An ENSO stability analysis. Part II: results from the twentieth and twenty-first century simulations of the CMIP3 models

In this study, a Bjerknes stability (BJ) index, proposed by Jin et al. (2006), is adopted to assess the overall stability of El Niño and Southern Oscillation (ENSO) in state-of-the-art coupled models. The twentieth and twenty-first century simulations of 12 coupled models among the coupled model intercomparison project phase 3 models used in the intergovernmental panel on climate change forth assessment report demonstrate a significant positive correlation between ENSO amplitude and ENSO stability as measured by the BJ index. The simulations also show a diversity of behavior regarding the ENSO stability among the coupled models, which can be attributed to different mean state and sensitivity of an oceanic and atmospheric response to wind and SST forcing from model to model. When respective components of the BJ index obtained from the coupled models are compared with those from observations, it is revealed that most coupled models underestimate the thermodynamic damping effect and the positive effect of the zonal advective and thermocline feedback. Under increased CO2 induced warm climate, changes, relative to the twentieth century simulations, in the damping and feedback terms responsible for the ENSO stability measured by the BJ index can be linked to mean state changes and associated atmospheric and oceanic response sensitivity changes. There is a clear multi-model trend in the damping terms and positive zonal advective feedback, thermocline feedback, and Ekman feedback terms under enhanced greenhouse gas conditions. However, the various behavior among the coupled models in competition between the positive feedback and negative damping terms in the BJ index formula prevent the formation of a definitive conclusion regarding future projections of ENSO stability using the current coupled models.