Climate Variability: MJO, ENSO, QBO and Sunspot Cycles
CSIR NET Part C meteorology: Madden–Julian Oscillation, ENSO dynamics and indices, the stratospheric QBO, and solar-cycle forcing of climate.
- 21 min
On this page · 14 sections
Prerequisites
- The Walker circulation, equatorial Kelvin and Rossby waves and the Gill response, from Tropical East–West Circulations.
- EP flux, wave–mean-flow interaction and the Brewer–Dobson circulation, from Zonally Symmetric and Asymmetric Circulation and Eddy Transports.
- Monsoon active and break spells and the Part B ENSO description, from Monsoon, Cyclones, Western Disturbances and ENSO.
- Spectra, periodograms and filtering, from Continuous and Discrete Signals, Fourier Series and Transforms.
The Hook: Four Clocks in One Atmosphere
The Four Modes at a Glance
| Mode | Time scale | Where | Essential mechanism | Main signature |
|---|---|---|---|---|
| MJO | 30–60 days (broadly 20–100) | tropical troposphere, Indian Ocean to central Pacific | convection coupled to planetary-scale circulation; propagates east | eastward-moving envelope of convection and 850/200 hPa wind anomalies |
| ENSO | 2–7 years, irregular | tropical Pacific ocean and atmosphere | Bjerknes feedback plus ocean wave adjustment | Niño 3.4 SST, SOI, thermocline depth |
| QBO | about 28 months on average | equatorial stratosphere, about 16–50 km | wave–mean-flow interaction with vertically propagating equatorial waves | alternating easterly and westerly regimes descending with time |
| Sunspot cycle | about 11 years (magnetic Hale cycle about 22 years) | the Sun | solar dynamo | total solar irradiance change about 0.1%; much larger UV change |
graph LR
A[Weeks: MJO] --> B[Months to a season: monsoon intraseasonal modes]
B --> C[About 2.3 years: QBO]
C --> D[2 to 7 years: ENSO]
D --> E[About 11 years: solar cycle]The Madden–Julian Oscillation
Discovery and structure
Roland Madden and Paul Julian (1971, 1972) found a 40–50-day oscillation in tropical winds and pressure and showed that it is a global-scale, eastward-moving disturbance. It is the dominant mode of tropical intraseasonal variability.
| Property | Typical value |
|---|---|
| Period | 30–60 days |
| Zonal scale | planetary: zonal wavenumbers 1–3 |
| Propagation | eastward at about 5 m s\(^{-1}\) while coupled to convection over the Indian Ocean and western Pacific |
| East of the date line | convection usually dies out; the circulation signal can continue faster as a largely dry, Kelvin-like disturbance |
| Vertical structure | first baroclinic: low-level and upper-level winds of opposite sign |
| Season | strongest in boreal winter and spring, when it stays close to the equator |
Circulation around the convective envelope (Gill-like):
- Low-level westerly anomalies west of the convection, easterly anomalies east of it; low-level convergence into the convection.
- Upper-level winds reversed, with divergence above the convection.
- Rossby-type cyclonic gyres either side of the equator to the west of the convection, and a Kelvin-type response to the east.
- Suppressed convection ahead (east) and behind (west) of the active envelope.
Life cycle and the RMM index
Wheeler and Hendon (2004) defined the Real-time Multivariate MJO (RMM) index from the first two empirical orthogonal functions of combined, equatorially averaged outgoing longwave radiation, 850 hPa zonal wind and 200 hPa zonal wind. The two principal components define a phase diagram with eight phases; amplitude below 1 is usually treated as “weak or no MJO”.
| RMM phases | Location of enhanced convection |
|---|---|
| 8 and 1 | Western Hemisphere and Africa |
| 2 and 3 | Indian Ocean |
| 4 and 5 | Maritime Continent |
| 6 and 7 | western Pacific |
The Maritime Continent barrier: MJO events often weaken or stall over the islands and seas of Indonesia, where the diurnal cycle of convection over land competes with the large-scale envelope.
Why it propagates (the theory is not settled)
No single theory is universally accepted. The main ideas:
| Idea | Core argument |
|---|---|
| Moisture mode | the MJO is a slow mode of column moisture; horizontal advection and surface fluxes moisten the region east of the convection, which then becomes the new centre |
| Frictional convergence | boundary-layer convergence in the Kelvin-wave region east of the convection preconditions new convection there |
| WISHE | wind-induced surface heat exchange enhances fluxes where the total wind is strongest |
| Skeleton models | interaction between convective activity and planetary-scale waves produces a slow, eastward neutral mode |
What is agreed: the MJO is much slower than a dry Kelvin wave because it is coupled to convection, which reduces the effective static stability.
Why the MJO matters
- Indian monsoon: in boreal summer the intraseasonal variability takes the form of the boreal summer intraseasonal oscillation (BSISO), whose bands of convection move northward from the equatorial Indian Ocean to the Indian subcontinent (roughly a degree of latitude a day) as well as eastward. It drives the active and break spells and can influence onset.
- Tropical cyclones: genesis is more likely in and just behind the active phase of the MJO, in the north Indian Ocean as elsewhere.
- ENSO: MJO-related westerly wind bursts over the western Pacific excite downwelling oceanic Kelvin waves that can trigger or amplify El Niño.
- Extratropics: MJO heating excites Rossby wave trains that alter middle-latitude weather 1–3 weeks later. This is a major source of sub-seasonal predictability.
- QBO link: in boreal winter the MJO has been observed to be stronger and more persistent when the lower-stratospheric QBO is easterly.
ENSO: the Coupled Oscillation
Indices
| Region or index | Definition |
|---|---|
| Niño 1+2 | 0–10° S, 90° W–80° W (coastal Peru and Ecuador) |
| Niño 3 | 5° N–5° S, 150° W–90° W |
| Niño 3.4 | 5° N–5° S, 170° W–120° W (the most widely used) |
| Niño 4 | 5° N–5° S, 160° E–150° W |
| Oceanic Niño Index (ONI, NOAA) | 3-month running mean of the Niño 3.4 SST anomaly; El Niño when it is at least +0.5 °C for at least five consecutive overlapping seasons, La Niña at −0.5 °C or below |
| SOI | standardised Tahiti-minus-Darwin sea-level pressure difference; negative in El Niño |
ENSO events tend to peak in boreal winter (roughly November–January): they are phase-locked to the annual cycle.
Growth: the Bjerknes feedback
A warm anomaly in the eastern Pacific weakens the zonal SST gradient, which weakens the trades, which deepens the eastern thermocline and reduces upwelling of cold water, which warms the east further. A cold anomaly runs the same loop in reverse (La Niña).
Turning around: ocean wave adjustment
A positive feedback alone would not oscillate. The reset comes from the ocean’s slow adjustment.
Delayed oscillator (Suarez and Schopf, 1988; Battisti and Hirst, 1989). Westerly wind anomalies in the central Pacific during El Niño excite:
- downwelling Kelvin waves travelling east, which deepen the eastern thermocline and reinforce the warming quickly;
- upwelling Rossby waves travelling west, which reflect at the western boundary as upwelling Kelvin waves that return east after a delay and raise the eastern thermocline, ending the event.
In non-dimensional form the idea is often written as
$$ \frac{dT}{dt} = T - T^3 - \alpha\,T(t-\delta) $$where the first term is the Bjerknes growth, the cubic term limits amplitude, and the delayed negative term is the reflected-wave signal arriving after a delay \( \delta \).
Recharge–discharge oscillator (Jin, 1997). The equatorial Pacific’s warm water volume (heat content above the thermocline) builds up before El Niño (recharge) and is exported poleward during it by Sverdrup transport (discharge), leaving a shallow thermocline that favours La Niña. Warm water volume therefore leads Niño 3.4 SST by several months and is a useful predictor.
| Model | Key negative feedback | Predictor it highlights |
|---|---|---|
| Delayed oscillator | reflected Rossby → Kelvin waves | wave travel time across the basin |
| Recharge oscillator | discharge of equatorial heat content | warm water volume |
| Western Pacific oscillator, advective–reflective (other variants) | off-equatorial wind and zonal advection | various |
ENSO diversity and predictability
- Eastern Pacific vs central Pacific El Niño. In some events the maximum warming lies in the central Pacific with cooler water on either side (the “El Niño Modoki” pattern described by Ashok and colleagues in 2007). Teleconnections differ, and central-Pacific warming has been associated with a stronger tendency for a weak Indian monsoon.
- Spring predictability barrier. Forecasts made before or through boreal spring lose skill rapidly, because the coupled system is weakest and most sensitive to noise (such as MJO wind bursts) in spring.
- Asymmetry. Strong El Niño events are typically stronger than strong La Niña events, while La Niña events more often last two years or more.
ENSO teleconnections
| Region | Typical El Niño impact |
|---|---|
| India | tends to weaken the summer monsoon, through anomalous descent from the shifted Walker cell (not every El Niño year is deficient) |
| Indonesia and Australia | drought |
| Peru and Ecuador | heavy coastal rain, collapse of upwelling-fed fisheries |
| Southern Africa | dry summer |
| North Pacific and North America | deepened Aleutian low; Pacific–North American (PNA) wave train |
| Atlantic | fewer Atlantic hurricanes (stronger vertical shear) |
The ENSO–monsoon relationship is examined further on The Ocean’s Role, Coupled Models and Monsoon Links.
The Quasi-Biennial Oscillation
What is observed
The QBO is an alternation of easterly and westerly zonal-mean zonal winds in the equatorial stratosphere.
| Property | Typical value |
|---|---|
| Mean period | about 28 months (individual cycles range from roughly 22 to 34 months) |
| Vertical range | roughly 100 hPa to a few hPa (about 16–50 km) |
| Latitudinal range | within roughly 10–15° of the equator; nearly symmetric |
| Amplitude | of order 20 m s\(^{-1}\) or more near 20–30 hPa; easterly phases typically stronger than westerly |
| Downward propagation | regimes appear at the top and descend at about 1 km per month, without loss of amplitude until the lowest levels |
| Shear zones | westerly shear zones descend faster and more regularly than easterly ones |
It was discovered from radiosonde winds around 1960 (Reed and colleagues in the United States and Ebdon in the United Kingdom, independently). It is not locked to the annual cycle, although transitions tend to occur in certain seasons. In 2015–16 the regular cycle was disrupted for the first time in the observed record, when an easterly layer formed unexpectedly within the descending westerlies; a similar disruption occurred in 2019–20.
Mechanism: waves drive the oscillation
Lindzen and Holton (1968) and Holton and Lindzen (1972) explained the QBO as wave–mean-flow interaction:
- Vertically propagating equatorial waves carry zonal momentum up from the troposphere: Kelvin waves carry westerly (eastward) momentum, mixed Rossby–gravity waves carry easterly momentum. Small-scale gravity waves of both signs are now known to provide much of the forcing.
- A wave is absorbed (deposits its momentum) near its critical level, where the mean wind approaches its phase speed, and is damped more strongly where the Doppler-shifted frequency is low.
- If the mean wind at some level is westerly, eastward-moving waves are absorbed just below it, depositing westerly momentum and pushing the westerly regime downward. Easterly waves pass through and are absorbed higher up, building the next easterly regime aloft.
- The descending shear zones alternate, giving an oscillation whose period is set by the wave momentum flux, not by any external clock.
The laboratory analogue of Plumb and McEwan (1978), a salt-stratified fluid in an annulus forced at the bottom by a standing wave of fixed frequency, produced descending, alternating mean flows with a period far longer than the forcing period, showing that the mechanism needs no external forcing at the oscillation period.
A mean upwelling (the tropical branch of the Brewer–Dobson circulation) opposes the descent and lengthens the period.
Why the QBO matters
| Effect | Description |
|---|---|
| Holton–Tan effect | the northern winter polar vortex tends to be weaker and more disturbed when the lower-stratospheric (about 50 hPa) QBO is easterly |
| Ozone and trace gases | a QBO signal in column ozone, through its induced secondary meridional circulation |
| MJO | MJO stronger in boreal winter during easterly QBO (observed in recent decades) |
| Prediction | the QBO is predictable a year or more ahead and is a source of long-range predictability |
Sunspot Cycles and Solar Forcing
The solar cycle
| Feature | Value or fact |
|---|---|
| Schwabe cycle | sunspot number varies with a period of about 11 years (individual cycles roughly 9–14 years) |
| Hale cycle | the Sun’s magnetic polarity reverses each Schwabe cycle, so the full magnetic cycle is about 22 years |
| Longer modulations | the Gleissberg cycle of roughly 80–100 years is often cited |
| Maunder Minimum | about 1645–1715, when very few sunspots were seen; it overlaps part of the Little Ice Age |
| Total solar irradiance (TSI) | about 1361 W m\(^{-2}\) at solar minimum; varies by about 1 W m\(^{-2}\), about 0.1%, between minimum and maximum |
| Ultraviolet | varies by several per cent at wavelengths around 200 nm, and much more at shorter wavelengths |
Why the Sun is brighter at sunspot maximum. Sunspots are dark, but at solar maximum the bright faculae that accompany them more than compensate, so TSI is higher at maximum.
From irradiance change to radiative forcing
A change \( \Delta S \) in TSI gives a global-mean radiative forcing
$$ \Delta F = \frac{\Delta S\,(1-\alpha)}{4} $$With \( \Delta S = 1 \) W m\(^{-2}\) and planetary albedo \( \alpha = 0.3 \), \( \Delta F \approx 0.18 \) W m\(^{-2}\). Multiplying by a climate sensitivity parameter of about 0.8 K per W m\(^{-2}\) gives an equilibrium response of about 0.14 K. Because the 11-year cycle is fast compared with the ocean’s response time, the actual surface signal is smaller, of order 0.1 K or less in observational analyses.
Where the solar signal is clearest
- Stratosphere: UV changes alter ozone production and heating, so the solar-cycle signal in upper-stratospheric temperature and ozone is relatively large. This is a “top-down” route that may influence the polar vortex and modes such as the North Atlantic Oscillation.
- Surface: a small, hard-to-detect signal, partly because ENSO and volcanic eruptions add variability at similar time scales.
- Cosmic rays and clouds: a proposed indirect route (galactic cosmic rays, modulated by the solar wind, influencing aerosol nucleation and cloud). Laboratory and observational tests so far indicate that the effect on clouds is too small to be important for recent climate change.
Solar forcing cannot explain recent warming
| Evidence | Why it rules out the Sun as the main cause |
|---|---|
| TSI trend | measured TSI has shown no upward trend since satellite records began in the late 1970s, while global temperature has risen strongly |
| Vertical fingerprint | a brighter Sun would warm both troposphere and stratosphere; observations show tropospheric warming and stratospheric cooling, the greenhouse-gas fingerprint |
| Diurnal and seasonal pattern | greenhouse warming is expected to be larger at night and in winter, which is what has been observed over many land regions; solar warming would be strongest by day and in summer |
Detecting periodicities in data
All four modes are often identified in Part C from spectra or time series:
- Period from a spectral peak: \( T = 1/f \). A peak at \( f = 0.0357 \) cycle per month means \( T = 28 \) months: the QBO. A peak at 0.091 cycle per year means 11 years: the solar cycle.
- Red-noise background: geophysical series have more variance at low frequencies; a peak must stand above a red-noise (first-order autoregressive) spectrum to be significant.
- Nyquist and record length: monthly data cannot resolve periods shorter than 2 months; an MJO study needs daily data. A record must contain several cycles (many decades for the solar cycle) to establish a periodicity.
- Band-pass filtering (for example 20–100 days for the MJO) isolates a mode before compositing.
Memory Tricks
Don’t Confuse
| These | Difference |
|---|---|
| MJO vs convectively coupled Kelvin wave | planetary scale, about 5 m s\(^{-1}\), 30–60 days vs smaller scale, about 15 m s\(^{-1}\) or more, a few days |
| MJO vs BSISO | eastward, boreal winter, near the equator vs eastward and northward, boreal summer, Indian and Asian monsoon |
| Delayed vs recharge oscillator | reflected wave delay vs discharge of warm water volume |
| ONI vs SOI | ocean (Niño 3.4 SST, El Niño positive) vs atmosphere (pressure, El Niño negative) |
| QBO vs SAO | about 28 months, 16–50 km vs 6 months, stratopause and mesopause |
| Schwabe vs Hale cycle | sunspot number, about 11 years vs magnetic polarity, about 22 years |
| TSI vs UV variation | about 0.1% vs several per cent or more |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Take the Earth’s radius as \( 6.371\times 10^{6} \) m.
Level 1: Foundation
- Which three fields make up the RMM index?
- What is the mean period of the QBO?
- Write the radiative forcing produced by a change \( \Delta S \) in total solar irradiance.
Answers
- Outgoing longwave radiation, 850 hPa zonal wind and 200 hPa zonal wind (equatorially averaged).
- About 28 months.
- \( \Delta F = \Delta S(1-\alpha)/4 \).
Level 2: Part B Style
Practice Q1. An MJO convective envelope moves along the equator from 60° E to 180° at 5 m s\(^{-1}\). The time taken is closest to:
(a) 10 days (b) 20 days (c) 31 days (d) 60 days
Answer: (c) about 31 days
Distance: \( 120° \times 111.2 \) km \( = 13{,}340 \) km. Time: \( 1.334\times 10^{7}/5 = 2.67\times 10^{6} \) s \( \approx 31 \) days. This is about half of a 60-day cycle, consistent with one active envelope crossing the Indo-Pacific while a suppressed one follows.
Practice Q2. A height–time section of equatorial zonal wind shows alternating easterly and westerly regimes between 70 hPa and 5 hPa that descend with time, each full cycle lasting about 2.3 years. The phenomenon and its main driver are:
(a) the SAO, driven by seasonal heating (b) the QBO, driven by vertically propagating equatorial waves (c) the MJO, driven by convection (d) ENSO, driven by Bjerknes feedback
Answer: (b)
Equatorial stratosphere, a period of about 28 months and downward-propagating regimes identify the QBO; Kelvin, mixed Rossby–gravity and gravity waves deposit momentum near their critical levels.
Practice Q3. If total solar irradiance rises by 1.0 W m\(^{-2}\) from solar minimum to maximum and the planetary albedo is 0.30, the global-mean radiative forcing is closest to:
(a) 0.07 W m\(^{-2}\) (b) 0.18 W m\(^{-2}\) (c) 0.70 W m\(^{-2}\) (d) 1.0 W m\(^{-2}\)
Answer: (b)
\( \Delta F = 1.0\times 0.7/4 = 0.175 \) W m\(^{-2}\). Option (c) forgets the factor 4 (sphere vs disc); (d) also ignores albedo.
Practice Q4. In the delayed-oscillator view of ENSO, what terminates an El Niño event?
(a) Increased cloud cover over the eastern Pacific (b) Upwelling Kelvin waves produced by reflection of westward-moving Rossby waves at the western boundary (c) Stronger trades caused by the MJO (d) A decrease in solar irradiance
Answer: (b)
The westerly wind anomalies that sustain El Niño also generate upwelling Rossby waves that travel west, reflect at the western boundary as upwelling Kelvin waves and raise the eastern thermocline after a delay. (a) is a damping effect but not the oscillator’s turnaround mechanism.
Level 3: Part C Style
Practice Q5. In a simple delayed-oscillator estimate, the equatorial Pacific is 15,000 km wide and the oceanic first-baroclinic Kelvin wave speed is 2.5 m s\(^{-1}\). The first-meridional-mode long Rossby wave travels at one-third of that speed. The time for a Rossby wave generated at the eastern side to cross to the western boundary and return as a Kelvin wave is closest to:
(a) 2 months (b) 4.5 months (c) 9 months (d) 2 years
Answer: (c) about 9 months
- Rossby leg: \( 1.5\times 10^{7}/(2.5/3) = 1.8\times 10^{7} \) s \( \approx 208 \) days.
- Kelvin leg: \( 1.5\times 10^{7}/2.5 = 6.0\times 10^{6} \) s \( \approx 69 \) days.
- Total: about 278 days \( \approx 9 \) months.
Waves generated in the central Pacific have shorter paths, but the result shows why the delay is a sizeable fraction of a year. Together with growth and decay phases this helps set ENSO’s 2–7-year time scale.
Practice Q6. Successive 3-month running means of the Niño 3.4 SST anomaly (°C) are:
| MJJ | JJA | JAS | ASO | SON | OND | NDJ | DJF | JFM |
|---|---|---|---|---|---|---|---|---|
| 0.3 | 0.5 | 0.7 | 0.9 | 1.1 | 1.3 | 1.4 | 1.2 | 0.8 |
Using the NOAA ONI rule, which statements are correct?
- An El Niño episode is identified.
- The episode peaks in boreal winter, consistent with ENSO phase locking.
- The episode began in MJJ.
- The SOI would be expected to be negative during OND.
(a) 1, 2 and 4 only (b) 1 and 3 only (c) 2 and 3 only (d) 1, 2, 3 and 4
Answer: (a) 1, 2 and 4 only
- 1 true: from JJA to JFM there are eight consecutive overlapping seasons at or above +0.5 °C, well over the five required.
- 2 true: the maximum (+1.4 °C) is in NDJ.
- 3 false: MJJ (+0.3 °C) is below the threshold; the episode starts in JJA.
- 4 true: El Niño is accompanied by a negative SOI.
Practice Q7. A periodogram of a 60-year monthly record of equatorial 30 hPa zonal wind shows a sharp peak at 0.0357 cycle per month. A second record, 150 years of annual mean sunspot number, peaks at 0.091 cycle per year. Which statements are correct?
- The first peak corresponds to a period of about 28 months.
- The second peak corresponds to a period of about 11 years.
- Monthly data could not resolve the MJO even with a long record.
- A 15-year record of annual sunspot number would be adequate to establish an 11-year periodicity with confidence.
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 1, 2 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
- 1 true: \( 1/0.0357 = 28.0 \) months.
- 2 true: \( 1/0.091 = 11.0 \) years.
- 3 true: monthly sampling has a Nyquist period of 2 months and monthly averaging smooths out 30–60-day variability; daily data are needed.
- 4 false: 15 years contains barely more than one cycle; establishing a periodicity against a red-noise background needs many cycles.
Quick Revision Table
| Idea | Fact or formula |
|---|---|
| MJO | 30–60 days; wavenumber 1–3; east at about 5 m s\(^{-1}\); Indian Ocean to western Pacific; westerlies west, easterlies east of convection |
| RMM | EOFs of OLR, U850, U200; 8 phases; 2–3 Indian Ocean, 4–5 Maritime Continent, 6–7 western Pacific |
| BSISO | northward propagation in boreal summer; active and break spells over India |
| ENSO indices | Niño 3.4 (5° N–5° S, 170°–120° W); ONI ≥ +0.5 °C for 5 seasons; SOI negative in El Niño |
| ENSO mechanisms | Bjerknes growth; delayed oscillator (wave reflection); recharge oscillator (warm water volume) |
| ENSO behaviour | 2–7 years; peaks Nov–Jan; spring predictability barrier; EP and CP types |
| QBO | about 28 months; 16–50 km; ±10–15°; descends about 1 km per month; Kelvin (W) and Yanai (E) plus gravity waves |
| Holton–Tan | easterly QBO at 50 hPa → weaker northern polar vortex |
| Solar cycle | Schwabe 11 yr; Hale 22 yr; Maunder Minimum about 1645–1715 |
| TSI | about 1361 W m\(^{-2}\); cycle change about 1 W m\(^{-2}\) (0.1%) → \( \Delta F \approx 0.18 \) W m\(^{-2}\) |
| Fingerprint | greenhouse: troposphere warms, stratosphere cools; solar would warm both |
Key Takeaways
What’s Next
- Grid-Point and Spectral GCMs: how models are built to simulate these modes.
- The Ocean’s Role, Coupled Models and Monsoon Links: ENSO, the IOD and the monsoon in coupled models.
- Long-term change and attribution in Climate Change.
- Back to the chapter overview.