Two Types of ENSO Varying in Tandem Facilitated by Nonlinear Atmospheric Convection

Abstract

The El Niño‐Southern Oscillation (ENSO) has two distinct regimes, one in the eastern Pacific (EP) featuring strong El Niño and the other in the central Pacific (CP) characterized by strong La Niña. Establishment of atmospheric convection in the climatologically cold and dry EP region provides a major source of the nonlinearity, but its role in the relationship between CP‐ and EP‐ENSO variability and their long‐term change remain unclear. By constructing a nonlinear two‐box recharge‐oscillator model, here we show that nonlinear atmospheric convection operates to boost strong EP El Niño, strong CP La Niña, and evolution in tandem of variability and skewness of CP‐ and EP‐ENSO. Further, the observed coherent long‐term increase in frequency of strong EP El Niño, strong CP La Niña, and variability of the two‐type ENSO over the past decades is facilitated by a corresponding increase in the nonlinear atmospheric convection.

Improving the Predictability of Two Types of ENSO by the Characteristics of Extratropical Precursors

“…Also, the correlations for all fields are higher for Niño3 than Niño4 in general, suggesting a closer connection of WWV with Niño3. This is consistent with previous studies reporting that EP ENSO is associated with stronger discharge/recharge processes than CP ENSO (e.g., Ren & Jin, 2013; Geng et al., 2020). However, we note that the characteristics of the two ENSO types are not easily distinguishable due to the high correlation between Niño3 and Niño4 (overlapping between the shading and contours).…”

“…(2019), both $${N}_{\text{EP}}$$ and $${N}_{\text{CP}}$$ show relatively lower skills compared to Niño3 and Niño4 indices, respectively, due to the spatial complexity of ENSO for $${N}_{\text{EP}}$$ and $${N}_{\text{CP}}$$ (Timmermann et al., 2018), respectively. According to the typical Recharge Oscillator framework (Jin, 1997), the system has only two degrees of freedom (SST and thermocline anomalies) for Niño3 and Niño4 indices (considering only SST anomalies), however, the system may include more than three degrees of freedom (further considering the spatial structure) for $${N}_{\text{EP}}$$ and $${N}_{\text{CP}}$$ (Geng et al., 2020). The reduction of ENSO prediction skill is not surprising due to the additional degrees of freedom.…”

“…However, the 𝐴𝐴 EPMTD prediction features a gradual decline with increasing lead months, particularly for EP type of ENSO (Figures 3a and 3c). (Jin, 1997), the system has only two degrees of freedom (SST and thermocline anomalies) for Niño3 and Niño4 indices (considering only SST anomalies), however, the system may include more than three degrees of freedom (further considering the spatial structure) for 𝐴𝐴 𝐴𝐴EP and 𝐴𝐴 𝐴𝐴CP (Geng et al, 2020). The reduction of ENSO prediction skill is not surprising due to the additional degrees of freedom.…”

Changing El Niño–Southern Oscillation in a warming climate

“…With understanding of ENSO diversity, it is possible to track observed changes in ENSO characteristics, providing important context to evaluate any future or projected changes. Since the late 1950s, CP and EP ENSO variability has increased [62][63][64] . Indeed, comparing the standard deviation of SST anomalies pre-1960 and post-1960 illustrates an approximately 20% increase in both EP and CP variability (fig.…”

Deep Residual Convolutional Neural Network Combining Dropout and Transfer Learning for ENSO Forecasting

“…However, predicting the strength of ENSO remains a challenge due to its complexity (Sun et al., 2016; Timmermann et al., 2018). Also, the increasing diversity of ENSO behavior since 2000 has led to a growing interest in the type of ENSO events (Geng et al., 2020). ENSO can be mainly divided into Eastern Pacific (EP) and Central Pacific (CP) types (Yeh et al., 2009), based on the distribution of the Sea Surface Temperature Anomaly (SSTA) during its maturation phase.…”