Skip to content

Ocean Circulation for Part C: Ekman Spiral, Upwelling, ENSO, IOD and Thermohaline Flow

CSIR NET Earth Sciences Part C: Ekman spiral and depth, upwelling and Ekman pumping, ENSO, Indian Ocean Dipole, thermohaline circulation.

  • 21 min read
  • Updated
On this page · 14 sections
  1. Prerequisites
  2. The Hook: A Warm Current at Christmas
  3. The Ekman Spiral
  4. Upwelling and Ekman Pumping
  5. El Niño–Southern Oscillation (ENSO)
  6. The Indian Ocean Dipole (IOD)
  7. Thermohaline Circulation
  8. Memory Tricks
  9. Don’t Confuse
  10. Common Mistakes
  11. Practice
  12. Quick Revision Table
  13. Key Takeaways
  14. What’s Next

Prerequisites

The Hook: A Warm Current at Christmas

The Ekman Spiral

Assumptions and result

Ekman (1905) solved for steady flow driven by a uniform wind over an infinitely deep, homogeneous ocean with constant vertical eddy viscosity \( A_z \), far from coasts, on an \( f \)-plane (\( f = 2\Omega\sin\phi \)). The balance is Coriolis against vertical friction:

$$-fv = A_z\frac{\partial^2 u}{\partial z^2}, \qquad fu = A_z\frac{\partial^2 v}{\partial z^2}$$

For a wind blowing toward the north (along \( +y \)) in the northern hemisphere, with \( z \) negative downward:

$$u = V_0\, e^{\pi z/D_E}\cos\left(\frac{\pi}{4} + \frac{\pi z}{D_E}\right), \qquad v = V_0\, e^{\pi z/D_E}\sin\left(\frac{\pi}{4} + \frac{\pi z}{D_E}\right)$$$$\boxed{D_E = \pi\sqrt{\frac{2A_z}{|f|}}} \qquad \boxed{V_0 = \frac{\sqrt{2}\,\pi\,\tau}{D_E\,\rho\,|f|}}$$

What this says:

PropertyValue
Surface current direction45° to the right of the wind (NH); 45° to the left (SH)
With depththe current turns further to the right and decays exponentially
At \( z = -D_E \) (the Ekman depth)current is opposite to the surface current and reduced to \( e^{-\pi} \approx 4\% \) of it
Hodographa spiral: the Ekman spiral
Net (depth-integrated) transport90° to the right of the wind (NH), magnitude \( \tau/(\rho f) \) per unit width, independent of \( A_z \)

The spiral shape depends on the assumed constant \( A_z \); real observations show a flatter spiral and a surface deflection less than 45°. The 90° net transport does not depend on the viscosity and is well confirmed.

Ekman also gave empirical forms from Nansen’s observations, for wind speed \( U_{10} \) (m s\(^{-1}\)) at latitude \( \phi \):

$$V_0 \approx \frac{0.0127\,U_{10}}{\sqrt{\sin|\phi|}}, \qquad D_E \approx \frac{7.6\,U_{10}}{\sqrt{\sin|\phi|}}\ \text{(m)}$$

Wind stress itself comes from the bulk formula \( \tau = \rho_{air}C_D U_{10}^2 \), with \( \rho_{air} \approx 1.2 \) kg m\(^{-3}\) and drag coefficient \( C_D \approx 1.2 \)–\( 1.5\times 10^{-3} \).

Upwelling and Ekman Pumping

Coastal upwelling

A wind blowing along a coast with the coast on its left (NH) drives Ekman transport offshore. Water must come from below to replace it. Upwelling is confined to a band roughly the width of the baroclinic Rossby radius of deformation (about 10–50 km), so

$$\boxed{w \approx \frac{U_E}{L}} \qquad U_E = \frac{\tau}{\rho f}$$
RegionWind and settingNotes
Eastern boundary upwelling systems: California, Humboldt (Peru–Chile), Canary, Benguelaequatorward winds along west coasts of continentstogether a small fraction of ocean area but a very large share of the world fish catch
Somalia and Oman (Arabian Sea)southwest monsoon (Findlater jet) blowing parallel to the coast, June–Septemberstrongest seasonal upwelling; cold SST off Somalia in summer
Southwest coast of Indiasummer monsoon winds; also remote forcing by coastal Kelvin wavesupwelling off Kerala in the southwest monsoon; supports the oil sardine fishery; mud banks form in the same season
Equatorialeasterly trades; Ekman transport poleward on both sides because \( f \) changes signcold tongue in the eastern Pacific and Atlantic; equatorial divergence
Open-ocean (Ekman pumping)cyclonic wind-stress curlsubpolar gyres, cyclonic eddies, beneath tropical cyclones

Ekman pumping: where Ekman transport converges or diverges because the wind varies in space,

$$w_E = \frac{1}{\rho}\,\hat{k}\cdot\nabla\times\left(\frac{\boldsymbol{\tau}}{f}\right)$$

Cyclonic (anticlockwise in the NH) wind-stress curl drives upwelling (Ekman suction); anticyclonic curl drives downwelling. Subtropical gyres are regions of Ekman downwelling; their thermocline is deep and their surface waters are nutrient-poor “ocean deserts”.

El Niño–Southern Oscillation (ENSO)

The normal state

graph LR
    A[Easterly trade winds along the equator] --> B[Warm water piled up in the west Pacific warm pool]
    B --> C[Deep thermocline in the west, shallow in the east]
    C --> D[Upwelling of cold water in the east: cold tongue]
    D --> E[East-west SST gradient]
    E --> F[Walker circulation: rising air over the warm pool, sinking in the east]
    F --> A
  • West Pacific warm pool: SST above about 28 °C; deep convection and heavy rain over Indonesia.
  • East Pacific: cold tongue, dry air, high surface pressure.
  • Sea level is higher in the west than in the east (by about half a metre), and the thermocline slopes up toward the east.

This is a positive feedback loop: stronger trades → more upwelling → colder east → larger SST gradient → stronger trades. Bjerknes identified it, and it bears his name: the Bjerknes feedback.

El Niño and La Niña

El Niño (warm phase)La Niña (cold phase)
Trade windsweaken; westerly wind bursts in the west Pacificstrengthen
Warm pool and convectionshift east to the central or eastern Pacificconfined farther west
Thermoclinedeepens in the east, shoals in the west (flattens)steeper than normal
Eastern Pacific SSTanomalously warm; upwelled water is warm, nutrient-pooranomalously cold
Peru fisheriescollapsestrong
Rainfallheavy on Peru and Ecuador coast; drought in Indonesia and Australiawetter than normal in Indonesia and Australia
Southern Oscillationnegative SOI: pressure falls at Tahiti, rises at Darwinpositive SOI
Indian summer monsoontends to be deficienttends to be normal or above

Measuring ENSO

IndexDefinitionEl Niño sign
Southern Oscillation Index (SOI)standardised difference of sea-level pressure, Tahiti minus Darwinpersistently negative
Niño 3.4 SST anomalySST anomaly averaged over 5°N–5°S, 170°W–120°Wpositive
Oceanic Niño Index (ONI), used by NOAA3-month running mean of the Niño 3.4 anomaly; an El Niño episode needs ONI ≥ +0.5 °C for at least five consecutive overlapping seasons (La Niña: ≤ −0.5 °C)≥ +0.5 °C

Events recur irregularly every 2–7 years and usually peak in the northern winter (November–January). The strong events of 1982–83, 1997–98 and 2015–16 are the reference cases.

Why ENSO oscillates

The ocean adjusts through equatorial waves that travel along the equatorial waveguide:

WaveDirectionSpeedRole
Equatorial Kelvin waveeastward only\( c = \sqrt{g'H} \), about 2–3 m s\(^{-1}\) for the first baroclinic modecarries a deepened thermocline (downwelling) east in about 2–3 months
Equatorial Rossby wave (first meridional mode)westwardabout \( c/3 \)carries off-equatorial thermocline anomalies west; reflects at the western boundary as Kelvin waves

Here \( g' = g\,\Delta\rho/\rho \) is the reduced gravity across the thermocline and \( H \) is the upper-layer depth.

  • Delayed oscillator: a westerly wind anomaly sends a downwelling Kelvin wave east (warming the east, reinforcing El Niño) and upwelling Rossby waves west; these reflect at the western boundary into upwelling Kelvin waves that return east months later and terminate the event.
  • Recharge oscillator (Jin): warm water volume along the equator builds up before El Niño (recharge) and is drained poleward during it (discharge), so the system swings back.

Flavours of El Niño

  • Eastern Pacific (canonical) El Niño: warmest anomalies near South America (1997–98).
  • Central Pacific El Niño (El Niño Modoki): warm anomaly in the central Pacific flanked by cooler water to the east and west, described in the 2000s by a group including Karumuri Ashok. Its teleconnections differ from the canonical type.

ENSO and the Indian monsoon

El Niño years tend to bring deficient Indian summer monsoon rainfall, through the eastward shift of the Walker circulation and subsidence over the Indian region. The link is strong but not deterministic: 1997–98, one of the strongest El Niños on record, came with a near-normal monsoon, which work from the early 2000s attributed partly to a strong positive IOD that year.

The Indian Ocean Dipole (IOD)

Described in 1999 (Saji and colleagues; and Webster and colleagues the same year), the IOD is an ocean–atmosphere mode of the tropical Indian Ocean.

$$\boxed{\text{DMI} = \text{SSTA}_{west} - \text{SSTA}_{southeast}}$$

with the western box at 50°E–70°E, 10°S–10°N and the south-eastern box at 90°E–110°E, 10°S–0° (off Sumatra and Java).

Positive IODNegative IOD
SSTcooler off Sumatra–Java (upwelling), warmer in the westwarmer in the east, cooler in the west
Winds along the equatoranomalous easterlies; south-easterlies along Sumatra–Javaanomalous westerlies
Thermoclineshoals in the eastdeepens in the east
Rainfallmore over East Africa; drought in Indonesia and southern Australiamore over Indonesia; drought in East Africa
Indian summer monsoontends to be favouredtends to be suppressed
  • The IOD usually develops in May–June, peaks in September–November and decays with the onset of the northeast monsoon. It is phase-locked to the seasons because the cooling feedback off Sumatra works only when the alongshore winds are south-easterly.
  • The Bjerknes feedback works here too, through Sumatra–Java upwelling, equatorial winds and SST gradient.
  • Strong positive events include 1994, 1997, 2006 and 2019. IOD events often, but not always, co-occur with El Niño.

Thermohaline Circulation

The equation of state, linearised

Density depends on temperature, salinity and pressure. Near a reference state,

$$\rho \approx \rho_0\left[1 - \alpha(T - T_0) + \beta(S - S_0)\right]$$

with thermal expansion coefficient \( \alpha \approx 2\times 10^{-4} \) K\(^{-1}\) (much smaller in cold water) and haline contraction coefficient \( \beta \approx 7.6\times 10^{-4} \) per salinity unit. The nonlinearity of the real equation of state (\( \alpha \) shrinking at low temperature and rising with pressure) causes cabbeling (mixing two waters of equal density gives a denser mixture) and thermobaricity.

Where deep water forms and why

Deep water forms in only a few places, where surface water becomes dense enough to sink through the stratification:

Water massWhereHow
North Atlantic Deep Water (NADW)Nordic Seas (overflows across the Greenland–Scotland Ridge), Labrador Seawinter cooling of salty water brought north by the Gulf Stream and North Atlantic Current; open-ocean convection
Antarctic Bottom Water (AABW)Weddell and Ross Seas, Adélie coastcooling plus brine rejection as sea ice forms in coastal polynyas
No deep waterNorth Pacificsurface too fresh: even at freezing point it is not dense enough

Deep water spreads as deep western boundary currents (Stommel–Arons theory: slow interior upwelling everywhere requires poleward interior flow and concentrated western boundary currents). It returns to the surface mainly by wind-driven upwelling in the Southern Ocean and by diapycnal mixing powered by winds and tides. Full overturning takes of order 1,000 years; radiocarbon shows the oldest deep water in the North Pacific.

The AMOC and its stability

The Atlantic Meridional Overturning Circulation carries heat northward across the equator. Stommel’s (1961) two-box model showed that a salinity-driven overturning can have two stable states: if freshwater input to the high-latitude box exceeds a threshold, the circulation can collapse into a weak or reversed state, with hysteresis. Two feedbacks compete:

  • Salt-advection feedback (positive): a strong overturning carries salt north, keeping northern water dense and the overturning strong. Freshening weakens it, which brings less salt, which weakens it further.
  • Thermal feedback (negative): a weaker overturning lets northern water cool more, raising its density.

Paleoceanographic records show that the AMOC weakened sharply during Heinrich events and the Younger Dryas, when meltwater freshened the North Atlantic. Continuous monitoring across 26.5°N began in 2004. Proxies for past AMOC strength are on Paleoceanographic Proxies and Paleo-SST Methods.

Three-component mixing

Below the mixed layer, \( T \) and \( S \) are conservative, so a sample is a mixture of source water types. For three types, the fractions \( x_1, x_2, x_3 \) follow from

$$x_1T_1 + x_2T_2 + x_3T_3 = T, \quad x_1S_1 + x_2S_2 + x_3S_3 = S, \quad x_1 + x_2 + x_3 = 1$$

Memory Tricks

Don’t Confuse

ThisvsThat
Ekman transport (90°, depth-integrated)Surface Ekman current (45°)
Ekman depth \( D_E \) (e-folding × \( \pi \))Mixed layer depth (set by turbulence and heat flux)
Coastal upwelling (alongshore wind)Ekman pumping (wind-stress curl, open ocean)
SOI (pressure, negative in El Niño)ONI / Niño 3.4 (SST, positive in El Niño)
El Niño (east Pacific warm)Positive IOD (east Indian Ocean cool)
Kelvin wave (eastward, equatorially trapped, fast)Rossby wave (westward, about three times slower)
Thermohaline: driven by surface buoyancy fluxesWind-driven: gyres, Ekman, boundary currents
NADW: salty, open-ocean convectionAABW: coldest, densest, brine rejection on shelves

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper.

Level 1: Foundation

  1. At the Ekman depth, in which direction does the current flow relative to the surface current?
  2. Which two stations define the classical Southern Oscillation Index?
  3. In which season does the Indian Ocean Dipole usually peak?
Answers
  1. Opposite to it, with about 4% (\( e^{-\pi} \)) of its speed.
  2. Tahiti and Darwin.
  3. September–November (boreal autumn).

Level 2: Part B Style

Practice Q1. A steady wind blows toward the north along a north–south coast in the northern hemisphere, with the ocean to the west of the coast. The net Ekman transport is:

(a) toward the north (b) toward the east (onshore) (c) toward the west (offshore) (d) toward the south

Answer: (b) toward the east

In the NH, transport is 90° to the right of the wind. Right of a northward wind is east, which here is onshore: downwelling. (c) uses the southern-hemisphere rule. A west-coast upwelling system like California needs an equatorward (southward) wind.

Practice Q2. Using \( D_E = 7.6\,U_{10}/\sqrt{\sin|\phi|} \), the Ekman depth for an 8 m s\(^{-1}\) wind at 30°N is about:

(a) 43 m (b) 61 m (c) 86 m (d) 122 m

Answer: (c) 86 m

\( \sqrt{\sin 30°} = 0.707 \); \( D_E = 60.8/0.707 \approx 86 \) m. (b) forgets the latitude factor; (d) multiplies by 2 instead of dividing by \( \sqrt{0.5} \).

Practice Q3. During a developing El Niño, which change is expected?

(a) the thermocline shoals in the eastern equatorial Pacific (b) downwelling Kelvin waves travel eastward along the equator (c) the SOI becomes strongly positive (d) convection shifts west over the maritime continent

Answer: (b)

Westerly wind bursts excite downwelling Kelvin waves that deepen the eastern thermocline. (a), (c) and (d) describe La Niña-like changes.

Practice Q4. Antarctic Bottom Water is the densest water mass mainly because of:

(a) high temperature and high salinity (b) evaporation in the subtropics (c) cooling and brine rejection during sea-ice formation (d) inflow of Mediterranean Water

Answer: (c)

Sea ice rejects most of its salt, so the water left beneath it near the freezing point becomes both very cold and saltier.

Level 3: Part C Style

Practice Q5. Monthly Niño 3.4 SST anomalies (°C) for May to January are:

MayJunJulAugSepOctNovDecJan
0.20.40.50.70.60.80.90.30.2

Using the NOAA rule (ONI = 3-month running mean; El Niño if ONI ≥ +0.5 °C for at least five consecutive overlapping seasons), which statements are correct?

  1. The MJJ (May–June–July) ONI is below the threshold.
  2. ONI is at or above +0.5 °C from JJA through OND.
  3. The episode qualifies as El Niño conditions under the rule.
  4. The highest ONI in the series is 0.9 °C.

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

Answer: (b) 1, 2 and 3 only

Running means: MJJ 0.37, JJA 0.53, JAS 0.60, ASO 0.70, SON 0.77, OND 0.67, NDJ 0.47.

  • 1: 0.37 < 0.5. True.
  • 2: JJA, JAS, ASO, SON and OND are all ≥ 0.5. True.
  • 3: five consecutive seasons meet the threshold. True.
  • 4: false. 0.9 is a single month; the highest 3-month mean is 0.77 (SON).

Practice Q6. An equatorward wind stress of 0.08 N m\(^{-2}\) blows along a coast at 20°N (\( \rho = 1025 \) kg m\(^{-3}\)). Upwelling is concentrated within 20 km of the coast. Which statements are correct?

  1. \( f \approx 5.0\times 10^{-5} \) s\(^{-1}\).
  2. The offshore Ekman transport is about 1.6 m\(^2\) s\(^{-1}\) per metre of coast.
  3. The upwelling velocity is about 7 m per day.
  4. Doubling the vertical eddy viscosity would double the offshore transport.

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

Answer: (a) 1, 2 and 3 only
  • \( f = 2\times 7.292\times 10^{-5}\times\sin 20° = 4.99\times 10^{-5} \) s\(^{-1}\). True.
  • \( U_E = 0.08/(1025\times 4.99\times 10^{-5}) = 1.56 \) m\(^2\) s\(^{-1}\). True.
  • \( w = 1.56/20{,}000 = 7.8\times 10^{-5} \) m s\(^{-1}\) = 6.8 m day\(^{-1}\) ≈ 7 m per day. True.
  • 4 false: Ekman transport is independent of \( A_z \).

Practice Q7. In a year with a positive IOD and no El Niño, which outcomes are most consistent with the composite patterns?

  1. Enhanced upwelling and negative SST anomalies off Sumatra–Java in September–October.
  2. Above-normal October–December rainfall in coastal East Africa.
  3. Drought tendency over Indonesia.
  4. A negative Dipole Mode Index.

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

Answer: (a) 1, 2 and 3 only

A positive IOD means a cool east (1), with convection shifted west so East Africa is wet (2) and Indonesia is dry (3). By definition the DMI (west minus south-east) is positive, so 4 is false.

Quick Revision Table

IdeaFact or formula
Coriolis parameter\( f = 2\Omega\sin\phi \), \( \Omega = 7.292\times 10^{-5} \) s\(^{-1}\)
Ekman depth\( D_E = \pi\sqrt{2A_z/\lvert f\rvert} \); empirical \( 7.6U_{10}/\sqrt{\sin\lvert\phi\rvert} \) m
Surface Ekman current45° right (NH); \( V_0 \approx 0.0127U_{10}/\sqrt{\sin\lvert\phi\rvert} \)
Ekman transport\( \tau/(\rho f) \), 90° right (NH); independent of \( A_z \)
Wind stress\( \tau = \rho_{air}C_DU_{10}^2 \)
Coastal upwelling\( w \approx U_E/L \), \( L \) about 10–50 km; metres per day
Ekman pumping\( w_E = \rho^{-1}\,\text{curl}(\boldsymbol\tau/f) \); cyclonic curl → upwelling
ENSO normal statetrades, warm pool west, cold tongue east, Walker cell, Bjerknes feedback
El Niñoweak trades, flat thermocline, warm east, negative SOI, ONI ≥ +0.5 °C for 5 seasons
Niño 3.4 box5°N–5°S, 170°W–120°W
WavesKelvin eastward \( c = \sqrt{g'H} \) (2–3 m s\(^{-1}\)); Rossby westward \( c/3 \)
ENSO theoriesdelayed oscillator; recharge oscillator
IODDMI = west (50–70°E, 10°S–10°N) minus south-east (90–110°E, 10°S–0°); peaks SON; 1999
Positive IODcool off Sumatra, wet East Africa, dry Indonesia; favours Indian monsoon
Deep waterNADW (Nordic and Labrador Seas), AABW (Weddell, Ross); none in North Pacific
AMOCStommel two-box model; salt-advection feedback; Heinrich and Younger Dryas slowdowns

Key Takeaways

What’s Next