El Niño Southern Oscillation in a Changing Climate

Abstract

The El Niño Southern Oscillation (ENSO) is characterized by being irregular or nonperiodic and asymmetric between El Niño and La Niña with respect to amplitude, pattern, and temporal evolution. These observed features suggest the importance of nonlinear dynamics and/or stochastic forcing. Both nonlinear deterministic chaos and linear dynamics subject to stochastic forcing and/or to non‐normal growth were introduced to explain the irregularity of ENSO, but no consensus has been reached to date given the short observational record. As a dominant source of stochastic forcing, westerly wind bursts play a role in triggering, amplifying, and determining the irregularity and asymmetry of ENSO, which are best treated as part of the deterministic dynamics or as a multiplicative noise forcing. Various nonlinear processes are responsible for the spatial and temporal asymmetry of El Niño and La Niña, which includes nonlinear ocean advection, nonlinear atmosphere‐ocean coupling, state‐dependent stochastic noise, tropical instability waves, and biophysical processes. In addition to the internal nonlinear processes, a capacitor effect of the Indian and Atlantic Oceans and atmospheric and oceanic teleconnections from extratropical Pacific could also contribute to the temporal and amplitude asymmetry of ENSO. Despite significant progress, most state‐of‐the‐art models are still lacking in simulation of the spatial and temporal asymmetry of ENSO. 1 Department of Atmospheric Sciences, Yonsei University, Seoul, Republic of Korea 2 Department of Earth and Planetary Sciences and School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA 3 Institute for Geophysics, Jackson School of Geosciences, The University of Texas at Austin, Austin, TX, USA 4 Department of Atmospheric Sciences/IPRC, University of Hawai’i at Ma ̄noa, Honolulu, HI, USA 154 EL NIÑO SOUTHERN OSCILLATION IN A CHANGING CLIMATE determine an atmosphere‐ocean coupled stability for ENSO system (T. Li, 1997b; An & Jin, 2000; Fedorov & Philander, 2000), and for example, depending on the coupling strength, ENSO system becomes a self‐sustained and possibly chaotic oscillator under a strong coupling and a damped oscillator under a weak coupling (An & Jin, 2001). It has been suggested that some decades may be characterized by a self‐sustained, possibly chaotic dynamics, while others show a damped ENSO cycle, excited by stochastic variability (Kirtman & Schopf, 1998). However, a bifurcation between stable and unstable regimes tends to be ambiguous in the presence of noise (e.g., Levine & Jin, 2010). Westerly wind bursts (WWBs) are episodic reversals of the equatorial trade winds with a strength of 5 to 7 ms–1, zonal extent of 20–40 degrees, duration of 5–30 days, and frequency of around 5 to 10 times per year (Harrison & Vecchi, 1997; L. Yu et al., 2003; Seiki & Takayabu, 2007a). These events, a dominant source of stochastic forcing, play a role in triggering, amplifying, and even determining the spatial pattern of ENSO events (Harrison & Vecchi, 1997; Eisenman et al., 2005; Levine & Jin, 2010; Rong et al., 2011; D. Chen et al., 2015; Hayashi & Watanabe, 2017). WWBs were initially considered as additive stochastic forcing (e.g. Moore & Kleeman, 1999), yet it became clear that they depend on the background SST and tend to occur more frequently during a developing El Niño (Verbickas, 1998; L. Yu et al., 2003; Eisenman et al., 2005). These events are thus best treated as part of the deterministic dynamics or as a state‐dependent multiplicative noise forcing, with important implications to amplitude and predictability of El Niño events. El Niño is not a simple mirror image of its opposite phase, La Niña. El Niño’s amplitude is on average greater than that of La Niña (Deser & Wallace, 1987; Burgers & tephenson, 1999; An & Jin, 2004). El Niño is often followed by a La Niña in the following year, but the opposite is much less common (Larkin & Harrison, 2002; M. Chen et al., 2016; An & Kim, 2017). After their mature phase, many La Niñas persist through the following year, but most of El Niños tend to decay rapidly by next summer (Ohba & Ueda, 2007; Okumura & Deser, 2010; Choi et al. 2013; DiNezio & Deser, 2014; An & Kim, 2018). Strong El Niños are mainly loaded over the eastern Pacific with focusing toward the equator, whereas strong La Niñas are mostly loaded over the central Pacific with a wider latitudinal extension (Hoerling et al., 1997; Kang & Kug, 2002; Takahashi et al., 2011; Dommenget et al., 2013). Such amplitude/duration/transition/pattern asymmetries between El Niño and La Niña may not be surprising given the nonlinear internal dynamics and/or selective external impacts (e.g., An & Kim, 2018). Asymmetrical internal nonlinear processes that are responsible for amplitude asymmetry include the vertical ocean temperature profile (Zebiak & Cane, 1986), ocean nonlinear advection (An & Jin, 2004; Su et al. 2010), asymmetric equatorial wind response to SST (Kang & Kug, 2002; Frauen & Dommenget, 2010; Choi et al., 2013), ocean wave response to the wind stress (An & Kim, 2017, 2018), outcropping thermocline nonlinearity (Battisti & Hirst, 1989; Galanti et al., 2002; An & Jin, 2004), state‐dependent stochastic forcing (Jin et al., 2007; Kug et al., 2008; Rong et al., 2011; Levine et al., 2016; Hayashi & Watanabe, 2017), tropical instability wave activity (J. Yu & Liu, 2003; An, 2008a, 2008b), biophysical feedback (Timmermann & Jin, 2002), shortwave feedback (Lloyd et al., 2012), etc. Transition/duration asymmetry has been attributed to a selective capacitor effect of the Indian and Atlantic oceans (Ohba & Ueda, 2007; Okumura & Deser, 2010; An & Kim, 2018), development of subtropical western Pacific atmospheric circulation during decaying phase of ENSO to boost ENSO transition (B. Wang et al., 1999; B. Wang et al., 2001; Y. Li et al., 2007; B. Wu et al., 2010a), and some of aforementioned internal nonlinear processes (Choi et al., 2013; Im et al., 2015; M. Chen et al., 2016; An & Kim, 2017, 2018; M. Chen & Li, 2018). This chapter focuses on the irregularity of ENSO and on its amplitude and evolution asymmetries. In section 7.2, the origin of irregularity will be addressed together with the role of westerly wind burst events. Mechanisms for amplitude asymmetry will be discussed in section 7.3. The cause of evolution asymmetry will be reviewed in section 7.4, and we include conclusion and discussion in section 7.5.