The El Niño/ Southern Oscillation (ENSO) Cycle

Michelle L’HeureuxNOAA Climate Prediction Center (CPC)

August 2013

Outline

(1) Brief Overview of the Ocean and Atmosphere (“Normal” vs. ENSO conditions) (2) The ENSO Cycle: A Coupled Ocean- Atmosphere System

(3) ENSO Teleconnections and Global Impacts

(4) ENSO Monitoring and Prediction at NOAA Climate Prediction Center (CPC)

ENSO in a Nutshell• An irregular, naturally occurring cycle (every 2-7 years) of warm (El Niño) or cold (La

Niña) conditions in the tropical Pacific Ocean. • Ocean changes occur alongside changes in the tropical atmosphere circulation & rainfall• On average, events last 9-12 months (La Niñas can persist longer) and peak in strength

during N. Hemisphere winter

Niño 3.4 sea surface temperatures (SST): Primary ENSO index or time series

Red colors: above average sea surface temps (SST)

Blue colors: below average sea surface temps (SST)

“Normal” SST: Major Features

Atlantic Warm Pool

Pacific Warm Pool Cold Tongues

“Normal” SST: Extremes in the Annual Cycle

Equatorial SSTs are warmest in April

Equatorial cold tongues are strongest in Jul.-Oct.

Sea Surface Temperatures:El Niño vs. La Niña

Equatorial cold tongue is weaker than average or absent during El Niño, resulting in positive SST anomalies

Equatorial cold tongue is stronger than average during La Niña, resulting in negative SST anomalies

Decadal Changes in Structure and Amplitude of El Niños

Eastern Pacific (EP) El Niño = Cold Tongue El Niño = Conventional or Canonical El Niño

The stronger El Niño events in the 80s and 90s

resembled EP El Niño

Central Pacific (CP) El Niño = Warm Pool El Niño = El Niño Modoki = Date Line El Niño

Since ~2000, El Niño has often resembled CP El Niño

2009-2010

“Normal” Precipitation: Major Features

SPCZ SACZ

ITCZ

Storm Tracks

“Normal” Precipitation: Extremes in the Annual Cycle

Precipitation:El Niño vs. La Niña

Enhanced rainfall occurs over warmer-than-average waters during El Niño.

Reduced rainfall occurs over colder-than-average waters during La Niña.

Sea Level Pressure: “Southern Oscillation”

El Niño: Positive SLP anomalies over the western tropical Pacific, Indonesia and Australia. Negative SLP anomalies over eastern tropical Pacific, middle and high latitudes of the North Pacific, and over U.S. Opposite pattern for La Niña. The pressure see-saw between the eastern and western tropical Pacific is known as the “Southern Oscillation.”

Low-Level Winds & Thermocline Depth: El Niño vs. La Niña

La Niña: stronger-than-average easterlies lead to a deeper (shallower)-than-average thermocline in the western (eastern) eq. Pacific.

El Niño: weaker-than-average easterlies lead to a deeper (shallower)-than-average thermocline in the eastern (western) eq. Pacific.

Animation of Subsurface Temperatures: 1996-1999

Outline

(1) Brief Overview of the Ocean and Atmosphere (“Normal” vs. ENSO conditions) (2) The ENSO Cycle: A Coupled Ocean- Atmosphere System

(3) ENSO Teleconnections and Global Impacts

(4) ENSO Monitoring and Prediction at NOAA Climate Prediction Center (CPC)

ENSO depends on coupling between the ocean and the atmosphere over the tropical Pacific Ocean.

COLDWARM

The Bjerknes feedback is a positive feedback between the ocean and atmosphere over the equatorial Pacific Ocean.

(1) Stronger SST gradients stronger winds.

wind

If surface water blows from east to west, cold water from the deep ocean upwells to the surface ocean to replace water

Water that travels westward over the wide expanse of the Pacific, warms up due to solar insolation

(2) Stronger winds stronger SST gradientsu

pw

ellin

g

What is “Normal?”

WarmCold

WarmCold

Winds and Sea Surface Temperature are COUPLED. The SSTs determine the winds and vice versa.

(1) Easterly trade-winds help push warm water to the western Pacific and upwell cold water in the eastern Pacific Ocean.

(2) Warm water heats the atmosphere and makes it rise, so low-level trade winds blow towards warm water to fill the gap. Subsiding air occurs in the eastern basin.

“La Niña”

ColdWarm

Warm Cold

Stronger Stronger

Upwelling

Enhanced

More Convection

becomes more shallow

• Convection becomes stronger over the far western Pacific Ocean/ Indonesia and more suppressed in the central

Pacific.

• Easterly trade winds strengthen

• Thermocline becomes more shallow and the cold water upwelling increases in the

eastern Pacific.

Cold

“El Niño”

WarmCold

Warm

ColdWarm

• Convection shifts eastward over the central and/or eastern Pacific Ocean. Convection becomes suppressed over the far western Pacific/

Indonesia.

• Easterly trade winds weaken

• Thermocline deepens and the cold water upwelling decreases in

the eastern Pacific.

NOTE: Location of the warmest SSTs (>28°C)

determines where tropical convection will

be located.

Features of the ENSO Cycle

• Irregular cycle with periods of warm (El Niño) and cold (La Niña) conditions

• Events tend to occur every 2-7 years • Generally episodes form during the spring or summer, peak

during the winter, and decay the following spring.• La Niña episodes can last multiple years (1-3 years). Less

common for El Niño, which last up to ~18months.• Often stronger El Niño events are followed by La Niña• Larger SST departures with strong El Niño episodes (relative

to strong La Niñas).

(1) ENSO is a stable, or damped, mode that requires stochastic (“weather”) forcing in order to occur.

- why ENSO events are so different from event to event.

- ENSO may have strict predictability limits because successful prediction would require skillful forecasts of the short-term atmospheric trigger

- implies a key role for shorter-term phenomenon such as the Madden Julian Oscillation (MJO). MJOs are often associated with “westerly wind bursts” that help drive the system forward.

Different Theories of ENSO

(2) ENSO is an unstable, naturally oscillatory mode that is self-sustained.

- Helps explain why ENSO has a 2-7 year period. ENSO is “made irregular” by weather perturbing this oscillation.

- implies that ENSO is predictable at longer lead times. Explains why models may be able to predict ENSO onset many months in advance.

Outline

(1) Brief Overview of the Ocean and Atmosphere (“Normal” vs. ENSO conditions) (2) The ENSO Cycle: A Coupled Ocean- Atmosphere System

(3) ENSO Teleconnections and Global Impacts

(4) ENSO Monitoring and Prediction at NOAA Climate Prediction Center (CPC)

ENSO Teleconnections

Schematic from Horel and Wallace (1981)

EXAMPLE:

Eastward expansion of warm sea surface temperatures during El Niño can result in an anomalous eastward shift of convection.

Enhanced thunderstorm activity in the central Pacific will perturb the upper-level flow resulting in an anticyclonic “couplet” straddling the equator.

Poleward of the ridge, an anomalous trough forms in the central North Pacific Ocean.

Tropical convection/heating can lead to “wavetrains” that can influence the global circulation.

Global El Niño Impacts

Impacts are generally more extensive during the northern winter.

Global La Niña Impacts

Mid-latitude impacts generally occur during the winter season (NH – DJF; SH- JJA).

Global ENSO Regression and Correlation Mapshttp://www.cpc.ncep.noaa.gov/products/precip/CWlink/MJO/enso.shtml#composite

• Gridded temperature anomalies (CPC GHCN) and precipitation anomalies (CPC Unified Precipitation) associated with the standardized Niño-3.4 index from 1948-2010.

• Assuming linearity so regression anomalies showing sign of El Niño (reverse for La Niña)

Outline

(1) Brief Overview of the Ocean and Atmosphere (“Normal” vs. ENSO conditions) (2) The ENSO Cycle: A Coupled Ocean- Atmosphere System

(3) ENSO Teleconnections and Global Impacts

(4) ENSO Monitoring and Prediction at NOAA Climate Prediction Center (CPC)

ENSO Monitoring and Prediction at NOAA Climate Prediction Center (CPC)

• NOAA CPC definitions for ENSO

• ENSO Alert System

• Forecasting ENSO and Model Skill

• Climate Change and ENSO

Evolution since 1950

El Niño

La Niña

neutral

• CPC provides weekly (every Monday) + monthly monitoring and prediction products for ENSO, which are available on our website:

http://www.cpc.ncep.noaa.gov/products/precip/CWlink/MJO/enso.shtmlIndices: http://www.cpc.ncep.noaa.gov/data/indices/

• The ENSO Diagnostics Discussion is released on the Thursday between the 4-10th of each month. Concurrent with that release, the “ENSO Alert System” and the official outlook is updated.

To receive monthly notification: ncep.list.enso-update@noaa.gov

Creation of the NOAA ENSO Outlook

El Niño or La Niña Watch: Favorable for development of ENSO within the next six (6) months.

El Niño or La Niña Advisory: conditions are observed and expected to continue.

Final El Niño or La Niña Advisory: conditions have ended.

NA: Not Active

El Niño conditions: one-month positive SST anomaly of +0.5 or greater in the Niño-3.4 region and an expectation that the 3-month ONI threshold will be met.

La Niña conditions: one-month negative SST anomaly of −0.5 or less in the Niño-3.4 region and an expectation that the 3-month ONI threshold will be met.

AND An atmospheric response typically associated with El Niño/ La Niña over the equatorial Pacific Ocean.

The ENSO Alert System is based on El Niño and La Niña “conditions” that allows the NOAA to be able to issue watches/ advisories in real-time.

The value of the ONI is to define episodes retrospectively.

What is the criteria for an ENSO Advisory?

Forecasting ENSO ENSO Forecasters rely on:

(1) Real-time data from the equatorial Pacific Ocean (collected from buoys, satellites, etc) and their knowledge of previous ENSO episodes

(2) Dynamical models: mathematical equations combined with current observations and run on a computer

- NCEP Climate Forecast System (CFS): a “coupled” computer model (ocean and atmosphere interact)

(3) Statistical models: use observations of the past to make predictions of the future

- Consolidated Forecast Tool (“CON”): statistically combines different models to take advantage of independent information provided by each model

Each forecaster individually provides probabilities of three categories (El Niño – Neutral – La Niña). Individual forecasts are averaged to create the “Consensus” probabilities and form the basis for the diagnostics discussion.

How is the probability of ENSO determined?

• Recently, dynamical models have slightly edged statistical models in forecast skill (see Barnston et al. BAMS, 2012)

• Models have trouble with transition timing and predicting amplitude of ENSO events.

• The transition to stronger ENSO events tends to be better predicted than transitions to weaker ones.

• “Spring prediction barrier:” historically, forecasts before the Northern Hemisphere Spring have low skill.

Primary features of ENSO model performance

Orange/Red Shading: Higher correlations (more skill)

White/Blue: Lower correlations ( 0 < r < 0.5)

Light Grey: Negative correlations (very poor skill!)

Prediction of Niño-3.4 Index by NCEP CFS from 2002-2011 (Post-processing/statistical corrections applied to model data after 2009)

From Barnston et al. (BAMS, 2012)

0

4

8

Lead Time

(mths)

Target (season you are predicting)

• Model skill is reduced during the N. Hemisphere spring when ENSO often emerges or decays

• CFS prediction improves to ~0.8 to 0.9 correlation for prediction of the N. Hemisphere winter (after ~June)

• RMSE (amount of error in amplitude) is ~0.5°C to 1.0°C in Niño-3.4

From Barnston et al. (BAMS, 2012)

The orange box designates the statistical models (the rest are dynamical)

Anomaly Correlations of ENSO models from 2002-2011 (from the IRI/CPC ENSO Prediction Plume)

• Skill for mid-year targets: Dynamical Models > Statistical models

-- Dynamical models have better initial conditions and ability to detect changes on shorter timescales than statistical models (often trained on monthly or seasonal data)

• For NH winter target, statistical and dynamical models are comparable.

• Short answer: Absolutely. There is uncertainty associated with seasonal forecasts even during strong ENSO episodes.

• This is why it is important to emphasize that the existence of ENSO means a “tilt in the odds” toward particular temperature/precipitation anomalies. These anomalies are never guaranteed, which is why climate outlooks are always probabilistic.

2-meter Temperature anomalies from Reanalysis data for Oct. and Dec. 2009 (during a moderate-strong El Niño).

This circulation and temperature pattern was reflective of a very strong negative Arctic Oscillation (AO) pattern.

L’Heureux et al. (2010, GRL)

Can other climate patterns overrule ENSO impacts?

See Ref. Collins et al. (2010), Vecchi and Wittenberg (2010), Guilyardi et al., (2009)

Will ENSO change due to Climate Change?

Models don’t agree on how ENSO changes.

ENSO feedbacks will likely change with warming, but hard to say which terms will dominate or cancel out.

IPCC-AR4: “No consistent indication at this time of discernible changes in projected ENSO amplitude or frequency in the 21st century.”

Continued ENSO variability in the future even with anthropogenic climate change

Summary

• ENSO is a naturally occurring coupled ocean-atmosphere phenomenon that has global impacts

• Equatorial Pacific fluctuates between warmer-than-average (El Niño) and colder-than-average (La Niña) conditions

• The changes in SSTs affect the distribution of tropical rainfall and atmospheric circulation features (Southern Oscillation)

• Changes in intensity and position of jet streams and storm activity occur at higher latitudes

• Monitoring and predicting ENSO is a key part of CPC’s monthly/seasonal temperature and precipitation outlooks

• Models do not agree on how ENSO will change with anthropogenic climate change.