Andrew Turner ([email protected]) Pete Inness, Julia Slingo

The effect of climate change and systematic model bias on the monsoon-ENSO system: the

TBO and changing ENSO regimes

Andrew Turner ([email protected])

Pete Inness, Julia Slingo

Outline

• Basis for study & heat flux adjustments.

• The effect of systematic bias in the future.

• Changing ENSO regimes and their properties.

• Why the eastward tendency?

• TBO in the model.

• Monsoon predictability and the observed record.

Background

• GCMs which generate mean climatologies close to observations are more likely to correctly represent interannual variability of precipitation. (Sperber & Palmer 1996).

• Equatorial Pacific SSTs important predictor of monsoon (Charney & Shukla 1981).

• Heat flux adjustments used at 1xCO2 to correct systematic biases (Turner et al. 2005).

• Systematic model errors likely to have large impact on future predictions.

• Same adjustments used at 2xCO2.

Annual mean heat flux into ocean

Peak-to-peak amplitude

Nino-3 region SSTA in the Flux adjusted model at 2xCO2

Nino-3 region SST annual cycle

Nino-3 variance cycle

Lag-correlation of TransNinoIndex with Nino3 SSTA, 12yr moving window

Possibilities for tendency to T-modes

• Weakening of trades allows deepening of thermocline in east Pacific (Federov & Philander 2001) dominance of vertical movements over zonal advection.

• Wang & An 2002 see similar pattern to 2xCO2 or FA response in 1976 climate shift, and with increased mean upwelling in east, decreased zonal temperature gradient.

Webster Yang South Asia DMI (JJAS)

Tendency to biennial period

• Subsurface Kelvin wave propagation more prevalent during biennial regime.

• Meridional confinement of wind stress response to ENSO (Kirtman 1997; Guilyardi et al. 2004) shortens period.

JJAS DMI vs AIR:

Irregular cc=0.31

Biennial cc=0.71

Summary

• Tendency to T-mode ENSO (together with associated Kelvin-wave behaviour at depth) when FA applied / CO2 doubled.

• Lack of broad meridional wind stress response in HadCM3 shortens period (only comes to light during strong T-mode behaviour).

Issues

• More understanding needs to be gained into the impact of different ENSO types on the monsoon system.

• Is modelled response to T-mode related to bienniality and/or incorrect representation in GCM?