• An, S-I., and F-F. Jin, 2004: Nonlinearity and asymmetry of ENSO. J. Climate, 17 , 23992412.

  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97 , 163172.

  • Burgers, G., and D. B. Stephenson, 1999: The “normality” of El Nino. Geophys. Res. Lett., 26 , 10271030.

  • Cox, M. D., 1984: A primitive equation, 3-dimensional model of the ocean. GFDL Ocean Group Tech. Rep. 1, Princeton, NJ, 143 pp.

  • Gadgil, S., P. V. Josph, and N. V. Joshi, 1984: Ocean atmosphere coupling over monsoon regions. Nature, 312 , 141143.

  • Glantz, M. H., 1996: Currents of Change: El Niño’s Impact on Climate and Society. Cambridge University Press, 194 pp.

  • Gordon, C., C. Cooper, C. A. Senior, H. Banks, J. M. Gregory, T. C. Johns, J. F. B. Mitchell, and R. A. Wood, 2000: The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Climate Dyn., 16 , 147168.

    • Search Google Scholar
    • Export Citation
  • Hannachi, A., D. Stephenson, and K. Sperber, 2003: Probablity-based methods for quantifying nonlineararity in the ENSO. Climate Dyn., 20 , 241256.

    • Search Google Scholar
    • Export Citation
  • Hoerling, M. P., A. Kumar, and M. Zhong, 1997: El Niño, La Niña, and the nonlinearity of their teleconnections. J. Climate, 10 , 17691786.

    • Search Google Scholar
    • Export Citation
  • Hosking, J. R. M., 1990: L-moments: Analysis and estimation of distributions using linear combinations of order statistics. J. Roy. Stat. Soc., B52 , 105124.

    • Search Google Scholar
    • Export Citation
  • Jin, F-F., S-I. An, A. Timmermann, and J. Zhao, 2003: Strong El Niño events and nonlinear dynamical heating. Geophys. Res. Lett., 30 .1120, doi:10.1029/2002GL016356.

    • Search Google Scholar
    • Export Citation
  • 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:10.1029/2001JD000393.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., 1999: Genesis and evolution of the 1997–1998 El Niño. Science, 283 , 950954.

  • Monahan, A. H., 2004: A simple model for the skewness of global sea surface winds. J. Atmos. Sci., 61 , 20372049.

  • Monahan, A. H., and A. Dai, 2004: Spatial and temporal structure of ENSO nonlinearity. J. Climate, 17 , 30263036.

  • Philander, S. G., 1990: El Niño, La Niña, and the Southern Oscillation. Academic Press, 289 pp.

  • Pope, V. D., M. L. Gallani, P. R. Rowntree, and R. A. Stratton, 2000: The impact of new physical parameterizations in the Hadley Centre climate model—HadAM3. Climate Dyn., 16 , 123146.

    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108 .4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Roberts, M. J., and Coauthors, 2004: Impact of an eddy-permitting ocean resolution on control and climate change simulations with a global coupled GCM. J. Climate, 17 , 120.

    • Search Google Scholar
    • Export Citation
  • Rodgers, K. B., P. Friederichs, and M. Latif, 2004: Tropical Pacific decadal variability and its relation to decadal modulation of ENSO. J. Climate, 17 , 37613774.

    • Search Google Scholar
    • Export Citation
  • Timmermann, A., and F-F. Jin, 2002: A non-linear mechanism for decadal El Niño amplitude changes. Geophys. Res. Lett., 29 .1003, doi:10.1029/2001GL013369.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., G. W. Branstator, D. Karoly, A. Kumar, N. C. Lau, and C. Ropelewski, 1998: Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophys. Res., 103 , 1429114324.

    • Search Google Scholar
    • Export Citation
  • Vialard, J., C. Menkes, J-P. Boulanger, P. Delecluse, E. Guilyardi, M. J. McPhaden, and G. Madec, 2001: A model study of oceanic mechanisms affecting equatorial Pacific sea surface temperature during the 1997/98 El Niño. J. Phys. Oceanogr., 31 , 16491675.

    • Search Google Scholar
    • Export Citation
  • Wang, B., and S-I. An, 2002: A mechanism for decadal changes of ENSO behavior: Roles of background wind changes. Climate Dyn., 18 , 475486.

    • Search Google Scholar
    • Export Citation
  • Wang, W., and M. J. McPhaden, 2000: The surface-layer heat balance in the equatorial Pacific Ocean. Part II: Interannual variability. J. Phys. Oceanogr., 30 , 29893008.

    • Search Google Scholar
    • Export Citation
  • White, H. G., 1980: Skewness, kurtosis and extreme values of Northern Hemisphere geopotential heights. Mon. Wea. Rev., 108 , 14461455.

    • Search Google Scholar
    • Export Citation
  • Wu, A., and W. W. Hsieh, 2003: Nonlinear interdecadal changes of the El Niño–Southern Oscillation. Climate Dyn., 21 , 719730.

  • View in gallery
    Fig. 1.

    (a) Time mean SST (°C), (b) depth–longitude cross sections of the mean upper-ocean temperature (°C), and (c) zonal current (cm s−1) in DJF along the equator (2.5°S–2.5°N) in the Pacific. Positive values are denoted by full lines and negative values are denoted by dashed lines.

  • View in gallery
    Fig. 2.

    Time series of monthly Niño-3 SST index anomalies (°C) for 100-yr data. (a) HadISST (1901–2000) and (b) HadCEM. Dotted lines are lines of zero and 1.5 standard deviation.

  • View in gallery
    Fig. 3.

    Distribution of the skewness of tropical Pacific seasonal mean SST for 100 yr: (a) HadISST (1901–2000) and (b) HadCEM. Positive values are denoted by full lines and negative values are denoted by dashed lines.

  • View in gallery
    Fig. 4.

    Composite SST anomalies in DJF at the peak of ENSO for observations (HadISST): (a) for El Niño, (b) for La Niña, and (c) their asymmetry. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% [90% in (c)] confidence level using the t test.

  • View in gallery
    Fig. 5.

    Same as in Fig. 4, but for the model simulation. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

  • View in gallery
    Fig. 6.

    Composite ocean temperature tendency (°C month−1) during the development of ENSO (Jan–Dec) due to the advection of anomalous temperature gradient by the mean current: (a) for El Niño, (b) for La Niña, and (c) their asymmetry. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

  • View in gallery
    Fig. 7.

    Same as in Fig. 6, but due to the advection of mean temperature gradient by the anomalous current.

  • View in gallery
    Fig. 8.

    Same as in Fig. 6, but due to the advection of anomalous temperature gradient by anomalous current (NDH).

  • View in gallery
    Fig. 9.

    Same as in Fig. 6, but due to (a), (b) the zonal advection, (c), (d) the meridional advection, and (e), (f) the vertical advection: (a), (c), (e) for El Niño and (b), (d), (f) for La Niña.

  • View in gallery
    Fig. 10.

    The composite evolution of upper-ocean temperature anomalies (°C) and 20°C isotherm (thick) along the equatorial tropical Pacific (2.5°S–2.5°N): (a), (d) four seasons before the peak; (b), (e) two seasons before the peak and (c), (f) at the peak of ENSO; (left) for El Niño and (right) for La Niña. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

  • View in gallery
    Fig. 11.

    Time (downward)–longitude cross sections of the composite monthly mean anomalies along the equatorial tropical Pacific (2.5°S–2.5°N): (a), (b) surface zonal wind stress (Nm−2); (c), (d) upper 50-m ocean current (cm s−1); and (e), (f) upper 200-m ocean heat content (OHC; °C) (left) for El Niño and (right) for La Niña. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

  • View in gallery
    Fig. 12.

    Composite zonal wind stress (Nm−2) and precipitation (mm day−1) anomalies in DJF at the peak of ENSO for the model simulation: (a), (c) for El Niño and (b), (d) for La Niña. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

  • View in gallery
    Fig. 13.

    Scatterplots of (a) monthly zonal wind stress (Nm−2) and (b) precipitation (mm day−1) anomalies averaged over the western and central tropical Pacific (10.0°S–2.5°N, 140°E–150°W) against Niño-3 SST anomalies (°C). The corresponding linear regression equations for positive and negative SST anomalies are shown on top and bottom, respectively. The results that do not include months during which the Niño-3 anomaly exceeds 1.5 of its standard deviation are in brackets. The 95% confidence intervals for the slope of regression are given in brackets after the regression equations.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 191 47 3
PDF Downloads 160 133 2

Asymmetry between El Niño and La Niña in a Global Coupled GCM with an Eddy-Permitting Ocean Resolution

Buwen DongHadley Centre for Climate Prediction and Research, Met Office, Exeter, United Kingdom

Search for other papers by Buwen Dong in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

Observations show the asymmetric nature of El Niño and La Niña sea surface temperature (SST) anomalies. Warm events are often stronger than cold events. This asymmetric behavior is an important feature that can be used to validate coupled models to test their ability to represent the climate system. The asymmetry of El Niño and La Niña SST anomalies has been investigated in a simulation of the Hadley Centre eddy-permitting coupled general circulation model. It is found that the asymmetric behavior is captured by the model with SST anomalies associated with strong El Niño events being greater than those associated with strong La Niña events. The pattern of the SST asymmetry also bears some similar characteristics to those based on observations despite the deficiency that SST anomalies associated with both El Niño and La Niña extend too far westward in the model.

Through a heat budget analysis of the ocean mixed layer, it is shown that nonlinear dynamic heating (NDH) is important in generating intense El Niño and the SST asymmetry between El Niño and La Niña events, especially in the eastern tropical Pacific Ocean. This nonlinear dynamic heating enhances the amplitude of El Niño and reduces the amplitude of La Niña, and therefore leads to the asymmetry between El Niño and La Niña events, with El Niño being stronger. However, the skewness and asymmetry in the model are relatively weak, being consistent with a relatively weak nonlinear dynamical heating. It is also shown that the eastward-propagating feature of subsurface anomalies provides a favorable phase relationship between temperature and current anomalies that results in strong nonlinear dynamical heating that tends to produce stronger El Niños. In addition, in the model simulation, the nonlinear nature of zonal wind stress anomalies between El Niño and La Niña events also plays an important role in the central tropical Pacific. These different mechanisms work constructively to determine the asymmetry between El Niño and La Niña events in the model, and they are similar to those proposed in recent studies based on observations. The ability of the model to simulate this asymmetric feature is encouraging and offers hope to the challenge of predicting the amplitude of strong El Niño events.

Corresponding author address: Dr. Buwen Dong, NCAS Centre for Global Atmospheric Modelling, Department of Meteorology, University of Reading, Reading RG6 6BB, United Kingdom. Email: swsdong@met.reading.ac.uk

Abstract

Observations show the asymmetric nature of El Niño and La Niña sea surface temperature (SST) anomalies. Warm events are often stronger than cold events. This asymmetric behavior is an important feature that can be used to validate coupled models to test their ability to represent the climate system. The asymmetry of El Niño and La Niña SST anomalies has been investigated in a simulation of the Hadley Centre eddy-permitting coupled general circulation model. It is found that the asymmetric behavior is captured by the model with SST anomalies associated with strong El Niño events being greater than those associated with strong La Niña events. The pattern of the SST asymmetry also bears some similar characteristics to those based on observations despite the deficiency that SST anomalies associated with both El Niño and La Niña extend too far westward in the model.

Through a heat budget analysis of the ocean mixed layer, it is shown that nonlinear dynamic heating (NDH) is important in generating intense El Niño and the SST asymmetry between El Niño and La Niña events, especially in the eastern tropical Pacific Ocean. This nonlinear dynamic heating enhances the amplitude of El Niño and reduces the amplitude of La Niña, and therefore leads to the asymmetry between El Niño and La Niña events, with El Niño being stronger. However, the skewness and asymmetry in the model are relatively weak, being consistent with a relatively weak nonlinear dynamical heating. It is also shown that the eastward-propagating feature of subsurface anomalies provides a favorable phase relationship between temperature and current anomalies that results in strong nonlinear dynamical heating that tends to produce stronger El Niños. In addition, in the model simulation, the nonlinear nature of zonal wind stress anomalies between El Niño and La Niña events also plays an important role in the central tropical Pacific. These different mechanisms work constructively to determine the asymmetry between El Niño and La Niña events in the model, and they are similar to those proposed in recent studies based on observations. The ability of the model to simulate this asymmetric feature is encouraging and offers hope to the challenge of predicting the amplitude of strong El Niño events.

Corresponding author address: Dr. Buwen Dong, NCAS Centre for Global Atmospheric Modelling, Department of Meteorology, University of Reading, Reading RG6 6BB, United Kingdom. Email: swsdong@met.reading.ac.uk

1. Introduction

The El Niño–Southern Oscillation (ENSO) phenomenon is the largest natural interannual climate fluctuation in the coupled ocean–atmosphere system (e.g., Bjerknes 1969; Philander 1990). These events have a major impact not only on the climate of the Pacific region where their centers of action lie, but also on the entire global climate system through atmospheric teleconnections from regions of anomalous tropical heating (Glantz 1996; Trenberth et al. 1998).

One of the major features of ENSO is the clear asymmetry between the warm phase, El Niño, and the cold phase, La Niña, with El Niño SST anomalies being greater than those of La Niña. Therefore, the average El Niño and La Niña events are not mirror images. This has been noted in several studies (Hoerling et al. 1997; Burgers and Stephenson 1999; Jin et al. 2003; Hannachi et al. 2003; An and Jin 2004; Monahan and Dai 2004; Rodgers et al. 2004). However, most current coupled models either fail to simulate this asymmetry or exhibit an asymmetry that is rather different in character from observations (Hannachi et al. 2003; Monahan and Dai 2004). Therefore, the correct simulation of the asymmetry between El Niño and La Niña events is still a challenge for coupled models.

The physical mechanisms responsible for the asymmetry between El Niño and La Niña are still not fully understood. Suggested mechanisms include nonlinearities in ocean dynamical heating (Jin et al. 2003; An and Jin 2004), nonlinearities in the dependence of tropical deep convection on the underlying SST (Hoerling et al. 1997), and asymmetry of wind stress in the western and central tropical Pacific Ocean (Kang and Kug 2002; Wu and Hsieh 2003; Monahan 2004).

Nonlinear interdecadal changes in the ENSO phenomenon are investigated by Wu and Hsieh (2003). They reveal notable interdecadal changes of ENSO behavior before and after the mid-1970s climate regime shift, with greater nonlinearity found during 1981–99 than during 1961–75. Spatial asymmetry (for both SST and wind stress anomalies) between warm El Niño and cold La Niña events was significantly enhanced in the later period when the ENSO characteristics tended to propagate predominantly eastward (Jin et al. 2003; An and Jin 2004) rather than westward in the previous period.

In two recent studies by Jin et al. (2003) and An and Jin (2004), it is proposed that the nonlinear dynamical heating (NDH) in the tropical Pacific Ocean is important in the generation of intense El Niño as well as the observed asymmetry between El Niño and La Niña events in the last decade. However, both of these studies are based on observations and an ocean analysis dataset. This study will focus on a coupled GCM with an eddy-permitting ocean component to consider the extent to which the coupled model can simulate the spatial structure of ENSO nonlinearity and to further elucidate the physical mechanisms. Two specific questions we address in this paper are as follows. 1) Does the model simulate the asymmetry between El Niño and La Niña events? 2) If it does, what are the physical mechanisms for this asymmetry?

The paper is organized as follows. The data and coupled model used are given in section 2. The model mean climate in the tropical Pacific is briefly discussed in section 3. The asymmetry between El Niño and La Niña events in the model simulation and its comparison with observations are presented in section 4. The physical mechanisms responsible for this asymmetry are elucidated in section 5. The features of subsurface temperature anomalies associated with El Niño and La Niña events and their relation to the ENSO asymmetry are discussed in section 6. A summary and conclusions are given in section 7.

2. Data and coupled model used

The observed monthly sea surface temperatures (SSTs) used are the Hadley Centre Sea Ice and Sea Surface Temperature Dataset (HadISST) from 1901 to 2000 (Rayner et al. 2003).

The model we use is the Hadley Centre Coupled Eddy-Permitting Model (HadCEM). The atmospheric component of HadCEM is a version of the Hadley Centre Atmospheric Model (HadAM3; Pope et al. 2000) run with a horizontal latitude–longitude grid spacing of 2.5° × 3.75° and 19 vertical levels using a hybrid vertical coordinate. The oceanic component of HadCEM is a development of the HadCM3 (Gordon et al. 2000) ocean component, comprising a Bryan–Cox-type model (Cox 1984) with 40 unevenly spaced levels in the vertical and horizontal resolution on a 1/3° × 1/3° latitude–longitude grid. The control run, which uses preindustrial atmospheric greenhouse gas concentrations, is run for 150 yr. Detailed model description and model climate are given in Roberts et al. (2004), and they also show that the coupled model remains close to radiative equilibrium without spinup procedure and flux correction. Here the monthly mean data of the last 100-yr simulation is used for the present study.

3. Model mean tropical Pacific climate

The SST, thermocline, and equatorial currents in the tropical Pacific are among the most important aspects of a coupled climate model. The model climatology of SST, upper-ocean temperature, and zonal current at the equatorial tropical Pacific are shown in Fig. 1. Errors in the simulated SSTs are less than 1.0°C except along the western coast of South America where the model is about 1.0°–2.0°C too warm. This warming is related to the underestimation of marine stratus cumulus cloud in the atmospheric model (Pope et al. 2000).

A pronounced feature in Fig. 1b is that the thermocline (the layer where the sharp vertical temperature gradient separates the upper ocean from the abyssal deep ocean) is at a depth of about 165 m in the west and slopes upward east of about 160°E, rising to about 50 m in the east tropical Pacific. This feature is very similar to an ocean analysis shown in Fig. 2b of An and Jin (2004).

The Equatorial Undercurrent (EUC) has a maximum eastward speed of 60 cm s−1 and surface westward current of about 50–60 cm s−1. These are similar to the analysis shown in Fig. 2c of An and Jin (2004) although the surface current is overestimated in the model. Overall, as with temperature, the upper-ocean current structure simulated in the tropical Pacific should permit an adequate representation of the evolution of ENSO-type disturbances. The model performance of present-day climate in details is documented in Roberts et al. (2004).

4. Asymmetry between El Niño and La Niña

The interannual variability of the Niño-3 SST index (the mean SST in the box 5°S–5°N, 150°–90°W) in the eastern central Pacific has been used extensively to represent observed ENSO events. Shown in Fig. 2 are monthly Niño-3 SST index anomalies based on HadISST from 1901 to 2000 (Rayner et al. 2003) and 100-yr periods for the model simulation. The observed Niño-3 index has a standard deviation of 0.79°C (Table 1). The modeled index exhibits a standard deviation of 1.14°C. With this measure, the amplitude of ENSO is clearly too strong in HadCEM, and the reason is discussed in B. Dong et al. (2005, unpublished manuscript). As shown in Roberts et al. (2004), the power spectrum of monthly Niño-3 SST anomalies in HadCEM has a rather narrow peak between 3 and 3.5 yr and rather low power at longer periods compared to the spectrum from HadISST. Another important feature in Fig. 2 is the asymmetry between El Niño and La Niña events. Observations show that the magnitude of strong El Niño events exceeds 2.0°C in terms of Niño-3 SST index, while no La Niña events have such a large amplitude. This indicates asymmetry between the largest El Niño and La Niña events with stronger warm events than cold events. The model simulation (Fig. 2b) shows a similar characteristic. Strong warm events have a Niño-3 index that exceeds 3.0°C, while nearly all cold events have magnitude less than 3.0°C. The model therefore simulates at least qualitatively the asymmetry between El Niño and La Niña events with El Niño stronger than La Niña in terms of Niño 3 SST anomalies.

Statistically, one possible way to quantify deviations from normal distribution is to examine the skewness (Burgers and Stephenson 1999). It is a measure of the asymmetry of a probability distribution function and is zero for normal distribution. The skewness is defined as the normalized third moment
i1520-0442-18-16-3373-eq1
where Xi is the ith sample of variable X, X is the mean, and n is the number of samples.

Figure 3 shows the skewness of seasonal SST anomalies over the tropical Pacific Ocean. Observation shows that there is strong positive skewness in the central and eastern tropical Pacific with amplitude of about 0.5°–1.0°C and weak negative skewness in the equatorial western tropical Pacific and over the subtropical western Pacific (0.1°–0.5°C). The pattern of skewness for the model simulation bears similarity to the observed pattern with positive skewness in the central and eastern tropical Pacific and negative skewness in the western tropical Pacific. However, the magnitude (about 0.3°–0.5°C) of positive skewness in the central and eastern tropical Pacific is weaker than observed. Positive skewness also extends westward to about 170°E.

The statistical significance of skewness can be estimated from the standard error of skewness (White 1980) if the number of independent samples in an analyzed variable is known. Given an ENSO period of about 4–4.5 (3.5) yr in observations (model) and assumption of independence of El Niño and La Niña (An and Jin 2004), observations (model) consist of approximately 45–50 (57) independent events. Thus according to White (1980), the threshold for significant skewness at the 95% confidence level is about ±0.7 (±0.65) for observations (model). Figure 4a indicates that skewness in observations over the eastern tropical Pacific is significant at the 95% confidence level. However, this is not true for the model simulation (Fig. 4b). As Hannachi et al. (2003) pointed out, skewness is a nonrobust measure and can be sensitive to outliers. More robust and resistant approaches, such as L moments (Hosking 1990), can be used to assess deviations from normal distribution. The statistics of Niño-3 indices for observations and model simulation are summarized in Table 1. The skewness of Niño-3 SST in model is 0.26, only about one-third of observation (0.74). However, the L-moment ratio τ3 is 0.061 for the model, and it is 1.5 times of that (0.044) based on observations.

Another way to show the asymmetry between El Niño and La Niña events is to construct composites. The outliers of 1.5 standard deviations of Niño-3 SST and only those that are phase locked to the annual cycle (i.e., Niño-3 SST anomalies maximize in the Northern Hemisphere winter) are selected for the composites. With this measure, nine El Niño and eight La Niña events are selected for observations. Shown in Fig. 4 are the composite SST anomalies for El Niño and La Niña events at the peak season of ENSO based on observations, and their asymmetry, defined as the sum of El Niño and La Niña composites. Both the El Niño and La Niña composites exhibit extremes along the equator in the eastern equatorial tropical Pacific from 100° to 160°W. The asymmetry between El Niño and La Niña is seen with positive SST anomalies of above 2.0°C over the eastern tropical Pacific for the El Niño composite and of −1.5°C for La Niña composite, indicating that the magnitude of SST anomalies associated with El Niño is stronger, as illustrated by Fig. 4c. The cold anomalies associated with La Niña events displace westward relative to warm SST anomalies associated with El Niño events. As a result, the skewness is negative in the western tropical Pacific (Fig. 3a), and it is also indicated by negative SST anomalies of the asymmetry (Fig. 4c).

The composites for HadCEM, based on 12 El Niño and 10 La Niña events, are shown in Fig. 5. Positive SST anomalies associated with El Niño and negative SST anomalies associated with La Niña extend too far westward to 150°E. Despite this deficiency, the model simulates the asymmetry between El Niño and La Niña events with SST anomalies associated with El Niño being about 0.5°–1.0°C stronger in the central and eastern tropical Pacific, while the SST anomalies associated with La Niña are stronger by up to 0.5°C in the western tropical Pacific due to westward displacement of negative SST anomalies associated with La Niña. This fact is in line with observation.

The results clearly indicate that the model simulates the asymmetry between El Niño and La Niña events with SST anomalies associated with El Niño being stronger by up to 1.0°C in the central and eastern tropical Pacific and weaker in the western tropical Pacific by 0.5°C. This is in agreement with observations although skewness as well as asymmetry in the model is weaker than in observations. A question that arises here is what are the main mechanisms responsible for the asymmetry in the model simulation? We try to address this in next sections.

5. The role of the dynamic heating through advections

As demonstrated in Jin et al. (2003) and An and Jin (2004) based on observations, the strong 1982/83 and 1997/98 El Niño events partly arose because of the nonlinearity in the dynamic heating in the heat budget of the upper ocean where the nonlinear advection warming has the same sign for El Niño and La Niña events. The heat budget of the upper ocean can be obtained using the following temperature equation:
i1520-0442-18-16-3373-eq2
where T, u, υ, and w are temperature, zonal, meridional, and vertical velocities, respectively. The overbar and prime denote the climatological mean and monthly anomaly. The contributions from surface heat fluxes, subgrid-scale (e.g., oceanic diffusion), and submonth-scale (e.g., heat advection due to tropical instability wave) processes are denoted by the residual term R′. The first bracket on the right of the equation is the advection of mean temperature gradient by anomalous current (hereafter referred to as the anomalous advection), and the second bracket is the advection of anomalous temperature gradient by mean current (hereafter referred to as the mean advection). The third bracket is the advection of anomalous temperature gradient by anomalous current. that is, the nonlinear advection term that is named as the NDH by Jin et al. (2003) and An and Jin (2004). The heat budgets have been performed for the upper 50 m of the ocean. The results for the upper 100 m of the ocean are similar.

The mean advection anomalies during the development of ENSO (from January to December of year 0) for the composite El Niño and La Niña events and its asymmetry are shown in Fig. 6. It shows a warming rate of 0.5°–0.7°C month−1 from the central to the eastern tropical Pacific for El Niño events and a similar magnitude of cooling rate for La Niña events. As a result, the nonlinearity of the mean advection anomalies is weak and does not play a dominant role in the asymmetry between El Niño and La Niña events. The mean advection is predominantly due to the vertical advection of anomalous temperature gradient by mean upwelling, with mean advection of anomalous zonal and meridional temperature gradient playing a secondary role (not shown). This is in agreement with the study of Wang and McPhaden (2000) for interannual variability of the equatorial surface-layer heat balance based on observations. During the growing phase of El Niño (La Niña), the largest temperature anomalies are around the thermocline (see next section). Therefore, dT ′/dz in the upper ocean above the thermocline is negative (positive) for El Niño (La Niña) and − wdT ′/dz is positive (negative) for El Niño (La Niña).

The anomalous advection is shown in Fig. 7. It indicates that this term is stronger for the composite El Niño in the central Pacific but not in the eastern tropical Pacific. Individual terms for the anomalous advection indicate that the zonal advection dominates in the central Pacific, while the vertical advection dominates in the eastern tropical Pacific (not shown). Therefore, the asymmetry of the anomalous advection in Fig. 7c in the central tropical Pacific is predominantly due to the asymmetry in the zonal advection of mean temperature gradient by anomalous current and therefore asymmetry in the upper-ocean current. The asymmetry in upper-ocean current anomalies between El Niño and La Niña events is associated with asymmetry in zonal wind stress anomalies in the central and western tropical Pacific. This will be discussed in section 6.

The NDH is shown in Fig. 8. It indicates that the warming tendency due to the NDH that centered in the equatorial eastern tropical Pacific appears for both El Niño and La Niña events. For the composite El Niño events, the NDH rate is about 0.1°–0.3°C month−1 in the equatorial eastern tropical Pacific, while it has a similar magnitude for the composite La Niña events. As a result, the asymmetry of the NDH (Fig. 8c) is about 0.5°C month−1 in the eastern tropical Pacific. The results shown here based on a coupled model simulation support the conclusions based on the ocean analysis by Jin et al. (2003) and An and Jin (2004). It shows that the NDH tends to enhance El Niño development, while it tends to damp La Niña development. As a result, the SST anomalies associated with El Niño are stronger in the tropical eastern Pacific than those associated with La Niña. However, NDH in the model simulation is relatively weak in comparison with observations (e.g., Jin et al. 2003; An and Jin 2004). This, in turn, leads to weak skewness and asymmetry over the eastern tropical Pacific in the model.

Note that the asymmetry of the anomalous advection (Fig. 7c) is largest in the central Pacific whereas the asymmetry of the NDH (Fig. 8c) is largest in the eastern tropical Pacific. As shown in Figs. 3b and 5c, the SST asymmetry in the model extends too far westward. This implies that the asymmetry due to the anomalous advection in the model is too strong in the central Pacific and also may extend too far westward because of its dominant role in the SST asymmetry in the western and central tropical Pacific.

Separating the NDH into individual terms indicates that the zonal advective term is the strongest among the NDH, and this is illustrated in Fig. 9. This aspect does not support the conclusions of Jin et al. (2003) and An and Jin (2004), who indicate that the vertical advection term is the largest and that zonal advection is the second largest term. This difference is mainly due to the difference in the zonal wind stress anomalies and associated upwelling anomalies in their studies and the model study. As indicated in Jin et al. (2003) and An and Jin (2004), during the mature phase of El Niño, westerly zonal wind stress anomalies occur in the central and western tropical Pacific whereas easterly wind anomalies occurs in the eastern Pacific. Associated with the wind stress anomalies are the anomalous downwelling in the central to western tropical Pacific and anomalous upwelling in the eastern tropical Pacific. This leads to an enhanced vertical advection of anomalous warm waters, therefore accelerating the surface warming. Similarly, during the mature phase of La Niña, there is reduced upwelling in the eastern equatorial Pacific due to westerly wind anomalies. The result is that upwelling of anomalous cold subsurface waters in the surface layer is reduced, therefore slowing down surface cooling. However, this phase relationship between the anomalous vertical temperature gradient and anomalous upwelling is weak in the model. This weak phase relationship in the model is associated with westward displacement of SST anomalies and zonal wind stress anomalies for both warm and cold ENSO events (Fig. 11a) in comparison with observations shown for 1982/83 (Fig. 8a of An and Jin 2004) and 1997/98 (Fig. 3a of Jin et al. 2003) El Niños. The westward displacement of SST and zonal wind stress anomalies associated with ENSO leads to the largest changes in upwelling occurring in the western and central tropical Pacific in the model (not shown) rather than in the central and eastern tropical Pacific as observations indicate (Fig. 3b of Jin et al. 2003), while the largest anomalous temperature gradient in the upper ocean occurs in the central and eastern tropical Pacific (Fig. 10). This, in turn, leads to weak −wdT ′/dz in the model.

6. Propagation characteristics of ENSO and the role of the NDH

Jin et al. (2003) and An and Jin (2004) concluded that ENSO having eastward-propagating features is favorable for strong NDH and therefore strong asymmetry between the El Niño and La Niña events based on ocean analysis data. In this section, the propagating characteristics of both El Niño and La Niña events are demonstrated for the model simulation. To examine how the ENSO events evolve within the upper thermocline, composites of the temperature anomalies in the upper ocean (top 200 m) averaged between 2.5°S and 2.5°N are shown in Fig. 10. During December–January–February (DJF−1), four seasons before the peak season of ENSO (Fig. 10a), a subsurface warm (cold) anomaly is found in the equatorial western and central Pacific in the thermocline, with small positive (negative) surface temperature anomalies for the El Niño (La Niña) events. From DJF−1 to June–July–August (JJA0), the subsurface temperature anomalies spread eastward and upward along the thermocline. By JJA0, subsurface temperature anomalies have intensified with maximum magnitude greater than 2.0°C. Additionally, a weak warm (cold) SST anomaly appears at the surface, marking the surface manifestation of the onset of the composite El Niño (La Niña) events. By DJF0, the season when ENSO peaks, the maximum warm (cold) anomaly appears in the far eastern Pacific at the thermocline. The maximum El Niño (La Niña) temperature anomalies at the thermocline in the eastern tropical Pacific are about 5.0°C (3.0°C) and 3.5°C (3.0°C) near the surface, marking the “mature phase” of ENSO and its asymmetry.

The asymmetry in the NDH due to anomalous zonal advection is further explained. As shown in Fig. 10, the positive (negative) temperature anomalies for El Niño (La Niña) events initiate around the thermocline in the tropical western Pacific. They spread eastward and upward along the thermocline during the development of El Niño (La Niña) events. During the development phase of ENSO, dT ′/dx is negative (positive) in the upper ocean eastward of maximum ocean temperature anomalies for El Niño (La Niña), while there is anomalous eastward (westward) ocean current associated with anomalous westerly (easterly) stress. That is, u′is positive (negative) during development of El Niño (La Niña). This is illustrated in Fig. 11. As a result, −udT ′/dx is positive during development of both El Niño and La Niña events. For westward propagation events, during the development phase of ENSO, dT ′/dx is positive (negative) in the upper ocean westward of maximum ocean temperature anomalies for El Niño (La Niña), while there is anomalous eastward (westward) ocean current associated with anomalous westerly (easterly) stress. That is, u′is positive (negative) during development of El Niño (La Niña). As a result, −udT ′/dx is negative during development of both El Niño and La Niña events for westward-propagating events. That is, NDH due to anomalous zonal advection tends to result in stronger La Niña events for westward propagation events. This indicates that only El Niño and La Niña events that have eastward propagation characteristics lead to asymmetry in the NDH that tends to produce stronger El Niño than La Niña events.

Another feature shown in Fig. 11 is the asymmetry in the zonal wind stress and upper-ocean current anomalies for the El Niño and La Niña events. The magnitude of anomalous westerly stress and the anomalous upper-ocean eastward current associated with the El Niño events are stronger than those of anomalous easterly stress and the anomalous upper-ocean westward current associated with the La Niña events. This asymmetry in wind stress anomalies and upper-ocean current anomalies leads to the asymmetry in the anomalous advection shown in Fig. 7 for the central tropical Pacific. Note that the asymmetry in the anomalous advection between the El Niño and La Niña events is due to the asymmetry in upper-ocean current and zonal wind stress anomalies. The asymmetry of zonal wind stress anomalies at the peak of ENSO is clearly indicated in Figs. 12a,b. Relative to the El Niño composite, the easterly stress anomalies associated with La Niña events displace westward, and the maximum of anomalies is about half of that associated with El Niño. This asymmetry in zonal wind stress anomalies is larger than the composites based on observations (Fig. 2 of Kang and Kug 2002). The asymmetry in zonal wind stress anomalies is associated with the asymmetry in convection (Figs. 12c,d). It shows that the maximum of precipitation anomalies during the peak of El Niño locates at about 180° with magnitude of about 10 mm day−1 whereas the maximum of precipitation anomalies (−4.0 mm day−1) during the peak of La Niña events locates in the western tropical Pacific at about 150°E. In terms of magnitude contrast of precipitation anomalies between El Niño and La Niña events, it is also stronger in the model than that based on observations (Fig. 5 of Hoerling et al. 1997). The results indicate that the asymmetry in convection and zonal wind stress anomalies in the western and central tropical Pacific between El Niño and La Niña events is overestimated in the model. This, in turn, leads to the strong asymmetry in the central tropical Pacific of the advection of mean temperature gradient by the anomalous current seen in Fig. 7 between El Niño and La Niña events.

However, the interpretation of the asymmetry in zonal wind stress and precipitation anomalies seen in Fig. 12 is complicated because the composites include the impact of the asymmetry of SST anomalies. Hoerling et al. (1997) showed that the tropical atmospheric responses are not linear for equal and opposite SST anomalies by idealized AGCM experiments. During El Niño, the convection anomalies appear over the central and eastern tropical Pacific whereas the anomalies associated with La Niña are more or less confined in the western central Pacific because of the dependence of active convection on the underlying value of SST (e.g., Gadgil et al. 1984). In the warm western tropical Pacific, small SST anomalies can induce large precipitation anomalies whereas positive SST anomalies with appreciable magnitude are required to induce convection in the cold eastern tropical Pacific. On the other hand, negative SST anomalies in the eastern tropical Pacific cold tongue have no further effect on regions having normally dry conditions (Hoerling et al. 1997). As a result, the convection anomaly during La Niña is shifted to the west in comparison to that of El Niño. This difference in convection anomalies can lead to a contrasting wind stress anomaly pattern for El Niño and La Niña (Hoerling et al. 1997; Kang and Kug 2002). The asymmetry in zonal wind stress and the convection shown in Fig. 12 is, therefore, partly reflecting the nonlinear atmospheric response to SST anomalies and partly reflecting the atmospheric response to the asymmetry of SST anomalies themselves for El Niño and La Niña. To further demonstrate that both processes are playing a role in the asymmetry in wind tress and precipitation anomalies, the scatterplots of zonal wind stress and precipitation anomalies against Niño-3 SST anomalies are shown in Fig. 13. The scatter diagram illustrates a stronger rate of change in the atmospheric response with respect to warm tropical Pacific SST forcing compared to cold SST forcing. This is clearly illustrated by the different linear regression equations for positive and negative SST anomalies. The linear regression coefficient of zonal wind stress (precipitation) against positive SST anomalies is 0.0082 Nm−2 °C−1 (2.154 mm day−1 °C−1), which is 1.50 (2.34) times that against negative SST anomalies. Excluding extreme months leads to the ratios of 1.30 and 2.18 for zonal wind stress and precipitation, respectively.

It is important to note, however, that the asymmetry in the NDH is not necessarily due to the asymmetry in upper-ocean current and zonal wind stress anomalies. Even for a purely symmetric distribution of upper-ocean current anomalies and ocean temperature anomalies of a warm and a cold event, this term is positive for both the warm and the cold event provided that both warm and cold events have eastward propagation characteristic. This eastward propagation characteristic is a prerequisite for the nonlinear dynamical heating to produce stronger SST anomalies in the eastern tropical Pacific for El Niño than La Niña events, similar to that illustrated in Jin et al. (2003) and An and Jin (2004) based on observations. They also show that for westward-propagating events, the nonlinear advective heating (vertical component) becomes much weaker because the phase relationship between the surface temperature and subsurface temperature becomes more in phase.

7. Conclusions and discussion

Observations show the asymmetric nature of El Niño and La Niña sea surface temperature anomalies. Warm events are often stronger than cold events. Therefore, the average El Niño and La Niña events are not mirror images. This has been noted in several studies (Hoerling et al. 1997; Burgers and Stephenson 1999; Timmermann and Jin 2002; Jin et al. 2003; Hannachi et al. 2003; An and Jin 2004; Monahan and Dai 2004; Rodgers et al. 2004). Recent studies by Jin et al. (2003) and An and Jin (2004) demonstrated that nonlinear dynamical heating is responsible for the asymmetry of El Niño and La Niña in the last two decades based on observations. Motivated by their study, this asymmetric behavior is investigated by the use of the Hadley Centre eddy-permitting coupled GCM. It has been shown that the model captures the observed asymmetry, with SST anomalies associated with strong El Niño events being greater than those associated with strong La Niña events, although both the skewness and asymmetry in the model are weaker than those in observations.

Through a heat budget analysis of the ocean mixed layer, it is shown that the nonlinear dynamic heating (NDH) in the eastern tropical Pacific ocean is important in generating intense El Niño and the asymmetry between El Niño and La Niña events as both the warm and cold events have eastward propagation characteristics. This eastward propagation characteristic is a prerequisite for strong nonlinear dynamical heating, which tends to produce strong El Niño events, as illustrated in Jin et al. (2003) and An and Jin (2004) based on observations. However, the asymmetry in the model is relatively weak, being consistent with a relatively weak nonlinear dynamical heating. In addition, in the model simulation, the nonlinear nature of zonal wind stress anomalies between El Niño and La Niña events also plays an important role in the central tropical Pacific. In this aspect, the model simulation is in agreement with studies by Kang and Kug (2002). The asymmetry in the zonal wind stress anomalies between El Niño and La Niña events in the model simulation is the combined effect of nonlinearities in the dependence of tropical deep convection on the underlying SST (Hoerling et al. 1997) and atmospheric response to the asymmetry of SST anomalies themselves. Further study is needed to assess their relative role quantitatively, and this is beyond the scope of this paper.

It has been recognized that ENSO behavior may be modulated by low-frequency decadal–interdecadal modes in the equatorial region (e.g., Wang and An 2002). Wang and An (2002) demonstrated that El Niño properties (period, amplitude, structure, and propagation) changed around 1976 in a coherent manner and further argued that these changes are associated with an interdecadal change in the background state. Before the 1976 climatic shift, ENSO events were characterized by westward-propagating anomalies, much less nonlinear heating during ENSO cycles, and therefore weak asymmetry between El Niño and La Niña events (Jin et al. 2003; An and Jin 2004). After the 1976 climatic shift, ENSO events were characterized by eastward-propagating anomalies with strong nonlinear dynamic heating during ENSO cycles. This nonlinear dynamic heating enhances the amplitude of El Niño and reduces the amplitude of La Niña, and therefore leads to asymmetry between El Niño and La Niña events with El Niño being stronger (Jin et al. 2003; An and Jin 2004).

The analysis of nonlinear dynamic heating provides only a rough indication of the importance of nonlinear dynamics in the tropical ocean–atmosphere interaction and the asymmetry between El Niño and La Niña. This asymmetry is therefore an additional measure to validate the ENSO in the coupled models. There are also other sources that can also induce the asymmetry between the El Niño and La Niña events. The tropical oceanic instability wave in the equatorial eastern Pacific is one of them. The increase of the meridional temperature gradient near the eastern tropical Pacific during La Niña enhances the tropical instability wave activity, which tends to enhance equatorward heat transport and to prevent the equatorial cold tongue from cooling down (Philander 1990; Vialard et al. 2001). During El Niño, the tropical instability wave is weak (Philander 1990; Vialard et al. 2001). Another source for the asymmetry is that the winds in the tropical western Pacific are characterized by westerly wind bursts (McPhaden 1999) but not easterly ones. An interesting question is to what extent is the SST anomaly asymmetry due to the westerly wind bursts in the western tropical Pacific associated with El Niño that are not matched by easterly wind bursts for La Niña. This clearly needs to be studied more in order to deepen our understanding of the asymmetry between El Niño and La Niña and to predict extreme events.

Most current coupled models either fail to simulate the asymmetry of SST anomalies between El Niño and La Niña or exhibit the asymmetry that is rather different in character from observations (Hannachi et al. 2003; Monahan and Dai 2004). It would be interesting to investigate the propagation features of ENSO and the nonlinearity of the atmospheric response to warm and cold SST anomalies in these models and to investigate reasons behind this failure.

Acknowledgments

This work is supported by the U.K. Government Meteorological Research Programme. Comments on the paper by Adam Scaife and particularly his suggestion of Fig. 13 are appreciated. The author also would like to thank Chris Folland for his comments on the paper and Michel Crucifix for his offline code to derive oceanic vertical velocity. The author is grateful to Fei-Fei Jin and an anonymous reviewer for their constructive comments. This paper is British Crown Copyright.

REFERENCES

  • An, S-I., and F-F. Jin, 2004: Nonlinearity and asymmetry of ENSO. J. Climate, 17 , 23992412.

  • Bjerknes, J., 1969: Atmospheric teleconnections from the equatorial Pacific. Mon. Wea. Rev., 97 , 163172.

  • Burgers, G., and D. B. Stephenson, 1999: The “normality” of El Nino. Geophys. Res. Lett., 26 , 10271030.

  • Cox, M. D., 1984: A primitive equation, 3-dimensional model of the ocean. GFDL Ocean Group Tech. Rep. 1, Princeton, NJ, 143 pp.

  • Gadgil, S., P. V. Josph, and N. V. Joshi, 1984: Ocean atmosphere coupling over monsoon regions. Nature, 312 , 141143.

  • Glantz, M. H., 1996: Currents of Change: El Niño’s Impact on Climate and Society. Cambridge University Press, 194 pp.

  • Gordon, C., C. Cooper, C. A. Senior, H. Banks, J. M. Gregory, T. C. Johns, J. F. B. Mitchell, and R. A. Wood, 2000: The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Climate Dyn., 16 , 147168.

    • Search Google Scholar
    • Export Citation
  • Hannachi, A., D. Stephenson, and K. Sperber, 2003: Probablity-based methods for quantifying nonlineararity in the ENSO. Climate Dyn., 20 , 241256.

    • Search Google Scholar
    • Export Citation
  • Hoerling, M. P., A. Kumar, and M. Zhong, 1997: El Niño, La Niña, and the nonlinearity of their teleconnections. J. Climate, 10 , 17691786.

    • Search Google Scholar
    • Export Citation
  • Hosking, J. R. M., 1990: L-moments: Analysis and estimation of distributions using linear combinations of order statistics. J. Roy. Stat. Soc., B52 , 105124.

    • Search Google Scholar
    • Export Citation
  • Jin, F-F., S-I. An, A. Timmermann, and J. Zhao, 2003: Strong El Niño events and nonlinear dynamical heating. Geophys. Res. Lett., 30 .1120, doi:10.1029/2002GL016356.

    • Search Google Scholar
    • Export Citation
  • 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:10.1029/2001JD000393.

    • Search Google Scholar
    • Export Citation
  • McPhaden, M. J., 1999: Genesis and evolution of the 1997–1998 El Niño. Science, 283 , 950954.

  • Monahan, A. H., 2004: A simple model for the skewness of global sea surface winds. J. Atmos. Sci., 61 , 20372049.

  • Monahan, A. H., and A. Dai, 2004: Spatial and temporal structure of ENSO nonlinearity. J. Climate, 17 , 30263036.

  • Philander, S. G., 1990: El Niño, La Niña, and the Southern Oscillation. Academic Press, 289 pp.

  • Pope, V. D., M. L. Gallani, P. R. Rowntree, and R. A. Stratton, 2000: The impact of new physical parameterizations in the Hadley Centre climate model—HadAM3. Climate Dyn., 16 , 123146.

    • Search Google Scholar
    • Export Citation
  • Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108 .4407, doi:10.1029/2002JD002670.

    • Search Google Scholar
    • Export Citation
  • Roberts, M. J., and Coauthors, 2004: Impact of an eddy-permitting ocean resolution on control and climate change simulations with a global coupled GCM. J. Climate, 17 , 120.

    • Search Google Scholar
    • Export Citation
  • Rodgers, K. B., P. Friederichs, and M. Latif, 2004: Tropical Pacific decadal variability and its relation to decadal modulation of ENSO. J. Climate, 17 , 37613774.

    • Search Google Scholar
    • Export Citation
  • Timmermann, A., and F-F. Jin, 2002: A non-linear mechanism for decadal El Niño amplitude changes. Geophys. Res. Lett., 29 .1003, doi:10.1029/2001GL013369.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., G. W. Branstator, D. Karoly, A. Kumar, N. C. Lau, and C. Ropelewski, 1998: Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophys. Res., 103 , 1429114324.

    • Search Google Scholar
    • Export Citation
  • Vialard, J., C. Menkes, J-P. Boulanger, P. Delecluse, E. Guilyardi, M. J. McPhaden, and G. Madec, 2001: A model study of oceanic mechanisms affecting equatorial Pacific sea surface temperature during the 1997/98 El Niño. J. Phys. Oceanogr., 31 , 16491675.

    • Search Google Scholar
    • Export Citation
  • Wang, B., and S-I. An, 2002: A mechanism for decadal changes of ENSO behavior: Roles of background wind changes. Climate Dyn., 18 , 475486.

    • Search Google Scholar
    • Export Citation
  • Wang, W., and M. J. McPhaden, 2000: The surface-layer heat balance in the equatorial Pacific Ocean. Part II: Interannual variability. J. Phys. Oceanogr., 30 , 29893008.

    • Search Google Scholar
    • Export Citation
  • White, H. G., 1980: Skewness, kurtosis and extreme values of Northern Hemisphere geopotential heights. Mon. Wea. Rev., 108 , 14461455.

    • Search Google Scholar
    • Export Citation
  • Wu, A., and W. W. Hsieh, 2003: Nonlinear interdecadal changes of the El Niño–Southern Oscillation. Climate Dyn., 21 , 719730.

Fig. 1.
Fig. 1.

(a) Time mean SST (°C), (b) depth–longitude cross sections of the mean upper-ocean temperature (°C), and (c) zonal current (cm s−1) in DJF along the equator (2.5°S–2.5°N) in the Pacific. Positive values are denoted by full lines and negative values are denoted by dashed lines.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 2.
Fig. 2.

Time series of monthly Niño-3 SST index anomalies (°C) for 100-yr data. (a) HadISST (1901–2000) and (b) HadCEM. Dotted lines are lines of zero and 1.5 standard deviation.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 3.
Fig. 3.

Distribution of the skewness of tropical Pacific seasonal mean SST for 100 yr: (a) HadISST (1901–2000) and (b) HadCEM. Positive values are denoted by full lines and negative values are denoted by dashed lines.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 4.
Fig. 4.

Composite SST anomalies in DJF at the peak of ENSO for observations (HadISST): (a) for El Niño, (b) for La Niña, and (c) their asymmetry. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% [90% in (c)] confidence level using the t test.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 5.
Fig. 5.

Same as in Fig. 4, but for the model simulation. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 6.
Fig. 6.

Composite ocean temperature tendency (°C month−1) during the development of ENSO (Jan–Dec) due to the advection of anomalous temperature gradient by the mean current: (a) for El Niño, (b) for La Niña, and (c) their asymmetry. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 7.
Fig. 7.

Same as in Fig. 6, but due to the advection of mean temperature gradient by the anomalous current.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 8.
Fig. 8.

Same as in Fig. 6, but due to the advection of anomalous temperature gradient by anomalous current (NDH).

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 9.
Fig. 9.

Same as in Fig. 6, but due to (a), (b) the zonal advection, (c), (d) the meridional advection, and (e), (f) the vertical advection: (a), (c), (e) for El Niño and (b), (d), (f) for La Niña.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 10.
Fig. 10.

The composite evolution of upper-ocean temperature anomalies (°C) and 20°C isotherm (thick) along the equatorial tropical Pacific (2.5°S–2.5°N): (a), (d) four seasons before the peak; (b), (e) two seasons before the peak and (c), (f) at the peak of ENSO; (left) for El Niño and (right) for La Niña. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 11.
Fig. 11.

Time (downward)–longitude cross sections of the composite monthly mean anomalies along the equatorial tropical Pacific (2.5°S–2.5°N): (a), (b) surface zonal wind stress (Nm−2); (c), (d) upper 50-m ocean current (cm s−1); and (e), (f) upper 200-m ocean heat content (OHC; °C) (left) for El Niño and (right) for La Niña. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 12.
Fig. 12.

Composite zonal wind stress (Nm−2) and precipitation (mm day−1) anomalies in DJF at the peak of ENSO for the model simulation: (a), (c) for El Niño and (b), (d) for La Niña. Positive values are denoted by full lines and negative values are denoted by dashed lines. Shading indicates that the anomalies are statistically significant at the 95% confidence level using the t test.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Fig. 13.
Fig. 13.

Scatterplots of (a) monthly zonal wind stress (Nm−2) and (b) precipitation (mm day−1) anomalies averaged over the western and central tropical Pacific (10.0°S–2.5°N, 140°E–150°W) against Niño-3 SST anomalies (°C). The corresponding linear regression equations for positive and negative SST anomalies are shown on top and bottom, respectively. The results that do not include months during which the Niño-3 anomaly exceeds 1.5 of its standard deviation are in brackets. The 95% confidence intervals for the slope of regression are given in brackets after the regression equations.

Citation: Journal of Climate 18, 16; 10.1175/JCLI3454.1

Table 1.

A summary of statistics of the monthly mean Niño-3 time series. Shown in the table are standard deviation (σ), skewness, and the L-moment ratio τ3 based on observations and model simulation.

Table 1.
Save