A. G. Turner * , P. M. Inness & J. M. Slingo

1. 1 Background: Integrations of the Met Office model HadCM3 at 1xCO2 show systematic model biases to have a detrimental impact on the monsoon-ENSO system and its predictability (Turner et al. 2005). Such systematic biases could cause uncertainties in future climate simulations (Federov & Philander 2000). An annual cycle of heat flux adjustments (FA) is applied to the equatorial Indo-Pacific in a 2xCO2 run of HadCM3 to reduce the biases. This reveals a greater climate change response than the uncorrected model (Turner et al. 2006). Additionally, HadCM3FA at 2xCO2 features distinct ENSO regimes (Fig. 1), which feature several similarities with theoretical studies (Jin 1997; Wang et al. 1999). The irregular regime features strong stochastic forcing whilst the biennial events are part of a self-sustained limit cycle in a more strongly coupled ocean-atmosphere system. The effect of doubled CO2 and model basic state biases on the monsoon-ENSO system: the TBO and changing ENSO regimes Fig. 1: Niño-3 SSTA in HadCM3FA 2xCO2. Irregular and biennial ENSO regimes are selected for comparison. 2 Regime characteristics: Similarities can also be drawn between the regimes and observed ENSO modes. The irregular regime features mild El Niño events located centrally and lasting several years (Fig. 2), akin to local-modes in observations (e.g. 2002). In the biennial regime, ENSO dominates over the annual cycle of cold upwelling in the east (Figs. 2,3), and more evidence of Kelvin wave propagation is revealed (Fig. 2), similar to observed basinwide thermocline modes (e.g. 1997). Central Pacific El Niño Fig. 4: Annual mean biennial minus irregular SST composite. Speckling indicates 95% significance. Annual cycle and ENSO compete to dominate East Pacific. Fig. 3: Normalized power spectra of Niño-3 SST in HadCM3FA 2xCO2. A. G. Turner*, P. M. Inness & J. M. Slingo Fig. 5: Lag correlation of Niño-3 SST with Trans-Niño Index in HadCM3(FA) integrations at 1x, 2xCO2. Basinwide eastward propagation The tendency toward eastward propagating basinwide modes can be related to slackening of the zonal SST gradient (Fig. 4), favouring vertical motion of the thermocline over zonal advection of SST anomalies (Federov & Philander 2001). This tendency is increased both by 2xCO2 forcing and flux adjustment (Fig. 5, shown using the Trans-Niño Index of Trenberth & Stepaniak 2001). Fig. 2: Equatorial Pacific 20ºC isotherm depth anomalies (dam), as a proxy for heat content. 3 Reasons for the biennial tendency: The biennial tendency in HadCM3 2xCO2 is in contrast with observed basinwide El Niño events which are often of 4-5 year period (Guilyardi 2006). Its cause can be understood by considering interannual variability in the Asian summer monsoon, to which ENSO is tied. The dynamical monsoon index (Webster & Yang 1992) allows selection of strong (weak) years of the tropospheric biennial oscillation (TBO), i.e., years which are stronger (weaker) than preceding and following years (Fig. 6.). A composite evolution of strong minus weak seasons shows strong coupling between the Asian monsoon and the Indian Ocean dipole and ENSO (Fig. 7). The decay of the Indian Ocean dipole to a basinwide anomaly of the same sign as ENSO is essential in reversing ENSO phase in the Pacific (Kug and Kang 2006). Strong monsoon forcing is increased by both 2xCO2 and flux adjustment, acting to increase coupling between the Indian and Pacific Oceans. Fig. 7: Strong minus weak composite evolution of precip (contours), SST (shading) and 850hPa winds. Fig. 6: Dynamical monsoon index of HadCM3FA 2xCO2. 4 Implications: Whilst the presence of biennial oscillations of the monsoon-ENSO system and strong coupling of the Indo-Pacific may suggest increased predictability, such extreme seasons are detrimental to the populations of South Asia. The existence of regime changes on interdecadal timescales suggests there may be greater uncertainty in projections of future climate. 5 References: Federov & Philander (2000) Science288; Federov & Philander (2001) J. Clim.14; Jin (1997) J. Atmos. Sci.54; Guilyardi (2006) Clim. Dyn. 26; Kug & Kang (2006) Geophys. Res. Lett.33; Trenberth & Stepaniak (2001) J. Clim. 14; Turner, Inness & Slingo (2005) Q. J. R. Meteorol. Soc.131; Turner, Inness & Slingo (2006) Submitted, Q. J. R. Meteorol. Soc.; Wang, Barcilon & Fang (1999) J. Atmos. Sci.56; Webster & Yang (1992) Q. J. R. Meteorol. Soc. 118. * A.G. Turner is supported via an EU-ENSEMBLES grant at NCAS-Climate, University of Reading, UK a.g.turner@rdg.ac.uk