A new approach to reproduce the 20th century ENSO variability in an OAGCM

Mathieu Joly ( Météo-France / CNRM )

PhD-Thesis Director: S. Janicot ( IPSL / LOCEAN )

Research Guidance: J.-F. Royer & A. Voldoire ( Météo-France / CNRM )

Framework of this study (1)

SST - African monsoon interannual teleconnections

IPCC Coupled SimulationsIPCC Coupled SimulationsThe Pacific teleconnection looks quite different from one model to another.(Joly et al. 2007)

Météo-France CNRM-CM3 coupled model (IPCC-AR4 version)

ObservationsObservations (1901– 2000)(1901– 2000)

Leading mode of a Maximum Covariance Analysis of JJAS west African precipitation & tropical SSTs

CRU & HadISST

Framework of this study (2)

This could be because of:

• The timing of the simulated ENSO and / or the position of the SST anomalies in the Pacific.

• The atmospheric teleconnection that links the Pacific SST anomalies with the monsoon.

In this study, we try to improve the ENSO variability in the coupled model, in order to assess the consequences for the teleconnection.

IPCC Coupled SimulationsIPCC Coupled SimulationsThe Pacific teleconnection looks quite different from one model to another.(Joly et al. 2007)

The pacemaker experiment would be a good solution,but we have developped an other method:

A reanalyzed Wind-Stress is used here to drive the ocean, while keeping all other fluxes fully coupled.

ARPEGE-Climat

OPA Ocean Model

CN

RM

-CM

3 V

ers

ion

OA

SIS

Cou

plin

g

Wind-Stress forcing: How?

ERA40 dailyWind-Stress

Forcing

- Over all the oceans (GLOBE)

ERA40Wind-Stress

Buffer Zone- Over the tropical Pacific (PACIF)

Spin-up: 10 yrsSimulation: 1960-2001

SST biases

GLOBE

PACIF

JAS (Monsoon Season) SST Mean

HadISST

Coupled REF

SST St-Dev

SST annual cycles

Niño-4 Niño-3

ENSO areaENSO area

Gulf of Guinea

Gulf of GuineaGulf of Guinea

Anomalies

Does it work? Where?

Grid-point correlations between Observed and Simulated SST anomalies

JFM

GLOBE experiment

JAS

Filtered Niño 3.4 SSTSpectrum

ENSO variability

Filtered Niño 3.4 SST index Band-Pass filter: 17.5 months – 7.3 years

YEARS

El Niño SST composite

AMJJAS JFMOND

HadISST

AMJJAS JFMOND

PACIF

AMJJAS JFMOND

Coupled REF

AMJJAS JFMOND

GLOBE

La Niña SST composite

AMJJAS JFMOND

HadISST

AMJJAS JFMOND

PACIF

AMJJAS JFMOND

GLOBE

AMJJAS JFMOND

Coupled REF

Wind-Stress forcing: Synthesis

Using a reanalyzed Wind-Stress to drive the ocean in the Coupled Model allows us:

• To yield rather realistic ENSO composites

What consequences for the African monsoon?

• To reproduce quite accurately the observed 1960-2001 ENSO timing

• To improve the SST biases and annual cycles in the equatorial Pacific,but also in the Gulf of Guinea (GLOBE experiment)

African monsoon interannual variability

CRU 26%

GPCC 30%

GLOBE

35%

PACIF

32%

Coupled REF

45%JAS Precipitation EOF leading mode(with prior filtering of the decadal variability)

Links with SST?

Coupled REF

GLOBE

PACIFCRU

GPCC

Correlation of the leading EOF timeseries with the JAS SST anomalies(with prior filtering of the decadal variability)

ENSO-monsoon teleconnection

Correlation of the leading precipitation EOF timeserieswith the lagged Niño-3.4 SST indices

PACIF

AMJJAS JFMOND JAS

HadISST

AMJJAS JFMOND JAS

ENSO-monsoon teleconnection: Synthesis (1)

Coupled REF

AMJJAS JFMOND JAS

PACIF

AMJJAS JFMOND JAS

HadISST

AMJJAS JFMOND JAS

Coupled REF

AMJJAS JFMOND JAS

PACIF

AMJJAS JFMOND JAS

Coupled REF

AMJJAS JFMOND JAS

ENSO-monsoon teleconnection: Synthesis (2)

GLOBE

AMJJAS JFMOND JAS

?

Conclusion

ENSO – African monsoon teleconnections are poorly reproduced in IPCC simulations (Joly et al. 2007)

Using a reanalyzed Wind-Stress to drive the ocean in our coupled model allows us to reproduce the historical ENSO timing

A new approach to reproduce the 20th century ENSO variability in an OAGCM

But this protocole has a strong impact on the ENSO – monsoon teleconnection,which remains to be understood in terms of atmospheric mechanisms…

The idea of such experiments is similar to the "pacemaker" protocol:

• Thanks to the realistically simulated ENSO timing it is now possible to make year-to-year comparisons between simulations and observations

• Furthermore, forcing the wind-stress preserves the heat budget at the ocean surface, and partially preserves the ocean-atmosphere interactions

End of this talk…

… Thank you for your attention.