Abstract

Observed SST anomaly (SSTA) in the equatorial eastern Pacific exhibits an asymmetric evolution characteristic between El Niño and La Niña. While El Niño is characterized by a rapid decay after its peak and a fast phase transition to a cold episode in the following winter, La Niña is characterized by a weaker decay after its peak and a reintensification of cold SSTA in the second year. The relative roles of dynamic (wind field) and thermodynamic (heat flux) processes in causing the asymmetric evolutions are investigated through a mixed layer heat budget analysis. The result shows both dynamic and thermodynamic processes contribute to the evolution asymmetry. The former is related to asymmetric wind responses in the western Pacific, whereas the latter is associated with asymmetric cloud–radiation–SST and evaporation–SST feedbacks. A strong negative SSTA tendency occurs during El Niño decaying phase, compared to a much weaker positive SSTA tendency during La Niña decaying phase. Such a difference leads to an SSTA sign change for El Niño but no sign change for La Niña by the end of summer of the second year. A season-dependent coupled instability kicks in during northern fall, leading to the development of a La Niña by end of the second year for El Niño, but the reoccurrence of a La Niña episode by end of the second year for La Niña. The overall heat budget analysis during the entire ENSO evolutions indicates the thermodynamic process is as important as the dynamic process in causing the El Niño–La Niña evolution asymmetry. The fundamental difference of the current result with previous theories is further discussed.

1. Introduction

El Niño–Southern Oscillation (ENSO) is the largest interannual variability in the tropics, and its fundamental dynamics (including its structure and evolution characteristics) have been studied a great deal in past decades (e.g., Rasmusson and Carpenter 1982; Cane and Zebiak 1985; Philander 1990; Neelin et al. 1998; Chang et al. 2006). Now it is widely accepted that ENSO amplification results from coupled ocean–atmosphere instability (Bjerknes 1969; Philander et al. 1984; Hirst 1986, 1988), and its oscillatory feature arises from either delayed oscillation dynamics in which the ocean waves play a critical role (e.g., Suarez and Schopf 1988; Battisti and Hirst 1989) or a discharge–recharge scenario in which zonal mean thermocline depth variation holds a key (Jin 1997; Li 1997).

An important dynamics question related to ENSO is what causes the amplitude asymmetry between El Niño and La Niña. A positive skewness of the SST anomaly (SSTA) was found in the equatorial eastern Pacific (EP; Burgers and Stephenson 1999; An and Jin 2004). A recent observational study by Su et al. (2010), using three sets of oceanic reanalysis data, showed that the positive skewness of ENSO-related SSTA in the equatorial EP arose from nonlinear zonal and meridional advection while nonlinear vertical advection played an opposite role. This result is opposite to An and Jin (2004), who emphasized the role of nonlinear vertical advection. The discrepancy between the two works above is attributed to ocean reanalysis data used. An and Jin (2004) used a test (beta) version of the Simple Ocean Data Assimilation (SODA) data that has a bias in the vertical velocity field (Su et al. 2010).

Compared to the amplitude asymmetry, the temporal evolution asymmetry between El Niño and La Niña has received less attention (Kessler 2002; Larkin and Harrison 2002; McPhaden and Zhang 2009; Kim et al. 2011). Kessler (2002) noticed that transitions from La Niña to El Niño are usually slower than those from El Niño to La Niña. Figure 1 shows the time series of the observed SSTA averaged in the Niño-3.4 region (5°N–5°S, 170°–120°W) during 1980–2013. Note that almost all El Niño events (except 1986/87 El Niño) terminated rapidly after their peak phase in boreal winter, whereas almost all La Niña events (except 1988 and 2005 La Niña) redeveloped into another La Niña event in the subsequent boreal winter.

Fig. 1.

Time series of SSTA (°C) in the Niño-3.4 region (colored bars) and MLTA in the equatorial EP region (5°N–5°S, 180°–80°W; solid curve) during 1980–2013.

Fig. 1.

Time series of SSTA (°C) in the Niño-3.4 region (colored bars) and MLTA in the equatorial EP region (5°N–5°S, 180°–80°W; solid curve) during 1980–2013.

Various studies have been devoted to the mechanism of understanding this El Niño and La Niña evolution asymmetry. By specifying tropical Indian Ocean (IO) SSTAs in a global coupled general circulation model (GCM), Ohba and Ueda (2007) showed that IO basin warming during El Niño peak winter can induce anomalous surface easterlies in the far equatorial western Pacific (WP), which accelerates a transition from El Niño to La Niña. Based on this modeling result, Okumura and Deser (2010) hypothesized that the El Niño and La Niña evolution asymmetry results from a combined effect of IO forcing and longitudinal shifting of Pacific heating anomalies associated with El Niño and La Niña; that is, whereas easterly anomalies in the equatorial WP induced by IO basin warming are canceled out by westerly anomalies induced by El Niño heating, westerly anomalies induced by IO basin cooling are only partly offset by La Niña heating–induced easterly anomalies because SSTA center associated with La Niña shifts farther toward the west compared to El Niño. Forcing an atmospheric GCM with specified SSTA, Okumura et al. (2011) suggested that IO and WP SSTA forcing is important in generating asymmetric zonal wind response in the far equatorial WP.

In both Ohba and Ueda (2007) and Okumura et al. (2011), the SSTA in IO is specified. In response to the basinwide IO SSTA forcing, a basinwide ascending motion or precipitation (convective heating) anomaly is generated [see Ohba and Ueda (2007), their Fig. 2c and Okumura et al. (2011), their Fig. 3, left], which contradicts the observed rainfall anomaly pattern in IO. Observations show a zonal dipole rather than a uniform pattern over IO (Wu et al. 2012; also see detailed discussion of this subject in section 5). This implies that the IO SSTA forcing effect illustrated by Ohba and Ueda (2007), Okumura and Deser (2010), and Okumura et al. (2011) is physically incorrect. The failure is attributed to the model experiment design because in a forced atmospheric model experiment with a specified SSTA, the SSTA plays an active role. However, in reality during El Niño mature winter, only the SSTA in western IO plays an active role (i.e., a positive SSTA is in phase with a positive precipitation anomaly), while the SSTA in the eastern IO plays a passive role (i.e., a positive SSTA is in phase with a negative precipitation anomaly). Thus the warming in the eastern IO is a result of atmospheric forcing, because of reduced precipitation (or anomalous subsidence) and thus increased surface shortwave radiation forcing (Wu et al. 2012).

McGregor et al. (2013) suggested that the southward shift of westerly anomalies during the mature phase of ENSO, which is more pronounced during strong El Niño but much less so in weak El Niño and La Niña, might contribute to the El Niño and La Niña evolution asymmetry. This zonal wind shifting was noted previously by Harrison and Vecchi (1999), who found a transition of zonal wind anomaly from an equatorially symmetric state during the earlier developing stage to an asymmetric state during El Niño mature phase. It was hypothesized that the southward shift of zonal wind anomaly in the central Pacific (CP) may hasten the El Niño–La Niña transition through an enhanced delayed oscillator oceanic wave effect (Harrison and Vecchi 1999). The cause of the wind shift was possibly related to the seasonal change of the mean state (Harrison and Vecchi 1999) or zonal wind acceleration due to reduced boundary layer friction coefficients (McGregor et al. 2012).

A problem with the hypothesis of McGregor et al. (2013) is that a transition from a warm to a cold episode happens not only to super El Niños but also to regular El Niños. In addition, shallow water model experiments in McGregor et al. (2013) contained the effect of not only asymmetric wind in CP but also asymmetric circulation in WP such as anomalous Philippine Sea anticyclone (PSAC) during El Niño, the latter of which has been emphasized by previous literature for its role in El Niño turnabout (e.g., Wang et al. 2000) and El Niño–La Niña evolution asymmetry (Wu et al. 2010a). Thus, it is unclear how important the asymmetric zonal wind shifting in CP is in causing the life evolution asymmetry between El Niño and La Niña.

In almost all previous El Niño–La Niña evolution asymmetry studies, emphasis was put on the dynamic effect of surface wind asymmetry. As shown by atmospheric modeling studies (e.g., Kang and Kug 2002; Ohba and Ueda 2009; Takahashi et al. 2011; Dommenget et al. 2013), there is a marked asymmetry (or nonlinearity) in basinwide circulation response to El Niño and La Niña forcing. Such a wind asymmetry affects not only the ocean dynamic processes (such as thermocline evolution and temperature advection) but also surface heat flux fields. The objective of the present study is to reveal the relative roles of dynamic (wind field) and thermodynamic (heat flux) forcing effects in causing the El Niño and La Niña evolution asymmetry. To achieve this goal, a quantitative analysis of each of the dynamic and thermodynamic terms is needed. In this study, a mixed layer heat budget analysis is carried out to examine specific dynamic and thermodynamic processes that give rise to the asymmetric El Niño and La Niña evolutions. In particular, we intend to address the following scientific questions: Why does El Niño decay more quickly than La Niña in the northern spring and summer seasons following their mature phase? What are the fundamental dynamic and thermodynamic processes responsible for the asymmetric damping rates during the El Niño and La Niña decaying phase? What causes the quick transition of an El Niño into a La Niña late in the second year, whereas La Niña tends to redevelop?

The remaining part of this paper is organized as follows. In section 2 we describe the datasets and analysis methods used. Section 3 shows the asymmetric evolution features between composite El Niño and La Niña. The physical causes of the evolution asymmetry between El Niño and La Niña are explored in section 4, based on a mixed layer heat budget analysis and the diagnosis of relevant dynamic and thermodynamic fields. A discussion of how and to what extent the IO SSTA influences the WP circulation is given in section 5. Finally, conclusions are given in the last section.

2. Data and methods

The primary datasets used in this study are the ocean reanalysis datasets derived from the National Centers for Environmental Prediction (NCEP) Global Ocean Data Assimilation System (GODAS; Saha et al. 2006) and University of Maryland SODA (Carton and Giese 2008), version 2.1.6 products (SODAv2.1.6). The GODAS product has an average 1° × 1° horizontal and 40-level vertical resolution with meridional resolution enhanced to ⅓° within 10°S–10°N, and it covers the period from 1980 to present. While the SODA dataset has an average 0.4° × 0.25° horizontal and 40-level vertical resolution with 10-m spacing in the upper ocean, and it is available from 1958 to 2008 (Smith et al. 1992).

The atmospheric and surface heat flux data are from NCEP–DOE AMIP-II reanalysis (NCEP-2; Kanamitsu et al. 2002) and the Woods Hole Oceanographic Institution (WHOI) objectively analyzed air–sea fluxes (OAFlux; Yu et al. 2008). The NCEP-2 dataset covers the period from 1979 to present, while the OAFlux data are available from 1984 to 2009. SST data are from the Extended Reconstructed Sea Surface Temperature, version 3b (ERSST.v3b), with a resolution of 2° × 2° (Smith et al. 2008). Observed outgoing longwave radiation (OLR) data are from National Oceanic and Atmospheric Administration (NOAA) at a resolution of 2.5° × 2.5° (Liebmann and Smith 1996).

To understand the relative roles of ocean dynamics (i.e., three-dimensional temperature advection) and surface heat fluxes in causing the SSTA evolution asymmetry between composite El Niño and La Niña episodes, we analyze the oceanic mixed layer heat budget. The mixed layer temperature anomaly (MLTA) tendency equation may be written as follows:

 
formula

where T denotes the mixed layer temperature; u, υ, and w represent three-dimensional (3D) ocean currents; ∂/∂x, ∂/∂y, and ∂/∂z denote the 3D gradient operator; a prime represents the interannual anomaly; a bar represents the climatological mean state; and the first nine terms on the right-hand side of the equation are 3D temperature advection terms (Hong et al. 2008). The variable Qnet denotes the ocean net surface heat flux (with a positive sign representing that the ocean receives heat), R denotes the residual term, ρ is the density of water (=103 kg m−3), CP is the specific heat of water (=4000 J kg−1 K−1), and H denotes the mixed layer depth, which is specified from the longitude-dependent climatological mean field (ranged from 20 to 90 m, derived from the ocean reanalysis data). All of the mixed layer terms in the equation above are calculated based on the vertical average within the mixed layer. Because the datasets mentioned above have different periods, in order to make a composite for the same period, all composite analyses are done for the period of 1980–2013. The mixed layer heat budget analysis results presented in this paper are the ensemble average of two ocean reanalysis datasets (GODAS and SODAv2.1.6) and two heat flux products (NCEP-2 and OAFlux).

3. Observed characteristics of evolution asymmetry between El Niño and La Niña

Because SSTAs associated with El Niño and La Niña are primarily confined in the equatorial EP, we examine the evolution of the composite MLTAs averaged over 5°N–5°S, 180°–80°W. The solid line of Fig. 1 shows that the MLTAs are consistent with the SSTAs. This gives us confidence to conduct the mixed layer heat budget analysis. For the period of 1980–2013, eight representative El Niño cases (1982/83, 1991/92, 1994/95, 1997/98, 2002/03, 2004/05, 2006/07, and 2009/10) and five representative La Niña cases (1983/84, 1995/96, 1998/99, 2007/08, and 2010/11) were selected for the subsequent composite analysis. Because the SODAv2.1.6 data end in 2008, only the GODAS dataset is used for the 2009/10 El Niño and 2010/11 La Niña cases.

The temporal evolutions of the MLTAs for both the El Niño and La Niña composites are displayed in Fig. 2. The El Niño and La Niña episodes bear many similarities during the developing year (year 0). For the composite El Niño, the positive MLTA starts to develop in April and reaches its peak in December. For the composite La Niña, the negative MLTA starts to develop in June and reaches its peak in December as well. The major evolution asymmetry happens in the second year (year +1). Following its peak phase, El Niño experiences a fast decay, and a negative MLTA occurs in July. In contrast, La Niña decays at a much slower rate, and by July of year +1 it still retains one-third of the peak MLTA value. During the northern fall of year +1, the cold MLTA rapidly reintensifies. As a result a La Niña reemerges in the following winter.

Fig. 2.

Composite temporal evolutions of MLTAs (°C) for selected (bars) and all (green curve) (a) El Niño and (b) La Niña events in the equatorial EP region (5°N–5°S, 180°–80°W). Red bars represent El Niño (La Niña) developing phase, blue bars represent El Niño (La Niña) decaying phase, and orange bars represent El Niño transition (La Niña reintensification) phase.

Fig. 2.

Composite temporal evolutions of MLTAs (°C) for selected (bars) and all (green curve) (a) El Niño and (b) La Niña events in the equatorial EP region (5°N–5°S, 180°–80°W). Red bars represent El Niño (La Niña) developing phase, blue bars represent El Niño (La Niña) decaying phase, and orange bars represent El Niño transition (La Niña reintensification) phase.

To illustrate that the evolutions above reflect the common features of ENSO, we also plotted the MLTA evolutions for all ENSO case composites, which include the special 1986/87 El Niño and 1988 and 2005 La Niña cases (see green lines in Fig. 2). As one can see, the two MLTA time series bear a great similarity, indicating that the asymmetric evolution feature between El Niño and La Niña is quite robust.

As shown in Fig. 2, a key difference between El Niño and La Niña evolutions lies in the distinctive MLTA decaying rates during the earlier part of year +1; that is, El Niño decays much quicker than La Niña after their peak phase. Because of this difference, the positive MLTA associated with El Niño has changed its sign (becoming a cold anomaly) by July of the second year, while the MLTA associated with La Niña keeps the same sign. We will address the dynamic and thermodynamic causes of the distinctive damping rates and examine the consequence of this difference in section 4.

To clearly demonstrate the roles of dynamic and thermodynamic processes in causing the evolution asymmetry, we define three phases for El Niño: April–November of year 0 (0 or the first year) as El Niño developing phase, January–May of year +1 (+1 or the second year) as El Niño decaying phase, and July–November of year +1 as El Niño–La Niña transition phase. Similarly, we also define three phases for La Niña: June–November of year 0 as La Niña developing phase, January–June of year +1 as La Niña decaying phase, and August–November of year +1 as La Niña reintensification phase. In the next section, we will diagnose the oceanic mixed layer heat budget and relevant atmospheric and oceanic fields during each of the three development phases, in order to understand key physical mechanisms responsible for the El Niño and La Niña evolution asymmetry.

4. Mechanisms responsible for the evolution asymmetry between El Niño and La Niña

Processes that control the ENSO amplitude growth during its developing phase have been extensively examined by many previous studies (e.g., Su et al. 2010). Because the major focus of the present study is the El Niño and La Niña evolution asymmetry, we will just describe the developing-phase MLTA budget analysis results briefly.

During both the El Niño and La Niña developing phases, the primary mechanism for the MLTA growth is 3D ocean temperature advection terms, while the net surface heat flux terms tend to suppress the growth. The 3D ocean temperature advection terms can be further decomposed into nine terms, following Su et al. (2010). The major contributors to the MLTA growth for both El Niño and La Niña are the anomalous zonal advection term (), meridional advection term (), and vertical advection terms ( and ).

To illustrate these positive feedback processes, we plotted the zonal–vertical cross section of ocean mean temperature and anomalous current fields along the equator (Figs. 3a,b). There are pronounced eastward (westward) mixed layer zonal current anomalies in the equatorial EP during El Niño (La Niña). Because the mean zonal temperature gradient is negative (), the eastward (westward) zonal current anomalies induce warm (cold) advection. The eastward (westward) current anomalies also induce anomalous downwelling (upwelling) in the far equatorial EP, leading to a warm (cold) vertical advection because the mean vertical temperature gradient is positive (). Following Su et al. (2010), we separate the surface geostrophic and Ekman currents. It is found that the geostrophic currents dominate the anomalous surface currents, while the wind-induced Ekman currents are small (Figs. 3c,d).

Fig. 3.

(a),(b) Composite ocean mean temperature (contours; °C) and anomalous current (vectors; u′ in m s−1 and w′ in 10−4 m s−1) along the equator (within 2°N–2°S); (c),(d) anomalies of ocean currents (solid line), geostrophic currents (dashed line), and Ekman currents (dotted line) along the equator (within 2°N–2°S; m s−1); and (e),(f) ocean mean meridional currents (vectors; in m s−1 and in 10−4 m s−1) and temperature anomalies (shading; °C) averaged in the EP region (180°–80°W) for (left) El Niño and (right) La Niña developing phase.

Fig. 3.

(a),(b) Composite ocean mean temperature (contours; °C) and anomalous current (vectors; u′ in m s−1 and w′ in 10−4 m s−1) along the equator (within 2°N–2°S); (c),(d) anomalies of ocean currents (solid line), geostrophic currents (dashed line), and Ekman currents (dotted line) along the equator (within 2°N–2°S; m s−1); and (e),(f) ocean mean meridional currents (vectors; in m s−1 and in 10−4 m s−1) and temperature anomalies (shading; °C) averaged in the EP region (180°–80°W) for (left) El Niño and (right) La Niña developing phase.

The meridional–vertical cross sections of anomalous ocean temperature and mean current fields are shown in Figs. 3e and 3f. The climatological meridional overturning circulation in the tropical Pacific is characterized by a subtropical cell (STC; McCreary and Lu 1994), with equatorward flow at the subsurface, upwelling over the equator, and poleward currents at the surface. Thus, the mean surface currents advect anomalous warm (cold) equatorial water poleward, leading to a positive (negative) MLTA tendency during the El Niño (La Niña) developing phase.

In general the anomalous circulation and SST patterns are mirror images between the El Niño and La Niña during their developing phase, as shown in Fig. 3. However, such a symmetric feature changes during their decaying phase. The SSTA amplitude change from December (0) to June (+1) in Fig. 2 suggests that El Niño decays at a rate twice as large as La Niña. During this period, El Niño amplitude decreases by 1.6 K, while La Niña amplitude decreases by 0.8 K. In the following, we will investigate the cause of the distinctive damping rates between El Niño and La Niña and the consequence of the difference in subsequent SSTA evolutions.

a. The cause of distinctive decaying rates between El Niño and La Niña

Table 1 shows the composite MLTA budget analysis during the decaying phase of El Niño and La Niña. As expected, the MLTA decaying rate during El Niño is about 2 times as large as that during La Niña. A further examination of the separate ocean advection and surface heat flux terms shows that both the dynamic and thermodynamic terms contribute equally to the El Niño and La Niña difference (Table 1). The dynamic term difference is primarily attributed to anomalous zonal advection (), while the thermodynamic term difference is mainly due to anomalous surface shortwave radiation and latent heat flux effects.

Table 1.

Composite mixed layer temperature anomaly (MLTA) budget analysis during the decaying phase of El Niño and La Niña averaged in the equatorial eastern Pacific (EP) region (5°N–5°S, 180°–80°W) (°C month−1). Adv denotes advection terms, Hflx represents heat flux terms, Sum is the sum of Adv and Hflx, sw′ denotes the anomalous shortwave radiation term, and lh′ is the anomalous latent heat flux term. The values are based on the ensemble average of two ocean reanalysis datasets (GODAS and SODAv2.1.6) and two surface heat flux products (NCEP-2 and OAFlux).

Composite mixed layer temperature anomaly (MLTA) budget analysis during the decaying phase of El Niño and La Niña averaged in the equatorial eastern Pacific (EP) region (5°N–5°S, 180°–80°W) (°C month−1). Adv denotes advection terms, Hflx represents heat flux terms, Sum is the sum of Adv and Hflx, sw′ denotes the anomalous shortwave radiation term, and lh′ is the anomalous latent heat flux term. The values are based on the ensemble average of two ocean reanalysis datasets (GODAS and SODAv2.1.6) and two surface heat flux products (NCEP-2 and OAFlux).
Composite mixed layer temperature anomaly (MLTA) budget analysis during the decaying phase of El Niño and La Niña averaged in the equatorial eastern Pacific (EP) region (5°N–5°S, 180°–80°W) (°C month−1). Adv denotes advection terms, Hflx represents heat flux terms, Sum is the sum of Adv and Hflx, sw′ denotes the anomalous shortwave radiation term, and lh′ is the anomalous latent heat flux term. The values are based on the ensemble average of two ocean reanalysis datasets (GODAS and SODAv2.1.6) and two surface heat flux products (NCEP-2 and OAFlux).

It is argued that the zonal current asymmetry between El Niño and La Niña arises primarily from the asymmetry of atmospheric wind responses in the WP to ENSO forcing. Figures 4a and 4b show the composite 925-hPa wind and streamfunction anomalies during the El Niño and La Niña mature winter [December–February (DJF)]. Note that during El Niño an anomalous off-equatorial anticyclone occurred over the western North Pacific (WNP; Wang et al. 2000; Chang et al. 2000a,b). Easterly wind anomalies south of the anticyclone tend to stimulate oceanic upwelling Kelvin waves (Weisberg and Wang 1997; Wang et al. 1999; Kim and Lau 2001; Wang and Picaut 2004; Li et al. 2007; Ohba and Ueda 2009), lifting the equatorial thermocline. The negative thermocline anomalies induce westward geostrophic ocean surface currents, leading to a cold advection that damps the warm SSTA. In contrast, an anomalous cyclone occurs in the WNP during La Niña. However, this anomalous cyclone center shifts westward (Fig. 4b). As a result, an anomalous easterly, rather than an anomalous westerly, appears in the far equatorial WP (140°E–180°). This distinctive zonal wind asymmetry appears responsible for the anomalous zonal advection difference between El Niño and La Niña during their decaying phase.

Fig. 4.

Composite 925-hPa wind anomalies (vectors; m s−1) and streamfunction anomalies (shading; 106 m2 s−1) during (a) El Niño and (b) La Niña mature winter (DJF) and asymmetric component estimated by the sum of them (c) during mature winter (DJF) and (d) following spring (MAM). (e),(f) The normalized asymmetric components are also shown (i.e., normalized according to the equatorial EP SSTA amplitude before calculating the asymmetric component). Areas marked with AC (C) denote anticyclonic (cyclonic) circulation.

Fig. 4.

Composite 925-hPa wind anomalies (vectors; m s−1) and streamfunction anomalies (shading; 106 m2 s−1) during (a) El Niño and (b) La Niña mature winter (DJF) and asymmetric component estimated by the sum of them (c) during mature winter (DJF) and (d) following spring (MAM). (e),(f) The normalized asymmetric components are also shown (i.e., normalized according to the equatorial EP SSTA amplitude before calculating the asymmetric component). Areas marked with AC (C) denote anticyclonic (cyclonic) circulation.

Following Wu et al. (2010a), the asymmetric (symmetric) wind field is defined as the sum of (difference between) the composite El Niño and La Niña wind fields. Figure 4c shows clearly that the asymmetric wind field in the WNP is dominated by the El Niño composite pattern. The cause of the asymmetric circulation between El Niño and La Niña in northern winter is caused by the SSTA asymmetry in the preceding summer and fall, as shown by Wu et al. (2010a).

An issue related to the asymmetric wind response shown in Fig. 4c is whether this asymmetry arises simply as a result of the SSTA amplitude difference between El Niño and La Niña. To address this issue, we calculated normalized wind fields (normalized with respect to the amplitude of the SSTA in the EP) for both El Niño and La Niña and then computed the asymmetric component. The so-derived asymmetric wind fields are shown in Fig. 4e. The asymmetry of the zonal wind fields between El Niño and La Niña is still clearly seen. Furthermore, the asymmetry becomes even stronger in March–May (MAM) (+1) (Figs. 4d,f).

The asymmetry in the equatorial wind field exerts a great impact on distinctive anomalous thermocline evolutions between El Niño and La Niña in year +1 (Fig. 5). For the El Niño composite, a stronger negative thermocline signal develops in the equatorial WP in the earlier part of year +1 and penetrates into the EP in May (+1) (Fig. 5, left). This negative thermocline anomaly continues strengthening as its center propagates to the east. This evolution feature is quite different from the La Niña composite (Fig. 5, right), in which a positive thermocline anomaly signal is so weak that it cannot change the sign in the equatorial EP. As a consequence of the weak wind forcing, an anomalous thermocline pattern with a minimum at the equator and two maximum centers off the equator emerges in May (Fig. 5f).

Fig. 5.

Composite thermocline depth anomalies (shading; m) for (left) El Niño and (right) La Niña episodes in year +1.

Fig. 5.

Composite thermocline depth anomalies (shading; m) for (left) El Niño and (right) La Niña episodes in year +1.

A much stronger equatorial thermocline damping during El Niño may have the following two effects on the SSTA tendency. First, shallower thermocline could induce greater westward geostrophic currents, which could lead to a cold advection anomaly () to cool the SST. Second, a stronger negative thermocline depth anomaly could also induce a stronger negative subsurface temperature anomaly, which further cools the SST through anomalous vertical advection by mean upwelling. Both the effects favor a faster decaying of El Niño. In contrast, during March–May of the La Niña decaying phase, a weaker thermocline depth anomaly at the equator leads to a weaker eastward current anomaly, which causes a weaker MLTA tendency and thus a slower decaying rate during the La Niña decaying phase.

Figure 6 is another view illustrating the asymmetric evolutions of the equatorial thermocline anomalies. During January–May of the El Niño decaying phase, the eastern edge of the negative thermocline depth anomalies reaches to 150°W, with a maximum center located at the date line. This negative thermocline anomaly moves all the way to the far EP later in the year (Fig. 6a). In contrast, during January–May of La Niña decaying phase, the eastern edge of the positive thermocline depth anomalies is confined to the west of the date line, with a maximum center located around 150°E. The positive thermocline anomalies continue staying in the WP throughout the year.

Fig. 6.

Evolutions of thermocline depth anomalies (shading; m) along the equator (within 2°N–2°S) for composite (a) El Niño and (b) La Niña in year +1.

Fig. 6.

Evolutions of thermocline depth anomalies (shading; m) along the equator (within 2°N–2°S) for composite (a) El Niño and (b) La Niña in year +1.

The longitude–depth cross sections of oceanic temperature anomalies illustrate a similar feature. A stronger wind forcing in the WP during El Niño excites upwelling Kelvin waves so that cold subsurface temperature anomalies can easily propagate into the equatorial EP (Figs. 7c,e). The eastern flank of the cold subsurface water anomalies extends to 120°W in May (Fig. 5e). In contrast, positive subsurface temperature anomalies are primarily confined in the WP and cannot penetrate into the EP during La Niña (Fig. 7, right).

Fig. 7.

Composite ocean currents (vectors; u′ in m s−1, and w′ in 10−4 m s−1) and temperature anomalies (shading; °C) for (left) El Niño and (right) La Niña episodes in year +1.

Fig. 7.

Composite ocean currents (vectors; u′ in m s−1, and w′ in 10−4 m s−1) and temperature anomalies (shading; °C) for (left) El Niño and (right) La Niña episodes in year +1.

In addition to the oceanic dynamic effect, the asymmetric thermodynamic damping effects also contribute greatly to the distinctive SSTA decaying rates between El Niño and La Niña (Table 1). Figure 8 shows the composite OLR anomaly patterns averaged during the decaying phase of El Niño and La Niña. A negative OLR anomaly, with a maximum center located in 160°W, occurs in the CP and EP during El Niño (Fig. 8a). In contrast, a positive OLR anomaly center during La Niña shifts westward by 40° in longitude (Fig. 8b). Given that SSTA centers are primarily located in the equatorial EP, this anomalous OLR center shift implies a weaker negative cloud–radiation–SST feedback during La Niña. Figure 8c illustrates the asymmetric component of the OLR anomaly pattern between El Niño and La Niña. The result confirms the east–west asymmetry of cloud responses between El Niño and La Niña.

Fig. 8.

Composite OLR anomaly patterns (shading; W m−2) during the decaying phase of (a) El Niño, (b) La Niña, and (c) their difference (i.e., their asymmetric component); (d) the normalized asymmetric component (normalized with respect to the SSTA amplitude). The rectangular box indicates the mixed layer heat budget region (5°N–5°S, 180°–80°W).

Fig. 8.

Composite OLR anomaly patterns (shading; W m−2) during the decaying phase of (a) El Niño, (b) La Niña, and (c) their difference (i.e., their asymmetric component); (d) the normalized asymmetric component (normalized with respect to the SSTA amplitude). The rectangular box indicates the mixed layer heat budget region (5°N–5°S, 180°–80°W).

To examine whether the OLR asymmetry is attributed to SSTA amplitude difference between El Niño and La Niña, we also calculated the normalized OLR fields and their asymmetric component (normalized with respect to the SSTA amplitude). Figure 8d shows the asymmetric OLR anomaly pattern per degree warming or cooling. One can see that even in the presence of same unit SSTA forcing, a stronger negative cloud–radiation–SST feedback is still presented during El Niño. Thus, the asymmetry is primarily caused by the OLR pattern asymmetry between El Niño and La Niña.

The asymmetry in the surface latent heat flux anomalies also contributes to the distinctive damping rates between the El Niño and La Niña decaying phase (Table 1). To examine whether the asymmetry is attributed to anomalous wind speed or anomalous sea–air specific humidity difference, we calculated their relative contributions separately. Table 2 shows that the major contribution arises from the anomalous sea–air specific humidity difference field, whereas the anomalous wind contribution to the asymmetric evaporative damping is relatively small.

Table 2.

Composite latent heat flux budget analysis (W m−2) during the decaying phase of El Niño and La Niña averaged in the equatorial EP region (5°N–5°S, 180°–80°W). Variable lh denotes latent heat flux term, ρ0 represents air density, Lυ is the latent heat of vaporization, Ce denotes the surface exchange coefficient, U represents surface wind speed, Qs is saturated specific humidity at SST, and Qa is air specific humidity at 10 m. The values are calculated based on the ensemble mean of NCEP-2 and OAFlux products.

Composite latent heat flux budget analysis (W m−2) during the decaying phase of El Niño and La Niña averaged in the equatorial EP region (5°N–5°S, 180°–80°W). Variable lh denotes latent heat flux term, ρ0 represents air density, Lυ is the latent heat of vaporization, Ce denotes the surface exchange coefficient, U represents surface wind speed, Qs is saturated specific humidity at SST, and Qa is air specific humidity at 10 m. The values are calculated based on the ensemble mean of NCEP-2 and OAFlux products.
Composite latent heat flux budget analysis (W m−2) during the decaying phase of El Niño and La Niña averaged in the equatorial EP region (5°N–5°S, 180°–80°W). Variable lh denotes latent heat flux term, ρ0 represents air density, Lυ is the latent heat of vaporization, Ce denotes the surface exchange coefficient, U represents surface wind speed, Qs is saturated specific humidity at SST, and Qa is air specific humidity at 10 m. The values are calculated based on the ensemble mean of NCEP-2 and OAFlux products.

As the sea–air humidity difference field is proportional to local SST (Li and Wang 1994), it is not surprising to observe a greater evaporative damping as the SSTA amplitude is greater during El Niño. Figure 9 shows the composite sea–air specific humidity difference field during the decaying phase of El Niño and La Niña. In the El Niño decaying phase, a positive sea–air specific humidity difference anomaly occurs in the equatorial EP, with a maximum center located in 160°W (Fig. 9a). In contrast to El Niño, the negative sea–air specific humidity difference anomaly center shifts westward during La Niña decaying phase (Fig. 9b). As a result, the specific humidity difference field shows a marked pattern asymmetry between El Niño and La Niña, characterized by a positive anomaly in the equatorial EP and a negative anomaly to its west (Fig. 9c). This asymmetric pattern implies a weaker evaporative damping during La Niña than during El Niño. Figure 9d further indicates that this asymmetry does not depend on the amplitude of El Niño and La Niña. Therefore, the asymmetry of the evaporative damping between El Niño and La Niña is primarily caused by the pattern asymmetry of sea–air specific humidity difference field, in a way similar to that of the cloud–shortwave radiation–SST feedback.

Fig. 9.

Composite sea–air specific humidity difference anomaly field (shading; g kg−1) during the decaying phase of (a) El Niño, (b) La Niña, and (c) their asymmetric component; (d) the normalized asymmetric component (normalized with respect to the SSTA amplitude). The rectangular box indicates the mixed layer heat budget region (5°N–5°S, 180°–80°W).

Fig. 9.

Composite sea–air specific humidity difference anomaly field (shading; g kg−1) during the decaying phase of (a) El Niño, (b) La Niña, and (c) their asymmetric component; (d) the normalized asymmetric component (normalized with respect to the SSTA amplitude). The rectangular box indicates the mixed layer heat budget region (5°N–5°S, 180°–80°W).

b. The impact of the distinctive damping rates in the earlier decaying phase on subsequent SSTA evolution

Because of a stronger decaying rate, the composite SSTA in the equatorial EP associated with El Niño is nearly zero by June of year +1 (Fig. 2a). In June, although the SSTA and the surface wind anomaly is small, the equatorial thermocline depth anomaly is strong and negative (Figs. 5 and 7), because of stronger wind forcing in the preceding winter and spring (Fig. 4). The negative thermocline anomalies induce westward geostrophic currents ug < 0), leading to cold temperature advections ( and ) (Figs. 7e,g). By October of year +1, a significantly large cold SSTA has been established (Fig. 2a).

Northern autumn is the season of the strongest coupled ENSO instability (Li 1997). This is because both the mean upwelling and the mean zonal temperature gradient along the equator reach a maximum in boreal fall, when the cold tongue is strongest (Li and Philander 1996; Philander et al. 1996). As a result, the Bjerknes thermocline–SST feedback () and the zonal advective feedback () are strongest in the season. Thus, whether or not the sign of the SSTA would change by summer of year +1 is critical in determining the subsequent ENSO development. For the composite El Niño case, a stronger damping earlier in the year leading to a fast transition to a negative SSTA by July of year +1 makes a “perfect” timing for further development of a cold episode through efficient coupled instability in northern fall of year +1. In contrast, a weaker damping during the La Niña decaying phase leads to no sign change of the SSTA by summer of year +1; as a consequence, the weak cold SSTA starts to develop in northern fall and reaches a peak in the subsequent winter. This explains why an El Niño, after reaching its peak, often transitions into a La Niña, whereas a La Niña often persists for 2–3 years.

It is worth mentioning that the surface heat flux anomaly works primarily in damping the SSTA perturbation, and it cannot change the SSTA sign. It is wind-induced dynamic forcing that is essential for causing the phase transition of the SSTA from a warm to a cold anomaly. A heat budget analysis during the El Niño transition phase confirms the argument above. Table 3 shows that the MLTA tendency during the El Niño transition phase is primarily caused by 3D ocean temperature advection terms, while the surface heat flux terms play an opposite role. In particular during June–July (+1), the major driving force for the negative MLTA tendency is the anomalous advection terms (figure not shown). The heat flux terms are negligible during the period.

Table 3.

As in Table 1, but for the transition phase of El Niño and reintensification phase of La Niña (°C month−1).

As in Table 1, but for the transition phase of El Niño and reintensification phase of La Niña (°C month−1).
As in Table 1, but for the transition phase of El Niño and reintensification phase of La Niña (°C month−1).

During the El Niño transition phase, the major contributors are anomalous zonal advection term (), meridional advection term (), and thermocline feedback term () (Table 3). Figure 10a shows that there are pronounced westward zonal current anomalies (u′ < 0) during this period. The westward currents are mainly induced by the thermocline depth anomaly related geostrophic currents in the equatorial EP (Fig. 10c). From the meridional–vertical cross section of anomalous ocean temperature and mean current fields (averaged within 180°–80°W; Fig. 10e), one may see that negative temperature anomalies occur at the equator. Because the anomalous temperature is larger in the subsurface than at the surface, the vertical gradient of the anomalous temperature is positive (∂T′/∂z > 0). The mean STC in the tropical Pacific, on one hand, advects the colder subsurface water upward and cools the SST and, on the other hand, advects the cold surface water poleward. Both the processes contribute to the strong negative SSTA tendency during the El Niño transition phase.

Fig. 10.

(a),(b) Composite ocean mean temperature (contours; °C) and anomalous current (vectors; u′ in m s−1 and w′ in 10−4 m s−1) along the equator (within 2°N–2°S); (c),(d) anomalies of ocean currents (solid line), geostrophic currents (dashed line), and Ekman currents (dotted line) along the equator (within 2°N–2°S; m s−1); (e),(f) ocean mean meridional currents (vectors; in m s−1 and in 10−4 m s−1) and temperature anomalies (shading; °C) averaged in the EP region (180°–80°W) for (left) El Niño transition phase and (right) La Niña reintensification phase.

Fig. 10.

(a),(b) Composite ocean mean temperature (contours; °C) and anomalous current (vectors; u′ in m s−1 and w′ in 10−4 m s−1) along the equator (within 2°N–2°S); (c),(d) anomalies of ocean currents (solid line), geostrophic currents (dashed line), and Ekman currents (dotted line) along the equator (within 2°N–2°S; m s−1); (e),(f) ocean mean meridional currents (vectors; in m s−1 and in 10−4 m s−1) and temperature anomalies (shading; °C) averaged in the EP region (180°–80°W) for (left) El Niño transition phase and (right) La Niña reintensification phase.

The strengthening of the cold SSTA during the La Niña reintensification phase is primarily caused by the anomalous zonal advection term (), meridional advection term (), and thermocline feedback term (), as shown in Table 3. The negative thermocline depth anomalies in the equatorial Pacific induce westward geostrophic currents, which cool the SST (Figs. 7j and 10b,d). They can also strengthen the surface cooling through advection of colder subsurface water by mean upwelling (Fig. 10f). All these positive feedback processes contribute to the La Niña reintensification.

To sum up, the asymmetry of El Niño and La Niña evolutions is primarily attributed to distinctive SSTA tendencies (caused by both the dynamic and thermodynamic processes) during their decaying phase. A stronger wind forcing during the El Niño decaying phase causes the negative thermocline anomaly to penetrate all the way to the EP (Fig. 5). This helps the transition to a negative SSTA by the boreal summer of year +1. The season-dependent coupled instability further strengthens the cold SSTA in northern fall, leading to a peak phase of La Niña in northern winter (+1). Because of weaker wind forcing associated with La Niña, the positive thermocline anomaly at the equator in earlier year +1 is weaker, and so are anomalous geostrophic currents and anomalous subsurface temperatures. As a result, the cold SSTA does not change sign by the northern summer of year +1. A stronger coupled instability in boreal fall reinforces the cold anomaly, leading to a peak phase of La Niña in the subsequent winter.

An interesting feature of the anomalous thermocline evolution during year +1 of the La Niña composite is the separation of two off-equatorial centers earlier in the year (Figs. 5d,f) and the reshaping of its structure with a maximum center right on the equator in September (Fig. 5j). This meridional structure change signifies a new phase of La Niña development, as so-induced geostrophic zonal and vertical current anomalies favor further development of La Niña (Su et al. 2010). What causes such a meridional structure change?

The first possible mechanism is the wind-forced equatorial wave dynamics. It is seen that the downwelling Kelvin wave that one expects from the delayed oscillator theory is making its way eastward during March–July (+1). This weakens the shallow thermocline anomaly near the equator. But the Kelvin wave is too weak to turn the event around. With reintensification of easterly wind anomalies in northern fall (because of the enhanced coupled instability), the thermocline shoals again toward the end of the year.

Another possible mechanism is the STC advection process. To illustrate clearly the split-reemergence process of the equatorial thermocline anomaly, we plotted the latitude–time cross section of thermocline depth anomalies averaged over 180°–120°W during the year +1 of the La Niña composite (Fig. 11). Note that in January, the anomalous thermocline center appears at the equator. After that, it begins to split poleward, owing to the eastward passage of weak positive thermocline signals forced by off-equatorial anomalous anticyclone in the preceding months. As a result, the negative thermocline anomalies are stronger off the equator. The corresponding anomalous geostrophic current is eastward during the period (Figs. 7d,f). This anomalous ocean current and associated upwelling attempted to weaken the cold SSTA. However, they are not strong enough to change the SSTA sign. As time progresses, the positive thermocline anomalies resulting from off-equatorial anticyclonic circulation forcing attenuate. Meanwhile, the off-equatorial negative thermocline anomalies (in particular the Southern Hemisphere branch) gradually move toward the equator as a result of the advection of the climatological mean STC. It is noted that the subsurface equatorward mean meridional current velocity averaged over 1°–4°S, 180°–120°W is 0.04 m s−1, or about 1° of latitude per month. The mean meridional velocity is consistent with the observed phase evolution of the negative thermocline anomalies. Figure 11 shows that it took 5 months (from May to October) for the off-equatorial cold subsurface waters to move from 5°S or 5°N to the equator. The consistent STC mean meridional velocity and the thermocline meridional phase speed suggest that the anomalous subsurface water is likely to be advected by the mean current. This subsurface advective process was previously emphasized in Su et al. (2014).

Fig. 11.

Evolutions of the thermocline depth anomaly (shading; m) averaged within 180°–120°W during La Niña year +1.

Fig. 11.

Evolutions of the thermocline depth anomaly (shading; m) averaged within 180°–120°W during La Niña year +1.

To further support the meridional advection mechanism, we conducted a subsurface heat budget analysis in the region of interest (i.e., 4°–1°S, 180°–120°W). Figure 12a shows that the cold off-equatorial subsurface water moves to the equator gradually during the period of May–October in La Niña year +1. In response to this northward movement, one would expect a negative temperature anomaly tendency to north of the negative temperature center, as shown in Fig. 12b. Our subsurface temperature budget analysis shows that the negative temperature anomaly tendency is primarily attributed to anomalous meridional temperature advection term, while the anomalous zonal and vertical advection terms are positive (figure not shown). Averaged over 4°–1°S for the period of May–October in year +1, the total subsurface temperature tendency is −0.1°C month−1, while meridional advection anomaly is −0.12°C month−1. A further diagnosis that separates mean and anomalous current contribution shows that the anomalous meridional advection is primarily determined by the advection of temperature perturbation by mean meridional current associated with the climatologic STC (see Fig. 12c). Therefore, both the mean current analysis and subsurface heat budget diagnosis above support the notion that the mean STC advects the off-equatorial anomalous cold subsurface water equatorward.

Fig. 12.

Time–latitude cross section of (a) subsurface (averaged at 100–150 m) temperature anomaly (shading; °C), (b) subsurface temperature anomaly tendency (shading; °C month−1), and (c) meridional advection of temperature perturbation by mean current (; shading; °C month−1) averaged within 180°–120°W during the period of April–November in La Niña year +1.

Fig. 12.

Time–latitude cross section of (a) subsurface (averaged at 100–150 m) temperature anomaly (shading; °C), (b) subsurface temperature anomaly tendency (shading; °C month−1), and (c) meridional advection of temperature perturbation by mean current (; shading; °C month−1) averaged within 180°–120°W during the period of April–November in La Niña year +1.

Thus, both the wind-forced equatorial wave dynamics and the mean STC advection are possible mechanisms to explain the reemergence of the negative thermocline anomaly center at the equator late in La Niña year +1. The former is related to the enhanced Bjerknes feedback in northern fall, while the latter depends on mean subsurface current velocity. Therefore, both the processes are closely related to the annual cycle mean state.

5. Discussion

a. Effect of the IO SSTA on wind response in WP during El Niño peak winter

We first examine the observed characteristics of composite wind, precipitation, and SST fields during El Niño mature winter (DJF). Figure 13 shows that although the basinwide SSTA pattern occurs in IO, rainfall and OLR anomalies show a clear east–west dipole structure, with enhanced (suppressed) convection over the western (eastern) IO. Prior to this season, the SSTA in IO is dominated by a zonal dipole, with strong cooling in the eastern pole [also seen in Wu et al. (2012)]. The rapid warming in eastern IO from northern fall to winter is attributed to both strengthened shortwave radiation forcing and ocean wave effect (Li et al. 2003; Hong et al. 2010).

Fig. 13.

Composite patterns of (a),(c) 925-hPa wind (vectors; m s−1) and SSTA (shading; °C) fields and (b),(d) OLR (shading; W m−2) and precipitation (contours; mm) fields for (top) El Niño and (bottom) La Niña during ENSO mature phase (DJF). Areas marked with AC (C) denote anticyclonic (cyclonic) circulation. Composite is based on 1980–2013.

Fig. 13.

Composite patterns of (a),(c) 925-hPa wind (vectors; m s−1) and SSTA (shading; °C) fields and (b),(d) OLR (shading; W m−2) and precipitation (contours; mm) fields for (top) El Niño and (bottom) La Niña during ENSO mature phase (DJF). Areas marked with AC (C) denote anticyclonic (cyclonic) circulation. Composite is based on 1980–2013.

The observed IO rainfall–SST relationship implies a distinctive role the SSTA plays in affecting local convection and circulation. While the local warming in the eastern IO plays a passive role (i.e., it is local atmosphere that influences the ocean), the SSTA in the western IO plays an active role in strengthening atmospheric convection (i.e., it is the ocean that primarily influences the atmosphere) (Wu et al. 2009, 2012). Given such a complex relationship, one needs to be cautious when designing numerical model experiments. Forced SSTA experiments such as those by Ohba and Ueda (2007) and Okumura et al. (2011) led to a basinwide precipitation anomaly over IO [see Ohba and Ueda (2007), their Fig. 2c and also Okumura et al. (2011), their Fig. 3, left panel of the second row], which, according to Gill’s (1980) solution, would lead to a response of low-level easterlies in the western equatorial Pacific.

To demonstrate how a specified SSTA experiment in IO might negatively impact the atmospheric response in WP, we conducted a control experiment (in which the climatologic monthly SST is specified everywhere in the ocean) and two sensitivity experiments using ECHAM4. In the first sensitivity experiment, the observed SSTA in IO (same as that shown in Fig. 13a) is superposed onto the climatologic SST field. In the second sensitivity experiment, a dipole heating pattern that has the same horizontal pattern as the precipitation anomaly shown in Fig. 13b and an idealized vertical profile that has a maximum at the midtroposphere is specified. Figure 14 shows the model simulation result from the two experiments. As expected, a basinwide precipitation anomaly occurs in the specified SSTA experiment; as a result the low-level wind anomaly in equatorial WP is dominated by easterlies. In contrast, a westerly anomaly appears in the equatorial WP in the specified heating experiment, because anomalous descending (ascending) motion in eastern (western) IO favors a reversed Walker circulation in both the tropical IO and Pacific.

Fig. 14.

The 850-hPa wind anomaly fields (vectors; m s−1) simulated from ECHAM4 (left) in response to a specified SSTA forcing in tropical IO and (right) in response to a specified diabatic heating anomaly in the tropical IO. The blue arrow denotes anomalous easterly wind, and red arrow denotes anomalous westerly wind.

Fig. 14.

The 850-hPa wind anomaly fields (vectors; m s−1) simulated from ECHAM4 (left) in response to a specified SSTA forcing in tropical IO and (right) in response to a specified diabatic heating anomaly in the tropical IO. The blue arrow denotes anomalous easterly wind, and red arrow denotes anomalous westerly wind.

The numerical experiments above indicate that a caution is needed in designing idealized atmospheric model experiments with specified SSTA forcing. It is important to compare observed and simulated rainfall anomalies in the specified SSTA region to make sure that the local heating anomaly pattern is realistic.

It is noted that the most pronounced circulation asymmetry feature is an anomalous anticyclone in the Philippine Sea during El Niño but an anomalous cyclone that is weaker and shifts to the west during La Niña (Figs. 4a,b and 13a,c). Different from many previous studies, we demonstrate quantitatively, through a mixed layer heat budget analysis, how important this wind asymmetry is in the subsequent El Niño and La Niña evolution (i.e., how much of the SSTA change is attributed to the wind asymmetry induced horizontal and vertical advection anomalies and surface heat flux changes). The cause of the circulation asymmetry between El Niño and La Niña is possibly attributed to the asymmetries in the equatorial heating anomaly and local SSTA in WP (Wu et al. 2010a).

b. Effect of southward wind shifting on El Niño–La Niña evolution asymmetry

How important is the southward shifting of zonal wind anomalies in CP in causing the El Niño–La Niña evolution asymmetry? Ocean GCM experiments by Harrison and Vecchi (1999) with idealized meridional wind profiles revealed that the southward shift did affect EP ocean thermocline evolution to a certain extent during the wind-shifting months, but the overall transition from a positive thermocline anomaly during El Niño mature winter to a negative thermocline anomaly in the subsequent winter remains unchanged [see Fig. 2 of Harrison and Vecchi (1999)]. This numerical result suggested that this CP wind shifting effect on overall ENSO phase transition might be limited. The shallow water model experiments by McGregor et al. (2013), on the other hand, included both the asymmetric wind anomalies in WP and CP, and thus their relative roles are unclear. It is necessary to examine their separate effects.

McGregor et al. (2013) found that the zonal wind shifting depends on El Niño amplitude, and a pronounced wind shifting becomes visible only during super El Niños. The result implies that this zonal wind shifting effect becomes effective only when super El Niños occur. However, the time series of Niño-3.4 SSTA (Fig. 15) shows that almost all El Niño events, regardless of their amplitude, exhibit a similar positive-to-negative SSTA transition (in a way similar to that shown in Fig. 2a). This implies that the zonal wind shifting in CP might not be a fundamental cause of the El Niño–La Niña evolution asymmetry. Again, the discussion here suggests that there is a need to conduct a further in-depth study to examine carefully the relative roles of the asymmetric wind patterns in WP and CP.

Fig. 15.

Time series of Niño-3.4 SSTA (°C) from 1980 to 2011. Red arrow indicates a quick El Niño–La Niña transition for super El Niño (denoted by S, defined as amplitude greater than 2.5 standard deviations), regular El Niño (denoted by R, defined as warm episodes other than the super El Niños and the CP El Niños), and CP El Niño [denoted by C, defined by Xiang et al. (2013) and Chung and Li (2013)] cases.

Fig. 15.

Time series of Niño-3.4 SSTA (°C) from 1980 to 2011. Red arrow indicates a quick El Niño–La Niña transition for super El Niño (denoted by S, defined as amplitude greater than 2.5 standard deviations), regular El Niño (denoted by R, defined as warm episodes other than the super El Niños and the CP El Niños), and CP El Niño [denoted by C, defined by Xiang et al. (2013) and Chung and Li (2013)] cases.

6. Conclusions

By analyzing the oceanic and atmospheric reanalysis datasets, we reveal the observed asymmetric evolution characteristics between El Niño and La Niña composites. While El Niño is characterized by a rapid decay after its peak and a fast phase transition from a positive to a negative SSTA within one year, La Niña is characterized by a weaker decay after its peak and a reintensification of the cold SSTA later in the second year (year +1).

The relative roles of dynamic (wind-induced thermocline and advection) and thermodynamic (heat flux) processes in causing the distinctive asymmetric evolutions are investigated, through a mixed layer heat budget analysis. The result shows that the faster (slower) decay of El Niño (La Niña) is attributed to both dynamic (ocean advection) and thermodynamic (heat flux) forcing. The former is primarily caused by the asymmetric atmospheric wind responses in WP between El Niño and La Niña. The latter is attributed to the asymmetric cloud–radiation–SST feedback and evaporation–SST feedback between El Niño and La Niña. The mixed layer heat budget analysis shows that the temperature tendency asymmetry between El Niño and La Niña is attributed equally to both the dynamic and thermodynamic processes (Table 1).

Because of the distinctive temperature tendencies during the ENSO decaying phase, the sign of the SSTA changes by July of the second year for the El Niño composite but remains the same for the La Niña composite. Because the coupled instability is strongest in northern fall (Li 1997), the negative SSTA in the El Niño composite grows and transitions into a La Niña in boreal winter of the second year. In contrast, a much weaker damping rate during the La Niña decaying phase leads to no sign change of the SSTA by boreal summer of the second year. Once the season-dependent coupled instability kicks in during northern fall, the negative SSTA reinvigorates itself and grows into another La Niña episode by the end of the second year.

The distinctive asymmetric wind responses in WP between El Niño and La Niña hold a key for subsequent asymmetric ocean thermocline responses. During the mature phase of El Niño, a positive SSTA in EP induces a cold SSTA and PSAC in WP through local thermodynamic air–sea feedback (Wang et al. 2000, 2003; Wu et al. 2010a,b). Easterlies to the south of the anomalous anticyclone may trigger upwelling Kelvin waves, which lead to subsurface cooling in the equatorial Pacific. The colder subsurface water further cools surface water through anomalous vertical advection by mean upwelling. The shoaling thermocline can also bring out anomalous eastward geostrophic currents and cools the surface water through anomalous zonal advection. This dynamic forcing effect eventually leads to a fast transition to a cold SSTA by boreal summer of the second year.

In contrast, during the mature phase of La Niña, a cold SSTA in the eastern equatorial Pacific induces an anomalous cyclone with its center located west of the Philippines. Because of this asymmetric response, the equatorial thermocline anomaly is much weaker during the La Niña decaying phase. This weak dynamic forcing, along with a weaker negative cloud–radiation–SST and evaporation–SST feedback, leads to a slow damping rate for La Niña. Because of the slowly decaying process, the original negative anomaly is weakly damped in the earlier half and redevelops in the latter half of year +1.

It is worth noting that the asymmetric thermodynamic (heat flux) forcing during the ENSO decaying phase is as important as the asymmetric wind effect. The former is often neglected in previous ENSO evolution asymmetry studies. The cause of the heat flux asymmetry arises from 1) the asymmetry of SSTA pattern between El Niño and La Niña and 2) nonlinearity of atmospheric wind, convection, and moisture responses to positive and negative SSTA forcing, as seen in Figs. 8 and 9.

It is worth mentioning that while the MLTA heat budget analysis helps us understand the relative roles of ocean dynamics and surface heat fluxes in causing the asymmetric temperature evolutions, however, caution is needed to interpret the budget result, as MLTA might differ from SSTA in some occasions. It is the SSTA that actually affects the atmosphere.

In this study the composites were made only for those El Niño and La Niña cases that have common evolution features. We intentionally excluded special El Niño and La Niña cases such as 1986/87 El Niño and 1988 and 2005 La Niña. The 1986/87 El Niño persisted its warming through the second winter and there was no phase transition. The 1988 and 2005 La Niña episodes decayed quickly and had no reintensification phase. Such special ENSO episodes have significantly different evolution characteristics from the common ENSO evolution features shown in Fig. 2. The physical processes involved in these special ENSO cases and the fundamental mechanism responsible for the difference between the typical and special ENSO cases will be investigated in future.

Acknowledgments

This work was supported by China National 973 Project 2015CB453200, NSFC Grant 41475084, ONR Grant N00014-1210450, the Jiangsu Shuang-Chuang Team, and the International Pacific Research Center. MC is also supported by the Research Innovation Program for College Graduates of Jiangsu Province (CXZZ13_0504).

REFERENCES

REFERENCES
An
,
S.-I.
, and
F.-F.
Jin
,
2004
:
Nonlinearity and asymmetry of ENSO
.
J. Climate
,
17
,
2399
2412
, doi:.
Battisti
,
D. S.
, and
A. C.
Hirst
,
1989
:
Interannual variability in a tropical atmosphere–ocean model: Influence of the basic state, ocean geometry and nonlinearity
.
J. Atmos. Sci.
,
46
,
1687
1712
, doi:.
Bjerknes
,
J.
,
1969
:
Atmospheric teleconnections from the equatorial Pacific
.
Mon. Wea. Rev.
,
97
,
163
172
, doi:.
Burgers
,
G.
, and
D. B.
Stephenson
,
1999
:
The “normality” of El Niño
.
Geophys. Res. Lett.
,
26
,
1027
1030
, doi:.
Cane
,
M. A.
, and
S. E.
Zebiak
,
1985
:
A theory for El Niño and the Southern Oscillation
.
Science
,
228
,
1085
1087
, doi:.
Carton
,
J. A.
, and
B. S.
Giese
,
2008
:
A reanalysis of ocean climate using Simple Ocean Data Assimilation (SODA)
.
Mon. Wea. Rev.
,
136
,
2999
3017
, doi:.
Chang
,
C.-P.
,
Y.
Zhang
, and
T.
Li
,
2000a
:
Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part I: Roles of the subtropical ridge
.
J. Climate
,
13
,
4310
4325
, doi:.
Chang
,
C.-P.
,
Y.
Zhang
, and
T.
Li
,
2000b
:
Interannual and interdecadal variations of the East Asian summer monsoon and tropical Pacific SSTs. Part II: Meridional structure of the monsoon
.
J. Climate
,
13
,
4326
4340
, doi:.
Chang
,
P.
, and Coauthors
,
2006
:
Climate fluctuations of tropical coupled systems? The role of ocean dynamics
.
J. Climate
,
19
,
5122
5174
, doi:.
Chung
,
P.-H.
, and
T.
Li
,
2013
:
Interdecadal relationship between the mean state and El Niño types
.
J. Climate
,
26
,
361
379
, doi:.
Dommenget
,
D.
,
T.
Bayr
, and
C.
Frauen
,
2013
:
Analysis of the non-linearity in the pattern and time evolution of El Niño Southern Oscillation
.
Climate Dyn.
,
40
,
2825
2847
, doi:.
Gill
,
A. E.
,
1980
:
Some simple solutions for heat-induced tropical circulation
.
Quart. J. Roy. Meteor. Soc.
,
106
,
447
462
, doi:.
Harrison
,
D. E.
, and
G. A.
Vecchi
,
1999
:
On the termination of El Niño
.
Geophys. Res. Lett.
,
26
,
1593
1596
, doi:.
Hirst
,
A. C.
,
1986
:
Unstable and damped equatorial modes in simple coupled ocean–atmosphere models
.
J. Atmos. Sci.
,
43
,
606
630
, doi:.
Hirst
,
A. C.
,
1988
:
Slow instabilities in tropical ocean basin–global atmosphere models
.
J. Atmos. Sci.
,
45
,
830
852
, doi:.
Hong
,
C.-C.
,
T.
Li
,
LinHo
, and
J.-S.
Kug
,
2008
:
Asymmetry of the Indian Ocean dipole. Part I: Observational analysis
.
J. Climate
,
21
,
4834
4848
, doi:.
Hong
,
C.-C.
,
T.
Li
,
LinHo
, and
Y.-C.
Chen
,
2010
:
Asymmetry of the Indian Ocean basinwide SST anomalies: Roles of ENSO and IOD
.
J. Climate
,
23
,
3563
3576
, doi:.
Jin
,
F.-F.
,
1997
:
An equatorial ocean recharge paradigm for ENSO. Part I: Conceptual mode
.
J. Atmos. Sci.
,
54
,
811
829
, doi:.
Kanamitsu
,
M.
,
W.
Ebisuzaki
,
J.
Woollen
,
S.-K.
Yang
,
J. J.
Hnilo
,
M.
Fiorino
, and
G. L.
Potter
,
2002
:
NCEP–DOE AMIP-II Reanalysis (R-2)
.
Bull. Amer. Meteor. Soc.
,
83
,
1631
1643
, doi:.
Kang
,
I.-S.
, and
J.-S.
Kug
,
2002
:
El Niño and La Niña sea surface temperature anomalies: Asymmetry characteristics associated with their wind stress anomalies
.
J. Geophys. Res.
,
107
,
4372
, doi:.
Kessler
,
W. S.
,
2002
:
Is ENSO a cycle or a series of events?
Geophys. Res. Lett.
,
29
,
2125
, doi:.
Kim
,
K.-M.
, and
K.-M.
Lau
,
2001
:
Dynamics of monsoon-induced biennial variability in ENSO
.
Geophys. Res. Lett.
,
28
,
315
318
, doi:.
Kim
,
W.
,
S.-W.
Yeh
,
J.-H.
Kim
,
J.-S.
Kug
, and
M.
Kwon
,
2011
:
The unique 2009–2010 El Niño event: A fast phase transition of warm pool El Niño to La Niña
.
Geophys. Res. Lett.
,
38
,
L15809
, doi:.
Larkin
,
N. K.
, and
D. E.
Harrison
,
2002
:
ENSO warm (El Niño) and cold (La Niña) event life cycles: Ocean surface anomaly patterns, their symmetries, asymmetries, and implications
.
J. Climate
,
15
,
1118
1140
, doi:.
Li
,
T.
,
1997
:
Phase transition of the El Niño–Southern Oscillation: A stationary SST mode
.
J. Atmos. Sci.
,
54
,
2872
2887
, doi:.
Li
,
T.
, and
B.
Wang
,
1994
:
A thermodynamic equilibrium climate model for monthly mean surface winds and precipitation over the tropical Pacific
.
J. Atmos. Sci.
,
51
,
1372
1385
, doi:.
Li
,
T.
, and
S. G. H.
Philander
,
1996
:
On the annual cycle of the eastern equatorial Pacific
.
J. Climate
,
9
,
2986
2998
, doi:.
Li
,
T.
,
B.
Wang
,
C.-P.
Chang
, and
Y.
Zhang
,
2003
:
A theory for the Indian Ocean dipole–zonal mode
.
J. Atmos. Sci.
,
60
,
2119
2135
, doi:.
Li
,
Y.
,
R.
Lu
, and
B.
Dong
,
2007
:
The ENSO–Asian monsoon interaction in a coupled ocean–atmosphere GCM
.
J. Climate
,
20
,
5164
5177
, doi:.
Liebmann
,
B.
, and
C. A.
Smith
,
1996
:
Description of a complete (interpolated) outgoing longwave radiation dataset
.
Bull. Amer. Meteor. Soc.
,
77
,
1275
1277
.
McCreary
,
J. P.
, and
P.
Lu
,
1994
:
Interaction between the subtropical and equatorial ocean circulations: The subtropical cell
.
J. Phys. Oceanogr.
,
24
,
466
497
, doi:.
McGregor
,
S.
,
A.
Timmermann
,
N.
Schneider
,
M. F.
Stuecker
, and
M. H.
England
,
2012
:
The effect of the south convergence zone on the termination of El Niño events and the meridional asymmetry of ENSO
.
J. Climate
,
25
,
5566
5586
, doi:.
McGregor
,
S.
,
N.
Ramesh
,
P.
Spence
,
M. H.
England
,
M. J.
McPhaden
, and
A.
Santoso
,
2013
:
Meridional movement of wind anomalies during ENSO events and their role in event termination
.
Geophys. Res. Lett.
,
40
,
749
754
, doi:.
McPhaden
,
M. J.
, and
X.
Zhang
,
2009
:
Asymmetry in zonal phase propagation of ENSO sea surface temperature anomalies
.
Geophys. Res. Lett.
,
36
,
L13703
, doi:.
Neelin
,
J. D.
,
D. S.
Battisti
,
A. C.
Hirst
,
F.-F.
Jin
,
Y.
Wakata
,
T.
Yamagata
, and
S. E.
Zebiak
,
1998
:
ENSO theory
.
J. Geophys. Res.
,
103
,
14 261
14 290
, doi:.
Ohba
,
M.
, and
H.
Ueda
,
2007
:
An impact of SST anomalies in the Indian Ocean in acceleration of the El Niño to La Niña transition
.
J. Meteor. Soc. Japan
,
85
,
335
348
, doi:.
Ohba
,
M.
, and
H.
Ueda
,
2009
:
Role of nonlinear atmospheric response to SST on the asymmetric transition process of ENSO
.
J. Climate
,
22
,
177
192
, doi:.
Okumura
,
Y. M.
, and
C.
Deser
,
2010
:
Asymmetry in the duration of El Niño and La Niña
.
J. Climate
,
23
,
5826
5843
, doi:.
Okumura
,
Y. M.
,
M.
Ohba
,
C.
Deser
, and
H.
Ueda
,
2011
:
A proposed mechanism for the asymmetric duration of El Niño and La Niña
.
J. Climate
,
24
,
3822
3829
, doi:.
Philander
,
S. G. H.
,
1990
:
El Niño, La Niña, and the Southern Oscillation.
Academic Press, 293 pp
.
Philander
,
S. G. H.
,
T.
Yamagata
, and
R. C.
Pacanowski
,
1984
:
Unstable air–sea interactions in the tropics
.
J. Atmos. Sci.
,
41
,
604
613
, doi:.
Philander
,
S. G. H.
,
D.
Gu
,
D.
Halpern
,
G.
Lambert
,
N.-C.
Lau
,
T.
Li
, and
R. C.
Pacanowski
,
1996
:
Why the ITCZ is mostly north of the equator
.
J. Climate
,
9
,
2958
2972
, doi:.
Rasmusson
,
E. M.
, and
T. H.
Carpenter
,
1982
:
Variations in tropical sea surface temperature and surface wind fields associated with the Southern Oscillation/El Niño
.
Mon. Wea. Rev.
,
110
,
354
384
, doi:.
Saha
,
S.
, and Coauthors
,
2006
:
The NCEP Climate Forecast System
.
J. Climate
,
19
,
3483
3517
, doi:.
Smith
,
R. D.
,
J. K.
Dukowicz
, and
R. C.
Malone
,
1992
:
Parallel ocean general circulation modeling
.
Physica D
,
60
,
38
61
, doi:.
Smith
,
T. M.
,
R. W.
Reynolds
,
T. C.
Peterson
, and
J.
Lawrimore
,
2008
:
Improvements to NOAA’s historical merged land–ocean surface temperature analysis (1880–2006)
.
J. Climate
,
21
,
2283
2296
, doi:.
Su
,
J.
,
R.
Zhang
,
T.
Li
,
X.
Rong
,
J.-S.
Kug
, and
C.-C.
Hong
,
2010
:
Causes of the El Niño and La Niña amplitude asymmetry in the equatorial eastern Pacific
.
J. Climate
,
23
,
605
617
, doi:.
Su
,
J.
,
T.
Li
, and
R.
Zhang
,
2014
:
The initiation and developing mechanisms of central Pacific El Niños
.
J. Climate
,
27
,
4473
4485
, doi:.
Suarez
,
M. J.
, and
P. S.
Schopf
,
1988
:
A delayed action oscillator for ENSO
.
J. Atmos. Sci.
,
45
,
3283
3287
, doi:.
Takahashi
,
K.
,
A.
Montecinos
,
K.
Goubanova
, and
B.
Dewitte
,
2011
:
ENSO regimes: Reinterpreting the canonical and Modoki El Niño
.
Geophys. Res. Lett.
,
38
,
L10704
, doi:.
Wang
,
B.
,
R.
Wu
, and
X.
Fu
,
2000
:
Pacific–East Asian teleconnection: How does ENSO affect East Asian climate?
J. Climate
,
13
,
1517
1536
, doi:.
Wang
,
B.
,
R.
Wu
, and
T.
Li
,
2003
:
Atmosphere–warm ocean interaction and its impacts on Asian-–Australian Monsoon variation
.
J. Climate
,
16
,
1195
1211
, doi:.
Wang
,
C.
, and
J.
Picaut
,
2004
:
Understanding ENSO physics: A review. Earth’s Climate: The Ocean–Atmosphere Interaction, Geophys. Monogr., Vol. 147, Amer. Geophys. Union, 21–48
.
Wang
,
C.
,
R. H.
Weisberg
, and
J. I.
Virmani
,
1999
:
Western Pacific interannual variability associated with the El Niño–Southern Oscillation
.
J. Geophys. Res.
,
104
,
5131
5149
, doi:.
Weisberg
,
R. H.
, and
C.
Wang
,
1997
:
A western Pacific oscillator paradigm for the El Niño–Southern Oscillation
.
Geophys. Res. Lett.
,
24
,
779
782
, doi:.
Wu
,
B.
,
T.
Zhou
, and
T.
Li
,
2009
:
Seasonally evolving dominant interannual variability modes of East Asian climate
.
J. Climate
,
22
,
2992
3005
, doi:.
Wu
,
B.
,
T.
Li
, and
T.
Zhou
,
2010a
:
Asymmetry of atmospheric circulation anomalies over the western North Pacific between El Niño and La Niña
.
J. Climate
,
23
,
4807
4822
, doi:.
Wu
,
B.
,
T.
Li
, and
T.
Zhou
,
2010b
:
Relative contributions of the Indian Ocean and local SST anomalies to the maintenance of the western North Pacific anomalous anticyclone during the El Niño decaying summer
.
J. Climate
,
23
,
2974
2986
, doi:.
Wu
,
B.
,
T.
Zhou
, and
T.
Li
,
2012
:
Two distinct modes of tropical Indian Ocean precipitation in boreal winter and their impacts on equatorial western Pacific
.
J. Climate
,
25
,
921
938
, doi:.
Xiang
,
B.
,
B.
Wang
, and
T.
Li
,
2013
:
A new paradigm for the predominance of standing central Pacific warming after the late 1990s
.
Climate Dyn.
,
41
,
327
340
, doi:.
Yu
,
L.
,
X.
Jin
, and
R. A.
Weller
,
2008
:
Multidecade global flux datasets from the objectively analyzed air–sea fluxes (OAFlux) project: Latent and sensible heat fluxes, ocean evaporation, and related surface meteorological variables. Woods Hole Oceanographic Institution Tech. Rep. OA-2008-1, 64 pp. [Available online at http://oaflux.whoi.edu/pdfs/OAFlux_TechReport_3rd_release.pdf.]

Footnotes

*

School of Ocean and Earth Science and Technology Contribution Number 9579, International Pacific Research Center Contribution Number 1169, and Earth System Modeling Center Contribution Number 091.