ENSO Asymmetry in Amplitude and Duration in a Linear Model with State Dependent Noise

The El Niño / Southern Oscillation (ENSO) phenomenon exhibits asymmetry in both amplitude and duration. In particular, positive ENSO events (El Niños) tend to exhibit larger amplitude and have different duration (depending on how they are defined) than negative ENSO events (La Niñas). This asymmetry can be captured by a linear model driven by state dependent noise, notably “correlated additive and multiplicative” (CAM) noise. In this study, a Linear Inverse Model (LIM) with CAM noise (a CAM-LIM) is derived from historical observations. The CAM-LIM includes noise amplitudes that are stronger for El Niño events than La Niña events, which is sufficient to produce amplitude and duration asymmetry that is consistent with the observed record. The CAM-LIM framework provides a null hypothesis for ENSO asymmetry. Results from the model will be shown, as will direct calculations of noise forcing from the NCAR Large Ensemble (LENS) simulations. Implications for predictability of ENSO events will be discussed.