Evolution of the 2006-07 El Niño:

The Role of Intraseasonal to Interannual Time Scale Dynamics

Michael J. McPhadenNOAA/PMEL

Seattle, Washington

CLIVAR ENSO WorkshopGuangzhou, China

26-28 November 2007

Correo - Perú

2 August 2006

“ENSO-neutral conditions are expected to continue for the next one to three months, with a 50% chance that weak El Niño conditions will develop by the end of 2006.”NCEP, 10 Aug 2006

Niño-3.4 Predictions from July 2006 Initial Condition

Compiled by the International Research Institute for Climate and Society

“ENSO-neutral conditions are expected to continue for the next one to three months, with a 50% chance that weak El Niño conditions will develop by the end of 2006.”

NCEP, 10 Aug 2006

“Despite considerable efforts in model and forecast system development, progress in improving prediction skill has been very modest in recent years…”

Science, 15 Dec 2006

Build up of excess heat content along equator is a necessary precondition for El Niño to occur.

The time between El Niños is determined by the time to recharge.

El Niño purges excess heat to higher latitudes, which terminates the event.

Upper Ocean Heat Content (Based Recharge Oscillator Theory of Jin, 1997)

Warm Water Volume (WWV): An Index for Upper Ocean Heat

Content Meinen & McPhaden, 2000

WWV based on BMRC analysis of TAO/TRITON, XBT and WWV based on BMRC analysis of TAO/TRITON, XBT and Argo dataArgo data

2006-07 El Niño

January 2006-September 2007

NCEP Forecast, 9 Mar ‘06: “La Niña conditions are expected to continue

during the next 3-6 months.”

What happened in the next 3-6 months: Unexpected episodic relaxation of the trade winds triggers onset of El Niño

January 2006-September 2007

Aug-Oct ‘06: Strong westerly wind bursts and downwelling Kelvin waves

amplify warming.

Indian | Pacific | Atlantic

Mar 2006

Feb 2007 cloudy/wet

clear/dry

Cloudiness & Rainfall (OLR, 5°N-

5°S) Slow eastward progression of convection along the equator punctuated by episodic intraseasonal convective flair ups.Some of the intraseasonal variability originates over the Indian Ocean in the form of the “Madden- Julian Oscillation.”

Peak Conditions

Peak Conditions

January 2006-September 2007

Late Dec ‘06-Feb ‘07: Sudden demise linked to strengthening trade winds and rapid

thermocline shoaling.

NCEP Forecast, 7 Dec ‘06: “El Niño conditions are likely to continue through May 2007.”

Demise: Indian Ocean Influence?

Linear Equatorial Wave Model(McPhaden and Yu, 1999)

Purpose: Compare magnitude of Kelvin waves generated at the western boundary via Rossby wave reflection (“delayed oscillator”) vs directly wind force upwelling Kelvin wave

Daily ECMWF wind forcing, 1979-2007 Rectangular basin, 80°W-120°E

Kelvin plus 6 Rossby waves (long wave approximation)

4 vertical modes

Linear damping (12 mon time scale for first vertical mode)

Boundary reflections 80% efficient.

Anomalies relative to mean seasonal cycle

Detrend wind forcing and sea level output

Model and Observations SST Heat Content T/P-Jason SL Model SL TAO/TRITON

Model Wind-Forced and Boundary Generated Kelvin and Rossby

Waves Rossby Kelvin Kelvin Rossby Wind-Forced W. Boundary Wind-Forced E. Boundary

Summary The 2006-07 El Niño was a weak-to-moderate amplitude “dateline” event.

Ocean heat content provided 1-2 season lead time forecast skill during developmental phase. However, onset, amplitude, and demise were not well predicted.

Intraseasonal MJO was a significant source of high frequency atmospheric and oceanic variability and one of the factors confounding attempts to forecast the evolution of the event.

Strongest westerly wind bursts in Aug & Oct 2006 occurred after onset of warming, consistent with the idea of “state dependent stochastic forcing.”

Strongest MJO event occurred coincident with basin scale warming in December 2006 following the termination of the IOD event.

MJO-related easterlies in Dec 2006 forced an upwelling Kelvin wave response that abruptly accelerated the demise of the El Niño.

Conclusions

One path to improved ENSO forecast skill is to improve the representation of high frequency atmospheric variability, especially that related to the MJO, in forecast models.

It’s important to accurately represent tropical Indian Ocean variability and Indo-Pacific atmospheric teleconnections in seasonal forecasts models.