Interactions between the Madden- Julian Oscillation and the North Atlantic Oscillation Hai Lin, Gilbert Brunet Meteorological Research Division, Environment.

Interactions between the Madden-Julian Oscillation and the North Atlantic Oscillation

Hai Lin, Gilbert BrunetMeteorological Research Division, Environment Canada

Jacques DeromeMcGill University

TTISS, Monterey, September 14, 2009

Outlines

• Observed MJO – NAO connection

Lin et al. 2009 (J. Climate)

• Intraseasonal variability in a dry GCM

Lin et al. 2007 (J. Atmos. Sci.)

NAO and MJO connection

• NAO: dominant large scale pattern in the extratropics with significant influence on weather from eastern North America to Europe

• MJO: dominant tropical intraseasonal mode, coupled with convections and variability in diabatic heating

• One-way impact, or two-way interaction?

• A possible mechanism for both the NAO and MJO

Data and methodology

Definition of the NAO: 2nd REOF of monthly Z500

NAO index: projection of pentad Z500 anomaly onto this pattern

Period: 1979-2003

Extended winter, November to April (36 pentads each winter)

Data and methodology

Definition of the MJO: combined EOF of OLR, u200 and u850 in the band of 15°S – 15°N (Wheeler and Hendon, 2004)

NAO index: RMM1 and RMM2

Period: 1979-2003

Extended winter, November to April (36 pentads)

Composites of tropical

Precipitation rate.

Winter half year

November-April

Xie and Arkin pentad data, 1979-2003

Pentads in MJO phases

Extended winter from 1979 to 2004

Phase 1 2 3 4 5 6 7 8

Number of pentads

55 79 78 78 63 71 87 66

Mean amplitude

1.67 1.66 1.81 1.78 1.66 1.70 1.62 1.75

Lagged composites of the NAO index

Phase 1 2 3 4 5 6 7 8

Lag −5 −0.39 0.28

Lag −4 −0.26 0.28

Lag −3 −0.29

Lag −2 0.26

Lag −1

Lag 0 −0.41

Lag +1 0.26 0.27 0.26 −0.25 −0.35

Lag +2 0.34 0.36 −0.31 −0.33 −0.29

Lag +3 0.35 −0.35 −0.41

Lag +4 −0.35 −0.31

Lag +5 −0.27

Lagged composites of the NAO index

Phase 1 2 3 4 5 6 7 8

Lag −5 −0.39 0.28

Lag −4 −0.26 0.28

Lag −3 −0.29

Lag −2 0.26

Lag −1

Lag 0 −0.41

Lag +1 0.26 0.27 0.26 −0.25 −0.35

Lag +2 0.34 0.36 −0.31 −0.33 −0.29

Lag +3 0.35 −0.35 −0.41

Lag +4 −0.35 −0.31

Lag +5 −0.27

Lagged probability of the NAO indexPositive: upper tercile; Negative: low tercile

Phase 1 2 3 4 5 6 7 8

Lag −5 −35% −40% +49% +49%

Lag −4 +52% +46%

Lag −3 −40% +46%

Lag −2 +50%

Lag −1

Lag 0 +45% −42%

Lag +1 +47% +45% −46%

Lag +2 +47% +50% +42% −41% −41% −42%

Lag +3 +48% −41% −48%

Lag +4 −39% −48%

Lag +5 −41%

Tropical influence

Wave activity flux and 200mb streamfunction anomaly

Lagged regression of 200mb U to NAO index

Extratropical influence

Lagged regression of 200mb U to NAO index

Extratropical influence U200 composites

Tropical intraseasonal variability (TIV) in a dry GCM

Model and experiment

• Primitive equation AGCM (Hall 2000).

• T31, 10 levels

• Time-independent forcing to maintain the winter climate (1969/70-98/99) all variabilities come from internal dynamics

• No moisture equation, no interactive convection

• 3660 days of integration

Unfiltered data

20-100 day band-pass

Zonal propagation

10S-10N

Model Result

Stronger in eastern Hemisphere

TIV in the dry model

• Kelvin wave structure

• Phase speed: ~15 m/s (slower than free Kelvin wave, similar to convective coupled Kelvin wave, but there is no convection)

What causes the TIV in the dry model?

• 3-D mean flow instability (Frederiksen and Frederiksen 1997)

• Tropical-extratropical interactions (all wave energy generated in the extratropics)

Moisture and convection related mechanisms are excluded

Possible mechanisms

ISO in a dry model

Linked to tropical eastward propagation in the eastern Hemisphere Global propagation of low-frequency wave activity

250 hPa PV’ and wave activity flux

Summary

• Two-way interaction between the MJO and the NAO

• Increase of NAO amplitude 5~15 days after the MJO-related convection anomaly reaches western Pacific

• Certain MJO phases are preceded by strong NAOs

• TIV generated in a dry GCM

• Tropical-extratropical interactions are likely responsible for the model TIV

Implication to the MJO

• A possible mechanism for the MJO: triggering, initialization

• Contribution of moisture and tropical convection: spatial structure, phase speed