Strong Eastern Pacific El Niño events largely determine the spread in ENSO Amplitude and Asymmetry in Climate Models

doi:10.21203/rs.3.rs-3349269/v1

Download PDF

Research Article

Strong Eastern Pacific El Niño events largely determine the spread in ENSO Amplitude and Asymmetry in Climate Models

https://doi.org/10.21203/rs.3.rs-3349269/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

The El Niño/Southern Oscillation (ENSO) exhibits a considerable asymmetry in Sea Surface Temperature anomalies (SSTa), as El Niño events tend to be stronger and centered further east than La Niña events. Here we show by observations and climate models that the southward shift of the Intertropical Convergence Zone (ITCZ) onto the equator during strong Eastern Pacific (EP) El Niño events is essential for ENSO asymmetry, as this shifts the wind-SST feedback eastward by up to 15° compared to Central Pacific (CP) El Niño events, explaining the more eastern center in SSTa. Further, the thermocline feedback is enhanced in the east and in turn the Bjerknes feedback, which explains the large amplitude of the EP El Niño events. These results are confirmed by the CESM1 Large Ensemble. Further, it shows a statistically significant and strong (r ~ 0.8) link between internally generated differences in the fraction of strong EP events in 50-years long realizations and their overall ENSO amplitude and asymmetry. Finally, we analyze ENSO amplitude and asymmetry in preindustrial control simulations of 32 models participating in the Coupled Model Intercomparison Project phase 6 (CMIP6). The spread in ENSO amplitude across the model ensemble can be largely explained by the fraction of the strong EP El Niños. Half of the models strongly underestimate the ENSO asymmetry, which can be partly explained by the equatorial Pacific cold SST bias that hampers the southward migration of the ITCZ onto the equator during El Niño events, leading to fewer strong EP El Niño events.

The El Niño/Southern Oscillation (ENSO) is the dominant climate mode on interannual time scales, with major socio-economic and environmental impacts in the Pacific region and beyond (McPhaden 1999; McPhaden et al. 2020). El Niño, its warm phase, is associated with warmer than normal sea surface temperature (SST) in the central equatorial Pacific (CEP) and/or eastern equatorial Pacific (EEP). The warming increases the atmospheric convection over the CEP and weakens the easterlies over the western equatorial Pacific (WEP) (Gill 1980; Philander 1990). The latter in turn leads to a reduction of the thermocline slope across the equatorial Pacific, where a deeper thermocline in the EEP amplifies the initial warming (Bjerknes 1969).

During La Niña, the cold phase of ENSO, the situation is mostly opposite to that during El Niño. There are however noticeable asymmetries between El Niño and La Niña. SST anomalies (SSTa) tend to be stronger and located further to the east during El Niño than during La Niña (Takahashi et al. 2011; Dommenget et al. 2013; Capotondi et al. 2014; Timmermann et al. 2018). Further, El Niño events can either have peak SSTa in the central Pacific (CP) or eastern Pacific (EP), and EP events tend to have larger amplitudes than CP events (Capotondi et al. 2014; Timmermann et al. 2018).

Due to nonlinearities in the climate system ENSO is asymmetric and recent studies suggest that the asymmetry is mainly caused by nonlinearities in the atmospheric and oceanic feedbacks (see e.g. the review by An et al. 2020). The amplifying zonal wind feedback is stronger during El Niño than during La Niña (Frauen and Dommenget 2010; Karamperidou et al. 2017) as well as the damping shortwave feedback (Bellenger et al. 2014; Bayr et al. 2021). These atmospheric nonlinearities are largely related to the nonlinearity of deep convection that causes a larger convective response during El Niño than during La Niña (Cai et al. 2014; Srinivas et al. 2022). Other atmospheric nonlinearities contribute to the ENSO asymmetry such as state-dependent stochastic forcing by westerly wind bursts (Jin et al. 2007; Levine et al. 2016; Hayashi and Watanabe 2017). Oceanic processes also contribute to ENSO asymmetry, like nonlinearities in the advection related to tropical instability waves and low-frequency nonlinear dynamical heating (Jin 2003; Imada and Kimoto 2012). Moreover, a stronger subsurface-surface coupling during El Niño (Lübbecke and McPhaden 2017) and a larger dynamical ocean response per unit zonal wind stress change due to the eastward-shifted wind feedback during El Niño (Im et al. 2015) are other factors that contribute to ENSO asymmetry. Further, a study with an idealised model suggests a dominant role of the nonlinearity in atmospheric deep convection that subsequently triggers oceanic nonlinearities (Geng et al. 2019). Finally, teleconnections from the tropical Indian and Atlantic Oceans (Okumura and Deser 2010; An and Kim 2018) and the extratropics also may contribute to the ENSO asymmetry (Li et al. 2007; Wu et al. 2010).

Instrumental SST over the tropical Pacific is available since the late 19th century and ENSO amplitude varied substantially over this period (Bunge and Clarke 2009; Okumura et al. 2017). The Pacific Decadal Oscillation (PDO) and Interdecadal Pacific Oscillation (IPO) are possible factors influencing ENSO. For example, the EEP SST warmed after 1977 and cooled after 2000 (Mantua and Hare 2002; England et al. 2014; Latif et al. 2023), concurrent with the phase switch in the PDO (Power et al. 2021). The switch in the PDO phase may explain the change from an EP El Niño-dominated to a CP El Niño-dominated period around 2000 (Sun and Yu 2009; Hu et al. 2013; Lübbecke and McPhaden 2014; Timmermann et al. 2018; Fedorov et al. 2020). However, there is an ongoing debate about cause and effect. It is unclear for example, whether anomalously warm decades simply result from more El Niño than La Niña, or if decades with more El Niño are the result from the changes in the background climate state (Power et al. 2021).

Although large progress has been made in understanding and simulating ENSO, state-of-the-art climate models still display a large spread in simulated ENSO amplitude and most of them underestimate ENSO asymmetry (Guilyardi et al. 2009; Dommenget et al. 2013; Bellenger et al. 2014; Planton et al. 2021). The causes are under debate, but recent studies have shown that mean-state biases in the coupled models play an important role. The cold equatorial SST bias (also termed Cold Tongue bias) hampers a realistic simulation of the ENSO dynamics by shifting the rising branch of the Pacific Walker Circulation (PWC) too far to the west (Bayr et al. 2018, 2020). As a consequence, the amplifying wind feedback and the damping heat flux are both too weak, resulting in an error compensation and biased ENSO dynamics (Guilyardi et al. 2009; Lloyd et al. 2009, 2011, 2012; Kim et al. 2014; Bayr et al. 2019, 2021).

Here we investigate the role of atmospheric spatial asymmetry related to strong EP El Niño events in generating the ENSO asymmetry. While amplitude asymmetries of the feedbacks are relatively well described in literature (see the references above), the spatial asymmetries are harder to measure and less discussed. We therefore use the Center of Heat Index (CHI, Giese and Ray 2011), providing the amplitude-weighted location and spatially-averaged amplitude of a pattern, to quantify the amplitude and spatial asymmetries of the ENSO feedbacks.

To further elucidate the role of strong EP El Niño events in simulating a realistic ENSO amplitude and asymmetry we analyze 40 historical runs with the CESM1 Large Ensemble (CESM-LE), exhibiting a large internal variability in ENSO amplitude and asymmetry (Zheng et al. 2018; Cai et al. 2020). Finally, we investigate the spread in ENSO amplitude and asymmetry in 500-year long preindustrial control (pictl) simulations from a number of CMIP6 models (Eyring et al. 2016), and its relation to strong EP El Niño events.

The manuscript is organized as follows: In Section 2, we describe the data and methods used in this study. We then investigate the importance of the southward shift of the ITCZ onto the equator over the EEP for strong EP El Niño events and ENSO amplitude and asymmetry in observations (Section 3) and in CESM-LE (Section 4). In section 5, we explain how the spread in ENSO amplitude and asymmetry in the CMIP6 models is related to the occurrence rate of strong EP El Niño and mean state biases. This study is concluded with a summary and a discussion in Section 6.

Observed SSTs for the time period 1880–2019 are taken from HadISST (Rayner et al. 2003), ERSSTv5 (Huang et al. 2017), COBE-SSTv2 (Ishii et al. 2005), and Kaplan SST (Kaplan et al. 1998). We use the SST data only from 1880 onwards, because SST is uncertain due to sparse observations in the tropical Pacific before that date (Bunge and Clarke 2009). The atmospheric variables, i.e. zonal wind stress (TAUX) and precipitation (PR) are taken from ERA5 reanalysis over the period 1970–2019 (Hersbach et al. 2020). The subsurface oceanic potential temperature is taken from ORAS5 reanalysis over the period 1970–2019 (Zuo et al. 2019). The thermocline depth is estimated from the depth of the 20°C isotherm (Z20).

We analyze 40 members of a large ensemble from the CESM1-CAM5-BGC model (Kay et al. 2015, hereafter named CESM-LE) over the 1920–1969 period. All 40 runs start from the same initial state that is slightly perturbed (10^–14 °C) to generate different internal variability due to the so-called butterfly effect (Lorenz 1969). The model employs the 1920–1969 CMIP5 historical forcing, common to all members. Additionally, we analyze a 500-years long preindustrial control experiment of the same model (CESM-ctl hereafter).

We also use 32 preindustrial control (pictl) simulations that are 500-years long, from the CMIP6 database (see list in Fig. 9). The SST, TAUX and PR of those models are interpolated to the same 2.5° x 2.5° regular grid.

We use the following common index regions: the Niño3 region (90°W-150°W and 5°S-5°N), the Niño3.4 region (120°W-170°W and 5°S-5°N) and the Niño4 region (160°E-150°W and 5°S-5°N). Monthly-mean values are used in all the analyses. In the long datasets of observed SST and pictl runs, anomalies are calculated with respect to a 30-year running climatology. In all other datasets, we use anomalies relative to a fixed climatology for the whole period where a quadratic trend has been removed from the anomalies. Using detrended anomalies from a fixed climatology for all datasets has only minor effects on the results.

ENSO amplitude is measured by the standard deviation (stdv) of the SSTa in the Niño3.4 region, as the SSTa in the Niño3.4 region represents both El Niño and La Niña events properly (Trenberth 1997). To calculate the ENSO asymmetry we consider composites of El Niño and La Niña events (Fig. 1a,b), with the SSTa in the Niño3.4 region exceeding + 1.2 stdv as selection criterion for El Niño months and < -1.2 stdv for La Niña months (Dommenget et al. 2013). A joint measure of the ENSO amplitude and spatial asymmetry, hereafter “total asymmetry” is calculated from the difference of the El Niño and La Niña patterns (Fig. 1a,b) as average eastern box (140°W-80°W, 5°S-5°N) minus western box (140°E-160°W, 5°S-5°N) (Fig. 1c, Dommenget et al. 2013).

We also define spatial and amplitude asymmetry separate metrics using the Center of Heat Index (CHI, Giese and Ray 2011). The CHI longitude is the amplitude-weighted center of mass of the SSTa pattern, calculated as the average over all amplitude-weighted longitudes that exceed 0.5 stdv of the Niño3.4 index. The CHI amplitude is the average amplitude over all grid points that exceed 0.5 stdv of the Niño3.4 index. The spatial asymmetry is defined as the difference between the CHI longitude of the El Niño pattern minus the CHI longitude of the La Niña pattern. The amplitude asymmetry is defined as the difference between the CHI amplitude of the El Niño pattern minus the CHI amplitude of the La Niña pattern.

The same method is applied to calculate the TAUX, PR and Z20 total, spatial and amplitude asymmetry. For those variables, 1/3 of the maximum value of the pattern is used as the threshold to define the area over which the center and amplitude is calculated. The boxes that are used to calculate the total asymmetry for each variable in the reanalysis data and the CESM-LE are given in Table 1.

Table 1

List of the boundaries of boxes to calculate the total asymmetry of the feedbacks in observations/reanalysis data and the CESM-LE.
	Observations/Reanalysis data		CESM-LE
Variable	Western box	Eastern box	Western box	Eastern box
SST	140°E-160°W, 5°S-5°N	140°W-80°W, 5°S-5°N	125°E-175°E, 5°S-5°N	175°W-80°W, 5°S-5°N
TAUX	120°E-160°E, 5°S-5°N	180°E-140°W, 8°S-2°N	120°E-160°E, 5°S-5°N	170°E-120°W, 10°S-0°
PR	130°E-170°E, 5°S-5°N	180°E-140°W, 5°S-5°N	120°E-155°E, 5°S-5°N	165°E-140°W, 5°S-5°N
Z20	150°E-120°W, 5°S-5°N	110°W-80°W, 10°S-5°N	150°E-120°W, 5°S-5°N	110°W-80°W, 10°S-5°N

We distinguish between EP and CP El Niño events based on an adapted Trans-Niño-Index (TNI, Trenberth and Stepaniak 2001), measuring the normalized SSTa difference between the eastern (120°W-80°W, 5°S-5°N) and western Pacific (Niño4 region). We use a larger eastern box than Trenberth and Stepaniak (2001) to account for the weaker SST variability close to the South American coast found in many climate models. A positive TNI indicates EP El Niños and a negative TNI CP El Niños and strong EP El Niño months are defined as EP El Niño months during which the SSTa in the Niño3 region is larger than 1.5 K.

A Bjerknes feedback analysis is used to estimate the strength of the ENSO dynamics, which consists of three components. The wind-SST feedback is calculated by regressing the Niño4 averaged TAUX anomalies onto the Niño3 averaged SSTa. The wind-thermocline coupling is calculated as the Niño3 Z20 anomalies regressed onto the Niño4 TAUX anomalies. The subsurface-surface coupling (thermocline feedback) is defined as the Niño3 SSTa regressed onto the Niño3 Z20 anomalies.

On average El Niño events are stronger and located further east than La Niña events (Fig. 1a,b), as highlighted by the amplitude and longitude of the CHI, namely + 1.3 K and 129°W for El Niño and − 1.0 K and 140°W for La Niña. The sum of the two composites (Fig. 1c) displays a positive pole in the east close to the South American coast and a negative horseshoe-like pattern in the west. ENSO asymmetry is measured by the difference between the eastern and western pole (the two index boxes are indicated in Fig. 1c), as suggested by Dommenget et al. (2013).

To highlight the diversity of El Niño and La Niña events, Fig. 1e) displays a scatterplot of the amplitude vs. longitude of the CHI for each El Niño and La Niña month during 1880–2019. Note that Fig. 1e) is plotted based on the ensemble mean of the four observational SST datasets but is very similar in all four observational datasets (Fig. S1). The scatterplot presents a horn-shaped pattern during El Niño months, with increasing amplitude as the SST patterns moves eastward, and a peak around 120°W. There is a slight tendency for the La Niña amplitude to increase when the SST pattern is centered around 145°W, but it is much less pronounced than for El Niño. This distribution is related to the larger amplitudes of EP versus CP El Niño events, with the largest amplitudes observed during the strong EP El Niños of 1982-83, 1997-98 and 2015-16 (indicated in Fig. 1e by the triangles, stars and squares, respectively). The strong EP El Niño of 2015-16 (red squares) had its center around 128°W and a smaller amplitude than the strong EP El Niño of 1997-98, consistent with the weaker PR anomalies over the Niño3 region in 2015-16 (Santoso et al. 2017). The intensity of La Niña events only weakly depends on their center. Therefore, we do not distinguish between CP and EP La Niña events. Further, a necessary condition for a large-amplitude El Niño event is that it extends sufficiently eastward: all strong El Niño events are EP events, but not all EP events are strong (e.g. Santoso et al. 2017).

Regarding the multidecadal modulation of the ENSO amplitude, we notice a period with generally lower ENSO amplitude (LOW period: 1910–1969 highlighted in blue in Fig. 1d,e) than the periods before and after (HIGH periods: 1880–1910 & 1970–2019 highlighted in red in Fig. 1d,e). In both LOW and HIGH, El Niño events have similar amplitudes for centers over the CEP, but the LOW period is missing the strong EP El Niño events with the eastward-shifted center and large amplitudes (Fig. 1e). Consequently, the horn-shaped pattern is well developed in HIGH while it is less pronounced in LOW. HIGH includes 7 strong EP El Niño events (1888-89, 1896-97, 1902-03, 1972-73, 1982-83, 1997-98 and 2015-16) while LOW includes only one (1930-31). The La Niña months in the HIGH periods exhibit slightly larger amplitudes around 145°W than the LOW period, but the difference is not as pronounced as for the El Niño months. This is exemplified by the blue and red lines, which are the means over 10° chunks, that is in HIGH considerably higher in the EEP during El Niño events, but only slightly higher in the CEP during La Niña events (Fig. 1e).

The ENSO amplitude has a statistically significant correlation (r = 0.59) with the ENSO asymmetry across decades (Fig. 1f), suggesting a relation with strong EP El Niño events. For example, the period 1980–1999 with the two strongest EP El Niños is outstanding in terms of both ENSO asymmetry and amplitude (Fig. 1f). Thus, the observations though relatively short suggest that epochs with a high ENSO amplitude and spatial asymmetry correspond to decades when more strong EP El Niño events occur. This suggests that the more frequent occurrence of strong EP El Niño events, may it be due to internal variability or external forcing, dominates multidecadal variations of the ENSO amplitude and asymmetry. This has however to be considered with some caution, as the spread between the four observational data sets is quite large prior to 1960 (Fig. 1d,f), and because the observational record is short (8 strong EP events in total).

We will next relate the SST asymmetries to those of the atmospheric and subsurface oceanic ENSO signature. We compute El Niño and La Niña composites of zonal wind stress anomalies (TAUXa), precipitation anomalies (PRa) and thermocline anomalies (Z20a) utilizing ERA5 and ORAS5 over the 1970–2019 period (second HIGH period). The composites are normalized by the average Niño3.4 SSTa, to obtain a feedback (unit of atmospheric or oceanic response per Kelvin), similar to a regression. During El Niño events the warm SSTa drives an eastward shift of the ascending branch of the PWC, resulting in enhanced PR over the CEP and a weakening of the TAUX in the WEP (Fig. 2a,d). This in turn deepens the Z20 in the EEP that reinforces the initial warming of the SSTa in the east (Fig. 2g).

During La Niña events the cold SSTa causes a westward shift of the PWC that leads to a strengthening TAUX over the WEP (Fig. 2b,e). In turn, the thermocline shoals in the EEP reinforcing the initial cooling of the SSTa in the east (Fig. 2h). The TAUXa, PRa and Z20a response per unit of SST are larger and eastward shifted for El Niño relative to La Niña (11°, 27° and 15° for TAUXa, PRa and Z20a, respectively, in Fig. 2 left and middle columns). As a result, the difference between El Niño and La Niña exhibits a dipole pattern in all three variables (Fig. 2c,f,i), which is related to the spatial and amplitude asymmetry of the wind-SST, rainfall-SST and thermocline feedbacks between El Niño and La Niña events.

We then consider the amplitude and center of each El Niño and La Niña month for the feedbacks to obtain information about feedback diversity within the two ENSO phases as well as the differences between them and how the feedbacks contribute to ENSO asymmetry. The amplitudes are not normalized by the SSTa, in order to reveal the interplay between spatial and amplitude asymmetry that are both important for SSTa asymmetry. The CHI of the SSTa from ERA5 over the period 1970–2019 (second HIGH period) shown in Fig. 3a) is very similar to that depicted in Fig. 1e). The amplitude and center distribution of the TAUXa response is rather different between El Niño and La Niña events (Fig. 3b). Further, the amplitude distribution is skewed to the east for El Niño events with a maximum around 160°W and skewed to the west for La Niña events with a maximum around 175°E.

The PRa display remarkable asymmetries between El Niño and La Niña events (Fig. 3c): the spread in the center and the overall amplitude is much larger for El Niño than for La Niña events. Especially the strong EP El Niño months stick out with their large amplitude and easternmost centers.

In comparison to the other variables (Fig. 3a-c), Z20a has a remarkable amplitude asymmetry between El Niño and La Niña (Fig. 3d), probably due to a saturation effect: the climatological thermocline is shallow in the east, so that the ocean surface limits negative Z20a during La Niña. The strong EP El Niño events clearly stick out with the largest amplitudes and centers in the far EEP.

We next investigate the spatial difference in the atmospheric response during La Niña, CP El Niño and strong EP El Niño using full-field composites (shading in Fig. 4a-f) at the peak of the ENSO events (the months November-February (NDJF)). During La Niña events, there is a pronounced Cold Tongue and the ITCZ and the South Pacific Convergence Zone (SPCZ) are relatively far away from the equator (Fig. 4a,d). During CP El Niño events, the Cold Tongue is much weaker and the ITCZ and the SPCZ migrate closer to the equator, inducing considerable PR over the Niño4 region (Fig. 4b,e,k). During strong EP El Niño events, the Cold Tongue further weakens, with a nearly vanishing meridional SST gradient over the EEP, and the ITCZ migrates onto the equator over the Niño3 region with PR of more than 5 mm/day (Fig. 4c,f,j,k). The southward shift of the ITCZ onto the equator is also possible, as during strong EP El Niño in large parts of the Niño3 region the SST crosses the 28°C-threshold of deep convection (thick white line in Fig. 4c).

The composites of the SSTa, PRa and TAUXa for La Niña, CP El Niño and strong EP El Niño (contours in Fig. 4a-f and shadings in Fig. 4g-i) demonstrate that there is hardly any spatial asymmetry between La Niña and CP El Niño events, while there is considerable spatial asymmetry between La Niña and strong EP El Niño events. The southward migration of the ITCZ onto the equator over the Niño3 region (Fig. 4f) causes the eastward shift by 15° of the TAUX feedback during strong EP El Niños (Fig. 4i), which in turn leads to the stronger and more eastward located SSTa (Fig. 4c).

In summary we have shown the importance of the southward migration of the ITCZ onto the equator during strong EP El Niños for ENSO asymmetry, as the asymmetry between La Niña and CP El Niño is much weaker. Further, there are indications that the eastward shift of the atmospheric response during strong EP El Niños is essential for ENSO asymmetry, as it seems to be important for the higher and more eastern SSTa during these events. These results of course have to be considered with some caution due to the short period for which we have reliable observations and reanalysis data sets. Therefore we next investigate climate model output to support these results.

The previous section suggests a link between the rate of occurrence of strong EP El Niño events and multidecadal variations of the ENSO amplitude and asymmetry. A statistically significant relation can however not be demonstrated due to the relatively short period of reliable observations and reanalysis products. Thus, we now investigate this relation in CESM-LE, focusing on differences between ensemble members instead of variations between different decades. As the external forcing from the historical scenario of CMIP5 is identical, differences between the 40 members of the CESM-LE are solely due to internal variability. We analyze here the detrended anomalies during 1920–1969, a period in which the anthropogenic forcing is relatively weak. Note that each member is 50-years long, consistent with the length of epochs considered for observations (40–60 years, except for the first one at the beginning of the instrumental record).

The CESM1 realistically simulates the ENSO asymmetry in SST (Fig. 5), as well as the asymmetry of the TAUXa, PRa and Z20a feedbacks (Fig. S2). We note that the atmospheric feedbacks are slightly weaker and located further to the west than in observations. The center of the El Niño and La Niña patterns are indeed roughly shifted by 10° westward relative to observations (Fig. 1a,b and Fig. 5). We will thus use an adapted box for the total asymmetry calculation (125°E-175°E, 5°S-5°N and 80°W-175°W, 5°S-5°N).

The ENSO amplitude varies between 0.68 K and 1.18 K across ensemble members in the CESM-LE (Fig. 6a). We define a HIGH sub-ensemble as the 10 members with the largest ENSO amplitude (1.08 K on average) and a LOW sub-ensemble as the 10 members with the smallest ENSO amplitude (0.78 K on average). The ENSO SSTa composites (Fig. 5) indicate that the difference in ENSO amplitude and spatial asymmetry between the HIGH and LOW sub-ensembles are mainly related to El Niño events: the El Niño composites differ much more in HIGH and LOW than the La Niña composites.

The ENSO asymmetry in the individual members is strongly related to the ENSO amplitude (r = 0.82, Fig. 6a). This is due to the strong relation between the occurrence rate of strong EP El Niño events and both the ENSO amplitude (r = 0.81, Fig. 6b) and overall asymmetry (r = 0.85, Fig. 6c). More internally-generated strong EP El Niño events also go along with a weaker meridional temperature gradient between the Niño3 region and the region to the north (Fig. 6d) and increased PR in the Niño3 region (Fig. 6e) during El Niño events. The total asymmetry in the SST is strongly related to the spatial asymmetry of the PR (Fig. 6f), again highlighting the importance of the southward migration of the ITCZ onto the equator for ENSO asymmetry.

The importance of the strong EP El Niños for ENSO amplitude and asymmetry can be further illustrated by the center and amplitude of each ENSO event month for selected variables in the LOW and HIGH sub-ensemble (Fig. 7). The model realistically simulates the horn-shaped relation between the SSTa center and amplitude of the for El Niño events (Fig. 7a). Like in observations, this relation is much less pronounced for La Niña (Fig. 7a). In fact, there are considerably more strong EP El Niño events in the HIGH than in the LOW sub-ensemble (shown as stars in Fig. 7a). The differences between the two sub-ensembles are much weaker for La Niña events. Many of the strong La Niña events in the HIGH ensemble actually occur in the year after a strong EP El Niño event (stars in the lower half of Fig. 7a). Like for SSTa, the higher amplitude and eastward shifted TAUXa, PRa and Z20a signatures in the HIGH ensemble mainly result from the strong EP El Niño events (stars in Fig. 7b-d). The eastward shift in the HIGH sub-ensemble is most pronounced in PRa but also clearly visible in the other variables. For the La Niña events, the differences between the HIGH and LOW sub-ensembles are again much less pronounced and largely due to the La Niña events occurring after strong EP El Niño events.

The composites of La Niña, CP El Niño and strong EP El Niño events in the HIGH sub-ensemble show - similar to observations - only a weak asymmetry between La Niña and CP El Niño events, while they confirm the importance of strong EP El Niño events for ENSO asymmetry (Fig. 8). Further, it underlines that the equatorward migration of the ITCZ during strong EP El Niño events is important for the more eastern TAUXa feedback and hence the large and eastern SSTa.

In summary, the CESM-LE ensemble confirms the multidecadal relation suggested from observations. It indeed indicates that internally-generated differences in the number of strong EP El Niño events control the overall ENSO amplitude and asymmetry across 50-years long climate realizations from CESM-LE. Ensemble members that have more strong EP events also display a larger ENSO amplitude, and more eastward-shifted SSTa, PRa, TAUXa and Z20a El Niño signatures, and hence a larger overall ENSO asymmetry.

In section 4, we have shown that the number of strong EP El Niño events in ~ 50-years long periods determine the overall ENSO amplitude and spatial asymmetry during that period in CESM-LE. Next, we will test the hypothesis that the overall ENSO asymmetry in climate models is also related to the occurrence rate of strong EP events. We will also relate biases in the overall ENSO asymmetry and number of strong EP events to mean state biases. To that end, we analyze the CESM-ctl run and all preindustrial control (pictl) integrations of the CMIP6 models that are 500 years long, to get in the light of large internal variability in ENSO a robust measure of ENSO amplitude and asymmetry. The same boxes as that in the observational analysis (Fig. 1c) are used in the calculation of the asymmetry for all models.

The CMIP6 models exhibit a large spread in amplitude ranging from 0.43 K to 1.52 K and asymmetry ranging from − 0.08 K to 1.09 K (Fig. 9a). The ENSO amplitude and asymmetry are strongly related to each other (r = 0.82) in the models (Fig. 9a) and both quantities strongly depend on the fraction of strong EP El Niño events (fsEP) in the model (r = 0.88 and 0.84, respectively; Fig. 9b,c). Similar to what we found for the individual members in CESM-LE, there is a strong relation between the overall ENSO amplitude, asymmetry and occurrence rate of strong EP events in 500-years CMIP6 simulations. Half of the CMIP6 models strongly underestimate the ENSO asymmetry, with an ENSO asymmetry of less than half of the observed, and we define these models as the LOW sub-ensemble (blue symbols in Fig. 9a). The most realistic models in terms of amplitude and asymmetry are defined as the models that have more than half of the observed asymmetry but an amplitude of less than 1.2 K (MEDIUM sub-ensemble, black symbols in Fig. 9a). The models with the largest (and overestimated) ENSO amplitude (> 1.2 K) are clustered into the HIGH sub-ensemble. We note that CESM1 is one of the models that is close to observations in terms of both ENSO amplitude, asymmetry and the fraction of strong EP El Niño events (Fig. 9).

To further elucidate why the fraction of strong EP El Niño events differs in the individual models, we first consider the mean-state biases in the LOW, MEDIUM and HIGH sub-ensembles (Fig. 10). The models in the LOW sub-ensemble exhibit the largest Cold Tongue bias, while the bias is considerably weaker in the MEDIUM and smallest in the HIGH sub-ensemble (Fig. 10a-c). As a result, the largest underestimation of the equatorial PR is observed in the LOW sub-ensemble. The underestimation is less in the MEDIUM and smallest in the HIGH sub-ensemble (Fig. 10d-f). Although the HIGH sub-ensemble has the most realistic equatorial SST and PR, it has considerably higher SST and PR north and south of the equator similar to the other sub-ensembles (Fig. 10d-f).

The mean-state biases translate into biases in the NDJF signature of strong EP El Niño events, with distinct biases for the three sub-ensembles (Fig. 11). As a result of the strong cold bias, the LOW sub-ensemble still depicts a clear Cold Tongue during strong EP El Niño events (Fig. 11a) unlike in observations (Fig. 4c). The equatorial PR response is too weak and too westward, as only in the western Pacific the SST crosses the 28°C-threshold of deep convection (thick white line in Fig. 11a) and the ITCZ remains north of the Niño3 region due to a too strong meridional SST gradient over the EEP (Fig. 11d,j,k). As the equatorial convective response is too far west and too weak, the TAUXa feedback also acts too far west and is too weak (Fig. 11g), making strong EP El Niño events unlikely, because the strong EP El Niño events require a strong and more eastward TAUXa feedback to force large SSTa in the east.

The MEDIUM sub-ensemble has a weaker Cold Tongue than the LOW during strong EP events, but still a slightly overestimated cold tongue relative to observations. It also has a slightly underestimated equatorial PR response, as the meridional SST gradient over the EEP is too strong and the ITCZ too far north of the equator (Fig. 11b,e,j,k). Further, only a small part of the Niño3 region crosses the 28°C-threshold of deep convection, which also leads to a weaker convective response over the EEP. Nevertheless, the equatorial convective response and TAUXa feedback during strong EP El Niño events are more eastward and stronger in MEDIUM than in the LOW sub-ensemble (Fig. 11h), explaining the higher fraction of strong EP El Niño events.

The HIGH sub-ensemble has only a weak Cold Tongue during strong EP El Niño events, so that the 28°C-threshold of deep convection is crossed over most of the EEP and the meridional SST gradient nearly vanishes, allowing the ITCZ to migrate southward onto the equator (Fig. 11c,f,j). Both the equatorial convective response and the TAUXa response are relatively far in the east during strong EP El Niño events (Fig. 11i), and the likelihood of strong EP El Niño events is thus highest.

To complement those composite analyses, let us now turn to individual models (Fig. 12). The cold SST bias in the Niño3.4 region explains a large part of the inter-model spread in the fraction of strong EP events ( r = 0.70, Fig. 12a). On the one hand the cold SST bias weakens the TAUX feedback (Fig. 12b) due to a too westward position of the rising branch of the PWC that leads to biased ENSO dynamics (Guilyardi et al. 2009, 2020; Bayr et al. 2019, 2021). On the other hand, the cold SST bias hampers the southward migration of the ITCZ onto the equator with the result of fewer strong EP El Niño events (Fig. 11, Capotondi et al. 2020). The spatial asymmetry of the PR feedback is strongly related to the ENSO amplitude and asymmetry in the models (Fig. 12c,d), which again emphasizes the importance of the southward migration of the ITCZ for a realistic representation of both. In summary, a large part of the spread in ENSO amplitude and asymmetry in the CMIP6 ensemble can be explained by the fraction of the strong EP El Niño events. Importantly, the models’ ability to simulate strong EP El Niño events is hampered by the equatorial cold SST bias.

In this study, we analyzed ENSO amplitude and asymmetry in observations and climate models. ENSO amplitude and asymmetry varied substantially over the last 140 years. This variation appears to be related to the occurrence rate of strong EP El Niño events. Much less strong EP events occurred during 1910–1969 than in the periods 1880–1909 and 1970–2019, associated with a weaker ENSO amplitude and asymmetry (both in terms of amplitude and spatial pattern). The atmospheric feedbacks exhibit a spatial asymmetry, which is due to the equatorial convective response during strong EP El Niño events in association with the southward migration of the ITCZ onto the equator. The equatorial convective response is up to 36° further east during strong EP El Niño events than during La Niña events. Regarding the thermocline, the amplitude asymmetry is more pronounced than the spatial asymmetry. The likely reason is that Z20 is more sensitive to the more eastern TAUXa during strong EP El Niños (Im et al. 2015). Further, there is a natural limit to the shoaling of the thermocline during La Niña events. Finally, composite analyses indicate that La Niña and CP El Niño are mostly symmetric and that the ENSO asymmetry is mainly caused by strong EP El Niño events.

The observational record is too short to unambiguously establish a relation between decadal variations of the rate of strong EP events and the ENSO amplitude and asymmetry. We hence used a large ensemble of the CESM1 model, which display a rather realistic ENSO amplitude and asymmetry. The CESM-LE support the major observational findings. The ENSO amplitude and asymmetry vary substantially internally among the ensemble members. Both parameters are highly correlated and strongly influenced by the number of strong EP El Niño events. More specifically, ~ 50 years periods with more internally generated strong EP El Niño events also have a larger ENSO amplitude, and a more pronounced eastward shift of El Niño relative to La Niña events.

We then investigate the relation between strong EP events, the ENSO amplitude and asymmetry and mean state biases in 500-years simulations from the CMIP6 ensemble. There is also a large spread in the ENSO amplitude and asymmetry between models. Again, there is a strong correlation between the ENSO amplitude and asymmetry, as both strongly depend on the fraction of strong EP El Niño events. Half of the CMIP6 models simulate less than half of the observed ENSO asymmetry, which is a well-known problem of the climate models (Dommenget et al. 2013; Capotondi et al. 2014; Timmermann et al. 2018; Guilyardi et al. 2020; Planton et al. 2021). This can partly be explained by the equatorial cold SST bias in those models, which causes a too westward mean-state position of the rising branch of the PWC and in turn weakens the wind-SST feedback resulting in biased ENSO dynamics (Guilyardi et al. 2009, 2020; Bayr et al. 2018, 2019, 2020). Further, the cold bias hampers the ITCZ to migrate onto the equator during El Niño events (Cai et al. 2014; Capotondi et al. 2020). These two factors explain the lower fraction of strong EP El Niño events in models with a large equatorial cold SST bias which then results in unrealistically low ENSO asymmetry.

There is a long-standing discussion about the relative roles of nonlinearities in the atmospheric and oceanic feedbacks for the ENSO asymmetry, as summarized by An et al. (2020). The results presented here highlight the importance of the southward migration of the ITCZ onto the equator during the strongest El Niño events, as the spatial asymmetry of PR strongly increases with increasing ENSO amplitude and asymmetry (Fig. 3c, 6f, 7c and 12c,d). The spatial asymmetry increases strongly in the CESM-LE from the LOW to the HIGH ENSO amplitude sub-ensemble, yielding the more eastern TAUXa feedback in the HIGH sub-ensemble. Further, Z20 seems to be more sensitive to the more eastern wind-stress forcing during strong EP El Niño events, which is supported by a Bjerknes Feedback analysis (Table 2). The larger amplitude of strong EP El Niño events can be explained by a stronger Bjerknes feedback, because all three elements, the wind-SST feedback, the wind-thermocline feedback and the thermocline feedback in the east are stronger than during La Niña or CP El Niño events in both reanalysis data and the CESM-LE.

Table 2

Components of the Bjerknes feedback for different ENSO event types; SST and Z20 is averaged over the Niño3 region and TAUX over the Niño4 region.
		All months	La Niña only	El Niño only	CP El Niño	Strong EP El Niño
Reanalysis	TAUX on SST [10^− 2 Pa/K]	1.29	1.25	1.36	1.30	1.52
	Z20 on TAUX [10^− 2 m/Pa]	6.35	5.84	6.80	6.77	8.07
	SST on Z20 [K/cm]	4.85	4.15	5.50	4.46	8.21
CESM-LE HIGH	TAUX on SST [10^− 2 Pa/K]	1.02	0.97	1.05	1.01	1.26
	Z20 on TAUX [10^− 2 m/Pa]	6.33	6.40	6.28	5.75	7.09
	SST on Z20 [K/cm]	7.05	6.54	7.56	7.07	10.12

Further, Figs. 3 and 7 highlight that the dynamics of strong EP El Niño events may be hard to capture with traditional stationary indices. For example, traditional measures of the TAUX feedback such as TAUXa in the Niño4 region regressed on the Niño3 SSTa cannot adequately represent the large shift of the TAUXa feedback center under strong EP El Niño events, which seems to be essential for the large SSTa amplitudes in the far eastern Pacific.

Finally, we have to note that observations indicate a relation between the fraction of strong EP El Niño events (fsEP), ENSO amplitude and asymmetry on multidecadal timescales, but these results have to be considered with same caution, as observational records are relative short. We could demonstrate that this relation is present in the CESM1-LE, but it needs to be demonstrated if it holds also in other climate models. Further, the open question remains if the multidecadal variations of fsEP, ENSO amplitude and asymmetry are purely random or linked to some background mean state variations related to the PDO or IPO, as suggested by Zheng et al. (2018) or even driven by a self-modulating effect of ENSO, as suggested by Cai et al. (2020). These questions are beyond the scope of this manuscript and left for future studies.

In summary we have shown the great importance of the southward migration of the ITCZ onto the equator and associated eastward shift of the wind-SST feedback to generate strong EP El Niños that in turn determine ENSO amplitude and asymmetry. Due to the equatorial cold SST bias many climate models are not able to simulate this southward shift of ITCZ during El Niño and therefore underestimate ENSO asymmetry.

Version 1.1 from 12.9.2023 submitted to Climate Dynamics

Competing Interests

The authors have no relevant financial or non-financial interests to disclose.

Author Contributions

All authors contributed to the study conception and design. Data analysis and the first draft of the manuscript was performed by Tobias Bayr. Joke Lübbecke, Jérôme Vialard and Mojib Latif commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Data availability

The datasets analyzed during the study are available at:

HadISST: https://www.metoffice.gov.uk/hadobs/hadisst/

ERSSTv5, COBE-SSTv2 and Kaplan-SST: https://psl.noaa.gov/data/gridded/tables/sst.html

ERA5: https://cds.climate.copernicus.eu/#!/search?text=ERA5%20monthly&type=dataset

ORAS5: https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-oras5?tab=form

CESM-LE: https://www.cesm.ucar.edu/community-projects/lens/data-sets

CMIP6: https://esgf-node.llnl.gov/projects/cmip6/

An S, Il, Kim JW (2018) ENSO Transition Asymmetry: Internal and External Causes and Intermodel Diversity. Geophys Res Lett 45:5095–5104. https://doi.org/10.1029/2018GL078476
An S, Tziperman E, Okumura YM, Li T (2020) Chap. 7: ENSO Irregularity and Asymmetry. In: El Nino/Southern Oscillation in a changing climate. pp 153–172
Bayr T, Dommenget D, Latif M (2020) Walker Circulation controls ENSO Atmospheric Feedbacks in Uncoupled and Coupled Climate Model Simulations. Clim Dyn. https://doi.org/10.1007/s00382-020-05152-2
Bayr T, Drews A, Latif M, Lübbecke J (2021) The interplay of thermodynamics and ocean dynamics during ENSO growth phase. Clim Dyn 56:1681–1697. https://doi.org/10.1007/s00382-020-05552-4
Bayr T, Latif M, Dommenget D et al (2018) Mean-state dependence of ENSO atmospheric feedbacks in climate models. Clim Dyn 50:3171–3194. https://doi.org/10.1007/s00382-017-3799-2
Bayr T, Wengel C, Latif M et al (2019) Error compensation of ENSO atmospheric feedbacks in climate models and its influence on simulated ENSO dynamics. Clim Dyn 53:155–172. https://doi.org/10.1007/s00382-018-4575-7
Bellenger H, Guilyardi E, Leloup J et al (2014) ENSO representation in climate models: From CMIP3 to CMIP5. Clim Dyn 42:1999–2018. https://doi.org/10.1007/s00382-013-1783-z
Bjerknes J (1969) Atmospheric teleconnections from the equatorial Pacific. Mon Weather Rev 97:163–172. https://doi.org/10.1175/1520-0493(1969)097<0163:atftep>2.3.co;2
Bunge L, Clarke AJ (2009) A verified estimation of the El Ninõ index Ninõ-3.4 since 1877. J Clim 22:3979–3992. https://doi.org/10.1175/2009JCLI2724.1
Cai W, Borlace S, Lengaigne M et al (2014) Increasing frequency of extreme El Niño events due to greenhouse warming. Nat Clim Chang 4:111–116. https://doi.org/10.1038/nclimate2100
Cai W, Ng B, Geng T et al (2020) Butterfly effect and a self-modulating El Niño response to global warming. Nature 585:68–73. https://doi.org/10.1038/s41586-020-2641-x
Capotondi A, Wittenberg AT, Kug J et al (2020) Chap. 4: ENSO Diversity. In: El Niño Southern Oscillation in a Changing Climate. pp 65–86
Capotondi A, Wittenberg AT, Newman M et al (2014) Understanding ENSO diversity. Bull Am Meteorol Soc 921–938. https://doi.org/10.1175/BAMS-D-13-00117.1
Dommenget D, Bayr T, Frauen C (2013) Analysis of the non-linearity in the pattern and time evolution of El Niño southern oscillation. Clim Dyn 40:2825–2847. https://doi.org/10.1007/s00382-012-1475-0
England MH, Mcgregor S, Spence P et al (2014) Recent intensification of wind-driven circulation in the Pacific and the ongoing warming hiatus. Nat Clim Chang 4:222–227. https://doi.org/10.1038/nclimate2106
Eyring V, Bony S, Meehl GA et al (2016) Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci Model Dev 9:1937–1958. https://doi.org/10.5194/gmd-9-1937-2016
Fedorov AV, Hu S, Wittenberg AT et al (2020) Chap. 8: ENSO Low-Frequency Modulation and Mean State Interactions. In: El Nino/Southern Oscillation in a changing climate. pp 173–198
Frauen C, Dommenget D (2010) El Niño and La Niña amplitude asymmetry caused by atmospheric feedbacks. Geophys Res Lett 37:L18801. https://doi.org/10.1029/2010GL044444
Geng T, Cai W, Wu L, Yang Y (2019) Atmospheric Convection Dominates Genesis of ENSO Asymmetry. Geophys Res Lett 46:8387–8396. https://doi.org/10.1029/2019GL083213
Giese BS, Ray S (2011) El Niño variability in simple ocean data assimilation (SODA), 1871–2008. J Geophys Res 116:C02024. https://doi.org/10.1029/2010JC006695
Gill AE (1980) Some simple solutions for heat-induced tropical circulation. Q J R Meteorol Soc 106:447–462. https://doi.org/10.1002/qj.49710644905
Guilyardi E, Capotondi A, Lengaigne M et al (2020) Chap. 9: ENSO Modeling: History, Progress, and Challenges. In: El Niño Southern Oscillation in a Changing Climate. pp 201–226
Guilyardi E, Wittenberg AT, Fedorov A et al (2009) Understanding El Nino in Ocean–Atmosphere General Circulation Models: Progress and Challenges. Bull Am Meteorol Soc 90:325–340. https://doi.org/10.1175/2008BAMS2387.1
Hayashi M, Watanabe M (2017) ENSO Complexity Induced by State Dependence of Westerly Wind Events. J Clim JCLI-D. https://doi.org/10.1175/JCLI-D-16-0406.1. -16-0406.1-52
Hersbach H, Bell B, Berrisford P et al (2020) The ERA5 global reanalysis. Q J R Meteorol Soc 146:1999–2049. https://doi.org/10.1002/qj.3803
Hu ZZ, Kumar A, Ren HL et al (2013) Weakened interannual variability in the tropical pacific ocean since 2000. J Clim 26:2601–2613. https://doi.org/10.1175/JCLI-D-12-00265.1
Huang B, Thorne PW, Banzon VF et al (2017) NOAA Extended Reconstructed Sea Surface Temperature (ERSST), Version 5. https://doi.org/doi:10.7289/V5T72FNM. NOAA Natl Centers Environ Inf
Im S-H, An S-I, Kim ST, Jin F-F (2015) Feedback processes responsible for El Niño-La Niña amplitude asymmetry. Geophys Res Lett 42:5556–5563. https://doi.org/10.1002/2015GL064853
Imada Y, Kimoto M (2012) Parameterization of tropical instability waves and examination of their impact on ENSO characteristics. J Clim 25:4568–4581. https://doi.org/10.1175/JCLI-D-11-00233.1
Ishii M, Shouji A, Sugimoto S, Matsumoto T (2005) Objective analyses of sea-surface temperature and marine meteorological variables for the 20th century using ICOADS and the Kobe Collection. Int J Climatol 25:865–879. https://doi.org/10.1002/joc.1169
Jin F-F (2003) Strong El Niño events and nonlinear dynamical heating. Geophys Res Lett 30:1–4. https://doi.org/10.1029/2002GL016356
Jin FF, Lin L, Timmermann A, Zhao J (2007) Ensemble-mean dynamics of the ENSO recharge oscillator under state-dependent stochastic forcing. Geophys Res Lett 34:1–5. https://doi.org/10.1029/2006GL027372
Kaplan A, Cane MA, Kushnir Y et al (1998) Analyses of global sea surface temperature 1856–1991. J Geophys Res Ocean 103:18567–18589. https://doi.org/10.1029/97JC01736
Karamperidou C, Jin FF, Conroy JL (2017) The importance of ENSO nonlinearities in tropical pacific response to external forcing. Clim Dyn 49:2695–2704. https://doi.org/10.1007/s00382-016-3475-y
Kay JE, Deser C, Phillips A et al (2015) The community earth system model (CESM) large ensemble project: A community resource for studying climate change in the presence of internal climate variability. Bull Am Meteorol Soc 96:1333–1349. https://doi.org/10.1175/BAMS-D-13-00255.1
Kim ST, Cai W, Jin FF, Yu JY (2014) ENSO stability in coupled climate models and its association with mean state. Clim Dyn 42:3313–3321. https://doi.org/10.1007/s00382-013-1833-6
Latif M, Bayr T, Kjellsson J et al (2023) Strengthening atmospheric circulation and trade winds slowed tropical Pacific surface warming. Commun earth Environ 4:1–10. https://doi.org/10.1038/s43247-023-00912-4
Levine A, Jin FF, McPhaden MJ (2016) Extreme noise-extreme El Niño: How state-dependent noise forcing creates El Niño-La Niña asymmetry. J Clim 29:5483–5499. https://doi.org/10.1175/JCLI-D-16-0091.1
Li Y, Lu R, Dong B (2007) The ENSO-Asian monsoon interaction in a coupled ocean-atmosphere GCM. J Clim 20:5164–5177. https://doi.org/10.1175/JCLI4289.1
Lloyd J, Guilyardi E, Weller H (2011) The role of atmosphere feedbacks during ENSO in the CMIP3 models. Part II: using AMIP runs to understand the heat flux feedback mechanisms. Clim Dyn 37:1271–1292. https://doi.org/10.1175/JCLI-D-11-00178.1
Lloyd J, Guilyardi E, Weller H (2012) The role of atmosphere feedbacks during ENSO in the CMIP3 models. Part III: The shortwave flux feedback. J Clim 25:4275–4293. https://doi.org/10.1175/JCLI-D-11-00178.1
Lloyd J, Guilyardi E, Weller H, Slingo J (2009) The role of atmosphere feedbacks during ENSO in the CMIP3 models. Atmos Sci Lett 10:170–176. https://doi.org/10.1002/asl.227
Lorenz EN (1969) The predictability of a flow which possesses many scales of motion. Tellus 21:289–307. https://doi.org/10.3402/tellusa.v21i3.10086
Lübbecke JF, McPhaden MJ (2017) Symmetry of the Atlantic Niño mode. Geophys Res Lett 44:965–973. https://doi.org/10.1002/2016GL071829
Lübbecke JF, McPhaden MJ (2014) Assessing the twenty-first-century shift in enso variability in terms of the bjerknes stability index. J Clim 27:2577–2587. https://doi.org/10.1175/JCLI-D-13-00438.1
Mantua N, Hare S (2002) The Pacific decadal oscillation. J Oceanogr 58:35–44
McPhaden MJ (1999) The child prodigy of 1997–1998. Nature 398:559–562
McPhaden MJ, Santoso A, Cai W (2020) El Niño Southern Oscillation in a Changing Climate
Okumura YM, Deser C (2010) Asymmetry in the Duration of El Nino and La Nina. J Clim 23:5826–5843. https://doi.org/Doi 10.1175/2010jcli3592.1
Okumura YM, Sun T, Wu X (2017) Asymmetric Modulation of El Niño and La Niña and the Linkage to Tropical Pacific Decadal Variability. J Clim JCLI-D. https://doi.org/10.1175/JCLI-D-16-0680.1. -16-0680.1
Philander S (1990) El Niño, La Niña, and the southern oscillation. Academic Press, San Diego, USA
Planton YY, Guilyardi E, Wittenberg AT et al (2021) Evaluating climate models with the CLIVAR 2020 ENSO metrics package. Bull Am Meteorol Soc 102:E193–E217. https://doi.org/10.1175/BAMS-D-19-0337.1
Power S, Lengaigne M, Capotondi A et al (2021) Decadal climate variability in the tropical Pacific: Characteristics, causes, predictability, and prospects. Science (80-) 374:. https://doi.org/10.1126/science.aay9165
Rayner NA, Parker DE, Horton EB et al (2003) Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J Geophys Res 108:4407
Santoso A, McPhaden MJ, Cai W (2017) The Defining Characteristics of ENSO Extremes and the Strong 2015/2016 El Niño. Rev Geophys 55:1079–1129. https://doi.org/10.1029/88EO01108
Srinivas G, Vialard J, Lengaigne M et al (2022) Relative Contributions of Sea Surface Temperature and Atmospheric Nonlinearities to ENSO Asymmetrical Rainfall Response. J Clim 35:3725–3745. https://doi.org/10.1175/JCLI-D-21-0257.1
Sun F, Yu J-Y (2009) A 10–15-Yr Modulation Cycle of ENSO Intensity. J Clim 22:1718–1735. https://doi.org/10.1175/2008JCLI2285.1
Takahashi K, Montecinos A, Goubanova K, Dewitte B (2011) ENSO regimes: Reinterpreting the canonical and Modoki El Niño. Geophys Res Lett 38:L10704. https://doi.org/10.1029/2011GL047364
Timmermann A, An S, Kug J et al (2018) El Niño-Southern Oscillation Complexity. Nature 559:535–545. https://doi.org/10.1038/s41586-018-0252-6
Trenberth KE (1997) The definition of El Niño. Bull Am Meteorol Soc 78:2771–2778
Trenberth KE, Stepaniak DP (2001) Indices of El Nino evolution. J Clim 14:1697–1701. https://doi.org/10.1175/1520-0442(2001)014<1697:LIOENO>2.0.CO;2
Wu B, Li T, Zhou T (2010) Asymmetry of atmospheric circulation anomalies over the western north Pacific between El Niño and La Niña. J Clim 23:4807–4822. https://doi.org/10.1175/2010JCLI3222.1
Zheng XT, Hui C, Yeh SW (2018) Response of ENSO amplitude to global warming in CESM large ensemble: uncertainty due to internal variability. Clim Dyn 50:4019–4035. https://doi.org/10.1007/s00382-017-3859-7
Zuo H, Balmaseda MA, Tietsche S et al (2019) The ECMWF operational ensemble reanalysis-analysis system for ocean and sea ice: A description of the system and assessment. Ocean Sci 15:779–808. https://doi.org/10.5194/os-15-779-2019

StrongEPElNinosupplementv1.0.pdf

Download PDF

Version 1

posted

You are reading this latest preprint version

Strong Eastern Pacific El Niño events largely determine the spread in ENSO Amplitude and Asymmetry in Climate Models

Status:

Version 1

Abstract

Figures

1. Introduction

2. Data and Methods

3. ENSO amplitude and asymmetry in observations

4. ENSO amplitude and asymmetry in CESM-LE

5. ENSO amplitude and asymmetry in CMIP6 models

6. Summary and Discussion

Declarations

Competing Interests

Author Contributions

Data availability

References

Supplementary Files

Status:

Version 1