscillation (QBO) of the tropical middle atmosphere. see fig-
ure 9. The QBO is known to affect the transport of aerosols in the lower
stratosphere, and has some influence on the strength of tropical cyclones,
through the change in vertical shear of the horizontal wind in the upper
troposphere, and possibly the monsoons (refs to be added later).

Figure 9: QBO zonal

If there were no land, and the equatorial plane were not inclined to the
plane of the ecliptic, then the ITCZ and the Hadley cells might be expected to
be rather zonally invariant. In reality the ITCZ tends to follow the annual
motion of the sun north and south of the equator. Figure 10 illustrates
roughly the most northerly and southerly movement of the ITCZ. Notice how
its migration is largest over the Indian Ocean and how restricted is its motion
south over the Pacific. The thermal effects of the Pacific and monsoon related
land/sea contrasts influence this. Indeed, close examination of precipitation
maps over the Pacific reveal a bi-modal (split) pattern extending East from
the warm pool.

9
 Figure 10: ITCZ annual

The ITCZ is, ofcourse, not a continuous, unbroken line around the Earth
and there are regions of enhanced convergence and preciptation over the
warmest waters and the tendency for convective systems to organize into
clusters. At any given portion of the ITCZ the phase velocity tends to be
westward. This is particularly the case later in the northern hemisphere
summer from North Africa westward across the Atlantic Ocean. It’s instruc-
tive to watch the movies at: http://goes.gsfc.nasa.gov. There also tends
to be enhanced precipitation over the tropical rainforest, as a result of the
recycling of moisture through evapotranspiration from plants.

4 Madden Julian Oscillation

A coherent structure of deep convective clouds, enhanced precipitation,
surface pressure and wind anomalies were observed in the 1970s with pe-
riod ranging from 30 to 60 days, propagating eastward across the Indian
Ocean and into the western Pacific. The characteristic wavelength is be-
tween 12,000 and 20,000 km. A great source for monitoring and news is:
http://www.bom.gov.au/climate/mjo/ There appears to be an ongoing
discussion about the detail on how the MJO forms and it’s evolution, but it
does have a baroclinic structure with upper tropospheric warming and low
level cooling during the westerly phase, with westerly anomalies in the lower
troposphere. This is consistent with the thermodynamics of deep convec-
tion. The maximum in precipitation occurs around the region of enhanced
low level convergence (just ahead of the max westerly anomaly). This zonal

10
flow anomaly can be seen in the context of a regionally modified Walker type
circulation. There is an associated feedback and response with surface ocean
temperature. The MJO tends to be at its strongest in the austral summer
(DJF) and in neutral ENSO and tends to be suppressed during either strong
El Niño or La Niña events. A schematic representation of the synoptic scale
flow and pressure structure are shown in figure 11.

Figure 11: MJO wind

And the structure can be tracked by observing the OLR as illustrated in

figure 12. There is much debate on the details of the MJO, for example in
repsect of the local feedback and impacts and dependencies upon the larger
scale motions, the phase of the ENSO and the monsoon circulations.
There has often been observed the tendency for regions of DCC and pre-
cipitation to be organised into clusters (MSC), in addition to the MJO, and
there are several theories as to how these form. These models are based
upon the notion of wave-like strctures forming in the tropical atmosphere,
the phase of the wave then resulting in an enhancement or reduction in
convergence or uplift, with resultant diabatic response (through latent heat
effects). There are eastward propogtaing Kelvin waves and westward prop-
agating Rossby waves. The Kelvin waves, see figure 13 are equatorially

11
trapped gravity waves which may be initiated by excess convection say over
Amazonia. The waves tend to propagate eastward and are convectively cou-
pled. We’ve come across Rossby waves as the dominant planetary wave
structure in the mid-latitudes. The restoring force is the gradient of po-
tential vorticity. The solutions to the linearized momentum equations on a
beta-plane for the equator provides solutions that are described as westward
propogating Rossby waves, see figure 14 . Note that one particular solution
provides for the possibility of mixed Rossby-gravity waves, see figure 15. In
the case of westward Rossby waves, the structure tends to be symmetric
about the equator, for mixed Rossby-gravity then tend to be anti-symmetric
about the equator. These waves are also convectively coupled (i.e. diabatic
heating influences its evolution) and can influence the mid-latitude atmo-
spheric circulation.
Intriguingly, and for those at the AGU 2014, it is postulated that an
eastward propogating MCS associated with the MJO, lead to a growing
Rossby wave-type structure that extended north from the tropical wave-
guide region, resulting in the so-called atmospheric river of moisture, see
figure 16. In my opinion it was necessary also to have the interaction with
a developing mid-latitude cyclone, or Rossby wave, at the same time and
place to advect the tropical air north eastward.

12
Figure 12: MJO OLR

13
Figure 13: Rossby Kelvin waves

14
Figure 14: Rossby waves

15
Figure 15: Mixed RG waves

16
Figure 16: pineapple

17
5 Tropical Cyclones
Otherwise known as typhoons (Japan and China), hurricanes (Atlantic),
these are the most energetic of tropical weather systems. They are most
prevelant and at their strongest in the western Pacific, they occur in the
northern Atlantic, with far fewer in the southern western Pacific and Indian
oceans. They form and grow most quickly when they are close to the equator
- but they never cross the equator. They either weaken and dissipate in situ
or become abosrbed in to the extra-tropical storm tracks and may transistion
into mature extra-tropical disturbances. See figure 17.

Figure 17: Global TC map

Tropical cyclones develop best if the SST to a depth of tens of meters > 26
C; if there is excess humidity in the low to mid-troposphere; the atmsophere
is conditionally (hydrostatically) unstable; there is weak vertical shear of the
horizontal winds; and it is at least 5 deg latitude away from the equator. To
a first approximation the cyclone is largely driven by conversion of latent
heat into kinetic energy.
The initial genesis of the cyclone depends upon local destabilization of
the ITCZ (through positive tendency of potential vorticity driven by the
overturning in convectove cells), or by quatorial trapped mixed Rossby grav-
ity waves (see above) - the proposed mechanism here is related to the spatial

18
phase of deepest convection relative to convergence (inflow) zone. Concern-
ing Atlantic hurricanes, the westward propogation of easterly waves from
north Africa over the warm waters sometimes leads to the generation of cy-
clones. Notwithstanding, these waves are often associated with MCSs in the
tropical Atlantic. If the wave remains coherent as it reaches the warmer
waters of the western tropical Atlantic and Gulf, they may have sufficient
energy available from the sea to grow. See figure 18 as an example of the
movement of an MCS complex as viewed from space and the development
into a hurricane.

Figure 18: MCS wave

The frequency of TCs can be measured in various ways, here we look at

the number of storms with winds greater than the given threshold averaged
over a hundred years and normalized. In figure 19, note how the storm
frequency varies between the ocean basins, that the Pacific has had a TC in
every month. Also shown are the average number of TC-days between the
two thresholds. The solid shading is for winds > 64 knts.
Within a given season, the likelihood of a TC developing (or at least

19
 Figure 19: TC annual cycle

gaining strength category 1 or higher) is influenced by the phase of the

MJO, more so in the Pacific basin. In general TCs are more likely in the
positive phase of the MJO (where convection is enhanced) and less likely - or
inhibited - during the negative phase of the MJO. As illustrated in figure 20.
The theory is that the deep convection of the +ve MJO maintains a reversal
of the PV gradient in the meridional direction (away from the equator),
which is conducive to TC spin up. There is also the possibility that if the
phase of the MJO and easterly waves coincide TC genesis is more likely.
In the Atlantic basin, there are occasions when low level air flows west
from the Sahara during the summer. This is a deep well mixed dry dusty
layer of air which mixes with humid maritime air. If this happens when an
MCS and easterly African wave form - there can be supression of convection

20
and decrease in the tendency for such distrubances to form TCs.
It is generally accepted that one of the major forcings of the modulation
of TC activity is the ENSO. During a positive (El Nino) phase, when deep
convection extends east into the central Pacific and extends the Indian ocean
zones, there tends to be increase vertical wind shear in the tropical western
Pacific and Atlantic basins. The TC genesis zones tend to move east further
into the central Pacific during a warm phase. Furthermore, due to increase
wind shear TC genesis in the Atlantic basin tends to suppressed. The oppo-
site tends to be true in the negative phase (La Nina). Figure 21 depicts the
actual record of Atlantic hurricanes during 14 El Nino cycles, according to
the phase of the year relative to the peak ENSO index.

21
Figure 20: TC MJO

22
Figure 21: ENSO hurricanes

23
6 El Nino Southern Oscillation
The El Nino Southern Oscillation (ENSO) of the Pacific basin is perhaps
one of the most iconic tropical climate phenomena that has global weather
consequences. As you’ll know, ENSO is characterized by a eastward shift in
the warm pool from Indonesian seas to the central Pacific and warming off
the west coast of contiental America. As the easterly trade winds weaken
the local sea surface currents change as does the Walker circulation. It has
a return period of 3 to 7 years. The southern oscillation (SOI) refers to the
swings in surface atmospheric pressure measured between Perth, and Tahiti.
The SOI is a proxy for the strength of trade winds, as pressure differences
determine wind speed. When there is a smaller pressure difference (low SOI)
there are El Niño conditions present; with larger differences (high SOI) and
La Niña conditions are present as illustrated in figure 22 .

Figure 22: SOI

I shalln’t attempt a description of the physical theory of ENSO, sufficient

to say that it appears to be a coupled atmosphere-ocean feedback effect due
to balance between the slope of the thermohaline and trade winds. The

24
’normal’ balance is illustrated in figure 23.

Figure 23: Pacific Walker

When the walker circulation is perturbed, westerly wind bursts occur

which generate downwelling Kelvin waves that propogate eastward, setting
up local upper ocean currents that advect the warmer water eastward. Figure
24 illustrates the time evolution of these effects. The key appears to be that
the westerly wind bursts need to be sustained and strong enough to generate
the Kelvin waves. At the eastern boundary, westward equatorially trapped
Rosby-type waves in the upper ocean acts to level out the thermocline (think
of this simply as a sloshing/seesaw of the upper ocean thermocline). Dis-
cussion continues about the role of excess heating in the warm pool and of
a "discharge - recharge" process resulting in Sverdrup transport below the
Ekman layer (refs).
There is much to learn about the details of precursurs, genesis meacha-
nisms and dissipation (re-normalization), but the point to note that ENSO
is not particular periodic and not particularly repeatable. The long-term
multivariate ENSO index from 1950 is shown in figure 25.

25
 Figure 24: Kelvin waves and ENSO

and comparison of six significant El Nino and la Nina events in figure

26 shows that the amplitude of the SST anomalies varies as does the areal
exent and ’center of gravity’, as shown in figure 27.
ENSO Impacts (TBD).

26
Figure 25: enso index history

Figure 26: Graph nino sst

27
Figure 27: Two strong Ninos

28
7 Decadal Variability
The Pacific Decadal Oscillation (PDO) is a 20 to 30 year oscillation of north
Pacific ocean and atmospheric anomalies. It tends to be correlated with the
ENSO in phase in the winter. The PDO signature in the tropics is secondary
to its signal in the North Pacific and North America. The PDO signal is
more distinct in the northern Pacific (cf PNA) than in the tropics, where
ENSO is the dominant signal.
Positive (negative) PDO values are usually associated with wetter (drier)
conditions in the Southwestern US. It also influences north Pacific fish popu-
lations, see figure 28. Causes of the PDO are not known, and much research
is on-going.

Figure 28: PDO

29
 The Atlantic Multi-decadal Oscillation (AMO) is a 20 to 40 year oscil-
lation of the north Atlantic ocean temperature. It appears to be related to
the strength of the over-turning circulation (AMOC) and perhaps also the
surface Gulf Stream. It appears when the AMOC weakens the north Atlantic
surface temperatures cool. Like the PDO the detailed mechanisms are not
known and much research is on-going. The AMO appears to influence the
strength and extent of the sub-tropical high pressure belts and mid-latitude
storm tracks. During a positive (warm) phase less rainfall is likely over north
America and more rainfall is likely over Europe. Some research suggests the
phase of the AMO influences the hurricane activity. Since the mid-1990s, the
AMO has been in a positive phase. Figure 29 shows the AMO SST signal
and correlation with gridded SST anomalies over the globe.

Figure 29: Upper: 10-year running mean detrended north Atlantic SST.
Lower: Correlation of AMO index to golbal SST anomaly. Bold line: zero.
Thin line: 95% significant.

30
8 Monsoons
The pattern and variability of rainfall in the tropical regions is a complicated
issue and varies depending on location, proximity to ocean etc. Many fac-
tors influence rainfall patterns including the diurnal cycle and annual cycles
(e.g. over tropical rainforests), the passage and strength of the MJO, the
ITCZ, tropical waves and larger scale effects due to the ENSO, PDO. In
some regions, such as those boardering deserts and mountain ranges, there
is a distinct wet and dry season. One of the seasonal patterns particularly
prevelant over the region including India to Northern Australia, and SW
USA, is the monsoon. The monsoon can be considered a modulation of the
ITCZ caused by the large thermal contrast between ocean and large land
mass compared to summer and winter seasons. It is important in describing
the physics of the monsoon to include local convective feedback, moisture
and rotation of the flow due to planetary motion. Recall, that the main fac-
tors required for precipitation are low level convergence (results in upward
flow) and moisture supply (latent heating). The areas of the world that are
considered to have a monsoon are illustrated in figure 30.

Figure 30: Global monsoons

As the northern summer season progreses, the continent warms up more

than the adjacent ocean, resulting in an increasing tendency for the air to rise
over the land and trade winds to shift from the ocean. When this coincides
with the arrival of the ITCZ trough then enhanced precipitation results.
This notion is illustrated in figure 31.
In the case of the Indian sub-continent, most precipitation occurs between

31
 Figure 31: monsoon circulation

May and October peaking around July, see figure 32

Considering the Indian and western Pacific sectors separately and follow-
ing the rainfall rates, you can clearly see the annual cycle, movement of the
ITCZ and enhancement of the monsoon, as shown in figure 33 In the lower
pane of this figure you can see that the northern territory of Australia is dry
when the Asian monsson is in progress.

32
Figure 32: seasonal rainfall

33
Figure 33: Asian monsoon

34
9 Conclusions and Points of interest
TBD

10 Acknowledgements
Numerous

11 References
Many

35