Skip to content

Tropical East–West Circulations

CSIR NET Part C meteorology: Walker circulation, velocity potential and divergent wind, equatorial waves, the Gill model, monsoon east–west outflow.

  • 18 min read
  • Updated
On this page · 15 sections
  1. Prerequisites
  2. The Hook: A Circulation Hidden in the Divergent Wind
  3. Why the Tropics Behave Differently
  4. Rotational and Divergent Wind
  5. The Walker Circulation
  6. Equatorial Waves: the Building Blocks
  7. The Gill Model
  8. The Northern-Summer East–West Circulations
  9. Memory Tricks
  10. Don’t Confuse
  11. Common Mistakes
  12. Practice
  13. Quick Revision Table
  14. Key Takeaways
  15. What’s Next

Prerequisites

The Hook: A Circulation Hidden in the Divergent Wind

Why the Tropics Behave Differently

PropertyMiddle latitudesTropics
Coriolis parameter \( f = 2\Omega\sin\phi \)about \( 10^{-4} \) s\(^{-1}\)small, zero on the equator
Rossby numbersmall: geostrophic balancelarge: geostrophy fails near the equator
Horizontal temperature gradientslarge (fronts, baroclinic zones)small: the weak temperature gradient (WTG) regime
Main balance in the thermodynamic equationhorizontal advection and vertical motiondiabatic heating balanced by adiabatic cooling of rising air
Main energy source of disturbancesbaroclinic conversionlatent heat release in convection

Weak temperature gradient balance. Because gravity waves rapidly spread any temperature anomaly over large distances when \( f \) is small, free-tropospheric temperatures in the tropics are nearly uniform horizontally. Heating by convection cannot raise the local temperature much; it is used instead to drive ascent:

$$ \frac{Q}{c_p} \approx w\,\frac{T}{\theta}\frac{\partial\theta}{\partial z} \quad\Longleftrightarrow\quad w \approx \frac{\dot\theta}{\partial\theta/\partial z} $$

So in the tropics heating sets vertical motion, and vertical motion requires divergence aloft and convergence below. That divergent flow is what builds the east–west cells.

Rotational and Divergent Wind

By Helmholtz’s theorem, any horizontal wind can be split into a non-divergent (rotational) part and an irrotational (divergent) part:

$$ \mathbf{V} = \mathbf{k}\times\nabla\psi + \nabla\chi, \qquad \nabla^2\psi = \zeta, \qquad \nabla^2\chi = D $$

with \( \psi \) the streamfunction, \( \chi \) the velocity potential, \( \zeta \) the relative vorticity and \( D \) the divergence.

QuantityWhat it showsUse
Streamfunction \( \psi \)rotational flow: highs, lows, jetsthe Tibetan anticyclone, subtropical highs, stationary waves
Velocity potential \( \chi \)divergent flow: where air spreads out or convergescentres of large-scale ascent and descent
Divergent wind \( \nabla\chi \)flows from minima to maxima of \( \chi \)the outflow branches of the east–west cells

With the convention \( \mathbf{V}_\chi = \nabla\chi \) used above, upper-level divergence (\( D > 0 \)) sits at a minimum of \( \chi \), and the divergent wind blows away from it. Some texts and data centres use \( \mathbf{V}_\chi = -\nabla\chi \), which flips the sign: check the convention before reading a map.

Because \( \chi \) is obtained by inverting a Laplacian, it is smooth and dominated by the largest scales, which is why 200 hPa \( \chi \) maps show the planetary-scale east–west circulations so clearly and are used to track the MJO.

The Walker Circulation

The Pacific cell

ElementNormal (non-El Niño) state
Sea-surface temperaturewarm pool in the western Pacific (often above 28–29 °C); cold tongue in the eastern equatorial Pacific
Surface windeasterly trades along the equator
Ascentover the Maritime Continent and western Pacific: deep convection, heavy rain
Upper-level flowwesterly return flow along the equator in the upper troposphere
Descentover the eastern Pacific: dry, with stratocumulus off South America
Oceanthermocline shallow in the east (upwelling), deep in the west; sea level higher in the west

The Bjerknes feedback links the ocean and the cell. Stronger trades push warm water west and raise the thermocline in the east, cooling the east; the stronger east–west SST gradient strengthens the trades. The positive feedback makes the coupled system unstable, and its oscillation is ENSO (next page).

Other east–west cells

In the annual mean the tropics have three main ascending regions, the Maritime Continent and western Pacific, tropical South America (Amazon) and equatorial Africa, separated by descending regions over the eastern Pacific, the Atlantic and the western Indian Ocean. Each pair forms an east–west cell.

OceanCellNotes
Pacificthe Walker cell properstrongest; tied to ENSO
Indianascent over the eastern Indian Ocean–Indonesia, descent over the western Indian Ocean–East Africaweaker; reverses partly during a positive IOD
Atlanticascent over the Amazon, descent over the eastern Atlanticinfluenced by Atlantic Niño and ENSO

Measuring the Walker circulation

  • Southern Oscillation Index: a standardised Tahiti-minus-Darwin sea-level pressure difference. Strong Walker circulation: high Tahiti pressure, low Darwin pressure, positive SOI (La Niña). Weak or shifted Walker: negative SOI (El Niño).
  • Equatorial SST gradient between the western and eastern Pacific.
  • Upper-level velocity potential and the zonal wind along the equator.

Under global warming

Models consistently project that the tropical overturning circulation weakens as the climate warms. The reasoning (Held and Soden, 2006): boundary-layer water vapour rises at roughly the Clausius–Clapeyron rate, about 7% per kelvin, while global precipitation, constrained by the energy budget, rises much more slowly, a few per cent per kelvin. If convective mass flux times moisture equals precipitation, the mass flux must fall. Observed trends in the Pacific Walker cell over recent decades have not clearly followed this projection, and natural variability is large; the projection concerns the long-term forced response.

Equatorial Waves: the Building Blocks

Linearised shallow-water equations on the equatorial \( \beta \)-plane (\( f = \beta y \), \( \beta = 2\Omega/a \approx 2.29\times 10^{-11} \) m\(^{-1}\) s\(^{-1}\)) give trapped waves (Matsuno, 1966). The trapping scale is the equatorial Rossby radius of deformation:

$$ \boxed{L_{eq} = \sqrt{\frac{c}{\beta}}} $$

where \( c = \sqrt{gh} \) is the gravity-wave speed for the equivalent depth \( h \). For an atmospheric first baroclinic mode with \( c \approx 50 \) m s\(^{-1}\), \( L_{eq} \approx 1{,}500 \) km (about 13° of latitude). For the ocean’s first baroclinic mode, \( c \approx 2.5 \) m s\(^{-1}\) gives about 330 km.

WavePropagationKey properties
Kelvineastward onlynon-dispersive, speed \( c \); no meridional wind; zonal wind and pressure in geostrophic balance across the equator
Equatorial Rossby (\( n = 1 \))westwardlong waves non-dispersive at speed \( c/3 \); twin vortices straddling the equator
Mixed Rossby–gravity (Yanai)westward phase, eastward group for some scalesantisymmetric about the equator; cross-equatorial meridional wind
Inertia–gravityeast or westhigh frequency

Only the Kelvin and Rossby waves matter for the steady response to heating.

The Gill Model

Gill (1980) solved the steady, linear, damped shallow-water equations on the equatorial \( \beta \)-plane for a fixed patch of heating (a mass sink in the lower layer, standing for convective heating). Damping by Rayleigh friction and Newtonian cooling at the same rate \( \varepsilon \) keeps the response finite.

Heating centred on the equator

graph LR
    W[West of heating: Rossby response] --- H[Heating on the equator]
    H --- E[East of heating: Kelvin response]
    W --> W1[Surface westerlies blowing into the heating]
    W --> W2[Twin surface cyclones straddling the equator]
    E --> E1[Surface easterlies blowing into the heating]
    E --> E2[Low pressure trough confined near the equator]
FeatureGill solution (lower troposphere)Upper troposphere
East of heatingKelvin wave region: easterly winds converging into the heat source; low pressure along the equatorwesterlies flowing away from the heat source
West of heatingRossby wave region: westerly winds into the heat source; a pair of cyclones either side of the equatoreasterlies; twin anticyclones
Over the heatingconvergence and ascentdivergence
Zonal extentKelvin response decays eastward over about \( c/\varepsilon \)Rossby response decays westward over about \( c/3\varepsilon \): three times shorter

The asymmetry comes from the speeds: Kelvin waves carry the signal east at \( c \), long Rossby waves carry it west at only \( c/3 \), so with the same damping the eastern response reaches three times farther.

Applied to the western Pacific warm pool, the Gill pattern gives the Walker cell: surface easterlies over the central and eastern Pacific converging into the warm pool, upper westerlies returning east.

Heating off the equator: the monsoon case

If the heating is centred well north of the equator, as in the Asian summer monsoon:

  • The Rossby response dominates and forms a surface cyclone to the north-west of the heating (the monsoon trough and heat low) with an anticyclone above it (the Tibetan anticyclone of the upper troposphere).
  • The Kelvin response is weaker and lies along the equator to the east.
  • Strong upper-level easterlies on the equatorward side of the upper anticyclone: the counterpart of the tropical easterly jet.
  • Low-level westerlies on the equatorward flank of the surface cyclone, feeding moisture into the heating: the counterpart of the southwest monsoon flow.

Rodwell and Hoskins (1996) used this framework to explain why deserts and Mediterranean-type summer climates lie to the west of monsoon heating. The Rossby wave response to the Asian monsoon heating, interacting with the westerlies on its western flank, produces adiabatic descent over the eastern Mediterranean, the Middle East and the eastern Sahara in summer. Monsoon heating in South Asia thus contributes to summer dryness thousands of kilometres to its west.

The Northern-Summer East–West Circulations

BranchWhereWhat it does
AscentSouth Asia, Bay of Bengal, Southeast Asia, western Pacificmonsoon convection; upper-level divergence (velocity-potential minimum)
Upper outflow westwardalong the tropical easterly jet, across India, the Arabian Sea and Africacarries mass west at 100–200 hPa
Descent (west)Arabian Sea, Middle East, North Africa, eastern Mediterraneandry subsidence; desert and Mediterranean summer
Upper outflow eastwardacross the Pacificlinks to the eastern Pacific descent
Descent (east)central and eastern Pacificstrengthens subtropical high and dry conditions
Low-level return flowwesterly monsoon flow over the Arabian Sea and Bay of Bengalsupplies moisture to the ascent region

In northern winter the main ascent centre moves to the Maritime Continent and northern Australia, and the Pacific Walker cell dominates.

How ENSO shifts the cells

During El Niño, warm water and convection move east to the central and eastern Pacific. The ascending branch of the Walker circulation shifts east, and anomalous descent develops over the Maritime Continent, and in many El Niño summers over South Asia as well. That anomalous descent is one route by which El Niño weakens the Indian monsoon. Studies of recent decades suggest that warming in the central Pacific is more effective in this respect than warming confined to the far eastern Pacific. During La Niña the Walker cell strengthens and convection over the Maritime Continent intensifies.

Memory Tricks

Don’t Confuse

TheseDifference
Hadley vs Walker circulationmeridional overturning (zonal mean) vs zonal overturning along the equator (zonal asymmetry)
Streamfunction vs velocity potentialrotational flow (\( \nabla^2\psi = \zeta \)) vs divergent flow (\( \nabla^2\chi = D \))
Kelvin vs equatorial Rossby waveeastward, speed \( c \), no meridional wind vs westward, speed \( c/3 \) for long waves, twin vortices
Atmospheric vs oceanic \( L_{eq} \)about 1,500 km (\( c \approx 50 \) m s\(^{-1}\)) vs about 330 km (\( c \approx 2.5 \) m s\(^{-1}\))
Tropical easterly jet vs subtropical westerly jetsummer, 100–150 hPa, over peninsular India vs winter, about 200 hPa, over north India
Walker circulation vs Southern Oscillationthe mean zonal cell vs its year-to-year see-saw of pressure

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Take \( \beta = 2.29\times 10^{-11} \) m\(^{-1}\) s\(^{-1}\) at the equator.

Level 1: Foundation

  1. Write the Helmholtz decomposition of the horizontal wind.
  2. In which direction do equatorial Kelvin waves propagate?
  3. Who named the Walker circulation?
Answers
  1. \( \mathbf V = \mathbf k\times\nabla\psi + \nabla\chi \), with \( \nabla^2\psi = \zeta \) and \( \nabla^2\chi = D \).
  2. Eastward only.
  3. Jacob Bjerknes (1969), after Gilbert Walker.

Level 2: Part B Style

Practice Q1. For an atmospheric first baroclinic mode with gravity-wave speed 50 m s\(^{-1}\), the equatorial Rossby radius of deformation is closest to:

(a) 330 km (b) 930 km (c) 1,500 km (d) 4,500 km

Answer: (c) about 1,500 km

\( L_{eq} = \sqrt{c/\beta} = \sqrt{50/2.29\times 10^{-11}} = \sqrt{2.18\times 10^{12}} \approx 1.48\times 10^{6} \) m. Option (a) is the oceanic value for \( c = 2.5 \) m s\(^{-1}\); (b) is for \( c = 20 \) m s\(^{-1}\).

Practice Q2. In the Gill model with heating symmetric about the equator, the low-level winds directly east and west of the heating are:

(a) easterly to the east, westerly to the west (b) westerly to the east, easterly to the west (c) easterly on both sides (d) westerly on both sides

Answer: (a)

Both flows converge into the heating. The Kelvin region east of the heating has easterlies; the Rossby region to the west has westerlies, with twin cyclones on either side of the equator.

Practice Q3. In a Gill-type model the damping time scale is 2.5 days and the gravity-wave speed 50 m s\(^{-1}\). The e-folding decay distances of the Kelvin response east of the heating and of the Rossby response west of it are about:

(a) 3,600 km and 10,800 km (b) 10,800 km and 3,600 km (c) 10,800 km for both (d) 1,500 km for both

Answer: (b)

Kelvin: \( c/\varepsilon = 50\times 2.5\times 86{,}400 = 1.08\times 10^{7} \) m \( = 10{,}800 \) km. Rossby: \( c/(3\varepsilon) = 3{,}600 \) km. The Kelvin response reaches three times farther.

Practice Q4. On a 200 hPa map, using the convention \( \mathbf V_\chi = \nabla\chi \), a strong minimum of velocity potential over the Bay of Bengal in July indicates:

(a) upper-level convergence and subsidence below (b) upper-level divergence and deep ascent below (c) a strong anticyclone, with no information about vertical motion (d) a cold-core low aloft

Answer: (b)

At a minimum of \( \chi \), \( \nabla^2\chi > 0 \), so \( D > 0 \): divergence aloft. By mass continuity that requires ascent below, here monsoon convection. (c) confuses \( \chi \) with \( \psi \): the anticyclone is in the streamfunction.

Level 3: Part C Style

Practice Q5. Over a tropical convective region, the column-mean diabatic heating is 2 K day\(^{-1}\) between 1000 hPa and 200 hPa, and the mean static stability is \( \partial\theta/\partial z = 3.5 \) K km\(^{-1}\). Assume weak temperature gradient balance and that the heating is all latent heat (\( L = 2.5\times 10^{6} \) J kg\(^{-1}\), \( c_p = 1004 \) J kg\(^{-1}\) K\(^{-1}\)). The mean vertical velocity and the implied rainfall rate are closest to:

(a) 0.66 cm s\(^{-1}\) and 6.5 mm day\(^{-1}\) (b) 6.6 cm s\(^{-1}\) and 65 mm day\(^{-1}\) (c) 0.66 cm s\(^{-1}\) and 65 mm day\(^{-1}\) (d) 0.066 cm s\(^{-1}\) and 0.65 mm day\(^{-1}\)

Answer: (a)

Vertical velocity: \( w \approx \dot\theta/(\partial\theta/\partial z) = (2/86{,}400)/0.0035 = 6.6\times 10^{-3} \) m s\(^{-1}\) \( = 0.66 \) cm s\(^{-1}\) (about 0.57 km per day).

Column heating: \( c_p\,\dot T\,\Delta p/g = 1004\times(2/86{,}400)\times 8\times 10^{4}/9.81 \approx 190 \) W m\(^{-2}\).

Rainfall: \( 190/2.5\times 10^{6} = 7.6\times 10^{-5} \) kg m\(^{-2}\) s\(^{-1}\) \( \approx 6.5 \) mm day\(^{-1}\).

Large-scale tropical ascent is only of order a centimetre per second, yet it sustains several millimetres of rain a day.

Practice Q6. In a given month the Tahiti sea-level pressure anomaly is −1.0 hPa, the Darwin anomaly is +1.2 hPa, and the standard deviation of the monthly Tahiti-minus-Darwin difference is 2.0 hPa. Which statements are correct?

  1. The standardised Tahiti-minus-Darwin index is about −1.1.
  2. The pattern indicates a weakened or eastward-shifted Pacific Walker circulation.
  3. Anomalous ascent over the Maritime Continent is the expected accompaniment.
  4. Such a month is consistent with El Niño conditions.

(a) 1, 2 and 4 only (b) 1 and 3 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4

Answer: (a) 1, 2 and 4 only
  • 1 true: \( (-1.0 - 1.2)/2.0 = -1.1 \).
  • 2 true: relatively high pressure in the west and low in the east weakens the east-to-west surface pressure gradient that drives the trades.
  • 3 false: high Darwin pressure goes with anomalous descent over the Maritime Continent as convection moves east.
  • 4 true: a strongly negative SOI is the atmospheric half of El Niño. (Operational SOI definitions differ in scaling; the sign is what matters here.)

Practice Q7. A simple model is forced by a patch of heating centred at 20° N, 90° E (summer monsoon). Which features would a Gill-type solution produce?

  1. A low-level cyclone to the north-west of the heating.
  2. An upper-level anticyclone above the low-level cyclone.
  3. Upper-level easterlies on the equatorward flank of the upper anticyclone.
  4. Descent to the west of the heating, over the Middle East and North Africa.

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4

Answer: (d)
  • 1 and 2 true: off-equatorial heating excites a dominant Rossby response, a baroclinic structure with a cyclone below and an anticyclone above, displaced to the north-west.
  • 3 true: easterlies on the southern flank of the upper anticyclone: the analogue of the tropical easterly jet.
  • 4 true: the Rossby wave response interacting with the westerlies produces descent to the west, the monsoon–desert mechanism of Rodwell and Hoskins.

Quick Revision Table

IdeaFact or formula
WTG balance\( w \approx \dot\theta/(\partial\theta/\partial z) \): heating drives ascent
Helmholtz\( \mathbf V = \mathbf k\times\nabla\psi + \nabla\chi \); \( \nabla^2\psi = \zeta \), \( \nabla^2\chi = D \)
Walker cellrise over warm pool, sink over cold tongue; surface easterlies, upper westerlies
Bjerknes feedbacktrades ↔ zonal SST gradient ↔ thermocline tilt
Other cellsIndian (weaker, IOD-linked), Atlantic (Amazon ascent)
Equatorial radius\( L_{eq} = \sqrt{c/\beta} \): about 1,500 km (air), about 330 km (ocean)
Equatorial wavesKelvin east at \( c \); long Rossby west at \( c/3 \)
Gill, equatorial heatingeasterlies east (Kelvin), westerlies and twin cyclones west (Rossby); Kelvin reaches 3× farther
Gill, off-equatorial heatingcyclone below, anticyclone above to the NW; TEJ analogue
Monsoon–desertRossby response gives descent to the west of monsoon heating
El Niñoascent shifts east; descent over the Maritime Continent and often South Asia
Warmingtropical overturning projected to weaken: water vapour rises faster than precipitation

Key Takeaways

What’s Next